Understanding Human Time (Oxford Studies of Time in Language and Thought) 019289644X, 9780192896445

Table of contents :
Cover
Contents
General preface
Preface
The contributors
1 Introduction: Metaphysical time, human time, and time in language
2 Tense and emotion
3 An exploration into construals of subjective time in poetry
4 The 2D past
5 Counterfactuality as pragmatic inference in perspectival readings of Past Conditional utterances with modal verbs: Evidence from French
6 Avertive/frustrative markers in Australian languages: Blurring the boundaries between aspectuo-temporal and modal meanings
7 On modelling the future
8 Perceiving direction in directionless time
9 Temporal transparency and the flow of time
10 Does human time really flow? Metaindexicality, metarepresentation, and basic concepts
11 Underspecified temporal semantics in Pirah"00E3: Compositional transparency and semiotic inference
References
Index

Citation preview

Understanding Human Time

OX F OR D STUDIES OF TIME IN L A NGUAGE AND THOUGHT General Editors Kasia M. Jaszczolt, University of Cambridge and Louis de Saussure, University of Neuchˆatel Advisory Editors Nicholas Asher, Université Paul Sabatier; Johan van der Auwera, University of Antwerp; Robert I. Binnick, University of Toronto; Ronny Boogaart, University of Leiden; Frank Brisard, University of Antwerp; Patrick Caudal, CNRS; Anastasia Giannakidou, University of Chicago; Hans Kronning, University of Uppsala; Ronald Langacker, University of California, San Diego; Alex Lascarides, University of Edinburgh; Peter Ludlow, Northwestern University; Alice ter Meulen, University of Geneva; Robin Le Poidevin, University of Leeds; Paul Portner, Georgetown University; Tim Stowell, University of California, Los Angeles; Henriëtte de Swart, University of Utrecht PUBLISHED Time Language, Cognition, and Reality Edited by Kasia M. Jaszczolt and Louis de Saussure Future Times; Future Tenses Edited by Philippe De Brabanter, Mikhail Kissine, and Saghie Sharifzadeh Time, Language, and Ontology The World from the B-Theoretic Perspective by M. Joshua Mozersky The Present Perfective Paradox across Languages by Astrid De Wit Understanding Human Time Edited by Kasia M. Jaszczolt

Understanding Human Time Edited by K A SI A M . J A S Z C Z O LT

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © editorial matter and organization Kasia M. Jaszczolt 2023 © the chapters their several authors 2023 The moral rights of the authors have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2022948347 ISBN 978–0–19–289644–5 DOI: 10.1093/oso/9780192896445.001.0001 Printed and bound in the UK by Clays Ltd, Elcograf S.p.A. Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

Contents General preface Preface The contributors

1. Introduction: Metaphysical time, human time, and time in language Kasia M. Jaszczolt

vii viii xi

1

2. Tense and emotion Simon Prosser

11

3. An exploration into construals of subjective time in poetry Anna Piata

30

4. The 2D past Graeme A. Forbes

60

5. Counterfactuality as pragmatic inference in perspectival readings of Past Conditional utterances with modal verbs: Evidence from French Louis de Saussure 6. Avertive/frustrative markers in Australian languages: Blurring the boundaries between aspectuo-temporal and modal meanings Patrick Caudal

85

107

7. On modelling the future M. Joshua Mozersky

174

8. Perceiving direction in directionless time Matt Farr

199

9. Temporal transparency and the flow of time Giuliano Torrengo

220

10. Does human time really flow? Metaindexicality, metarepresentation, and basic concepts Kasia M. Jaszczolt

244

vi

CONTENTS

11. Underspecified temporal semantics in Pirahã: Compositional transparency and semiotic inference Daniel L. Everett

276

References Index

319 348

General preface The series Oxford Studies of Time in Language and Thought identifies and promotes pioneering research on the human concept of time and its representation in natural language. Representing time in language is one of the most debated issues in semantic theory and is riddled with unresolved questions, puzzles, and paradoxes. The series aims to advance the development of adequate accounts and explanations of such basic matters as (i) the interaction of the temporal information conveyed by tense, aspect, temporal adverbials, and context; (ii) the representation of temporal relations between events and states; (iii) human conceptualization of time; (iv) the ontology of time; and (v) relations between events and states (eventualities), facts, propositions, sentences, and utterances, among other topics. The series also seeks to advance time-related research in such key areas as language modelling in computational linguistics, linguistic typology, and the linguistic relativity/universalism debate, as well as in theoretical and applied contrastive studies. The central questions to be addressed concern the concept of time as it is lexicalized and grammaticalized in the different languages of the world. But its scope and the style in which its books are written reflects the fact that the representation of time interests those in many disciplines besides linguistics, including philosophy, psychology, sociology, and anthropology.

Preface Linguists, philosophers, psychologists, poets, physicists, but often also people across every known profession want to get to the bottom of the meaning of time. What does it mean that time flows? Scientists tell us that it is just a dimension of spacetime, so why can’t we conceptualize time as something akin to space (unless we try very, very hard)? Also, why are sci-fi novelists, film scriptwriters, and their respective readers and viewers so fascinated by time travel and its scientific underpinnings? Why does time fly when we want it to slow down and drags on and on when we want it to speed up? And, what exactly does it mean that ‘time speeds up’? Passing faster than a second per second? Most of all, how do all these queries and scraps of information and experience sum up to what we would like to call the human concept of time? Or, perhaps, ‘concepts’ if cross-cultural differences are better foregrounded than ironed out at some universal level? In fact, how much information about such a level of universal concepts do natural languages reveal? Or, indeed, how much do they reveal about culture- and language-specific concepts? All contributors to this unique collection have their own preferred selections and rankings of questions about time, spanning different areas of linguistics and philosophy. What unites us here is the goal of shedding more light on the human concept of time, approaching it from various angles. However, the resulting collection is more (not less) closely knit through its multidimensionality. Where most cross-disciplinary collections are found wanting is their unfulfilled aspiration to being truly inter-disciplinary. The idea behind composing this volume is that true interdisciplinarity can sparkle when ideas of different contributors are juxtaposed in a way that forms a particular roadmap through the tangle of questions about the human concept of time. Therefore, in the Introduction, I offer a range of possible dimensions that can be picked and chosen for such a roadmap, followed by one sample roadmap selected for this volume, pointing out at every step how a different path could have been taken, had different dimensions of inquiry been foregrounded at that particular step. The idea is simple: we know that meaning is dynamic and co-constructed, and as such does not break down neatly to particular speakers’ meanings (see e.g. Elder and Haugh 2018 for one such model and Jaszczolt and

PREFACE

ix

Berthon forthcoming for an overview). This also applies to the meta-level of concept construction or concept unpacking—and in this instance the human concept of time and its kins. Next, such a meta-level co-construction can itself be a many-phase process—beginning with externalizing ideas and brainstorming and ending with fitting the outcomes in a particular roadmap (see the order of the chapters in the Contents), but also, here, suggesting the options of other roadmaps (see Introduction), and finally leaving it to the readers to bounce the ideas off again for themselves. This produces an interactive approach to the search for understanding human time. In other words, the process begins in the most familiar way, with different individual (but calibrated) research questions and proposed solutions (phase 1), followed by bouncing them off other contributors’ questions and solutions—if successful, then to the benefit of such questions and solutions (phase 2). Here is where our workshop comes in: Understanding Human Time, held on 9–10 April 2021—sadly, not in Newnham College, Cambridge as planned, but virtually, due to the pandemic, but hopefully with no less interest and enthusiasm. The discussions it generated take us to phase 3—writing what follows, further thought through with the invaluable help of Reviewers to whom I hereby send my words of gratitude. Phase 4 is to be attributed to yours truly, the editor: extracting a range of dimensions of inquiry and using them as a guide for ordering the contributions and making them speak to each other in order to create a roadmap through them. The Reader can follow the roadmap or, with the help of my other examples of roadmaps in the Introduction, follow their own.¹ Doing the latter repeats the process, with all the advantages of such a multi-roadmap approach. This is where there is scope for interdisciplinarity to evolve further: depending on the selected roadmap and the dimensions according to which it is drawn, one can begin to address novel questions about human time and stumble across novel perspectives and answers. In addition to acknowledging the help of the Reviewers, I would also like to thank the OUP Commissioning Editor Julia Steer for her support and patience—it took longer than planned to put this volume together, facing various Covid-related restrictions and delays. Next, I owe thanks to Mireia Cabanes Calabuig for her editorial assistance, especially with collating the references. But most of all, I would like to thank my colleagues—linguists and philosophers who accepted my invitation to contribute to this rather special

¹ I followed this idea of a personalized journey on a much larger scale in my authored journey through semantics, pragmatics, and philosophy (Jaszczolt, 2023) to demonstrate how gathered wisdom can point to different novel ideas depending on a path taken.

x

PREFACE

workshop and the resulting volume. The book has also benefited from my participation in the research project CHRONOS: Rethinking and Communicating Time (PID2019-108762GB-I00) funded by the Spanish Ministry of Science and Innovation (MCIN/AEI/10.13039/501100011033). Cambridge, 1 August 2022

The contributors Patrick Caudal (PhD Université de Paris-Diderot 2000) is a research scientist at CNRS and U. Paris, Laboratoire de Linguistique Formelle, France. He is a specialist in the semantics and pragmatics of tense, aspect, and modality (TAM), and an Australianist focusing notably on the verb systems of non-Pama-Nyungan languages. He has extensively worked on semantic change, language comparison/typology and fieldwork linguistics applied to Australian languages as well as Romance languages, with a triple quantitative, experimental and formal perspective. He currently coordinates the FEMIDAL (‘Formal/Experimental Methods and In-depth Description of Australian Indigenous Languages’) CNRS international research network (2021–2025), as well as several Outreach/Science in Society research projects dedicated to Iwaidja, a non-Pama-Nyungan language spoken in the Cobourg Peninsula area. Daniel L. Everett is Trustee Professor of Cognitive Sciences, Bentley University in Waltham, Massachusetts, USA. He is best known for his work on Pirahã language and culture and for his arguments against sentential recursion postulated in Chomsky’s universal grammar. In addition to academic publications, he authored several popular science books, including How Language Began: The Story of Humanity's Greatest Invention (2017); Dark Matter of the Mind: The Culturally Articulated Unconscious (2016); Language: The Cultural Tool (2012); and Don't Sleep, There Are Snakes: Life and Language in the Amazonian Jungle (2008). He is currently working on two books on the linguistics and philosophy of Charles Sanders Peirce. Matt Farr is a philosopher of science and lecturer in the Department of History and Philosophy of Science at the University of Cambridge. His research focuses on the metaphysics and epistemology of time and causation, particularly what it means for time to have a direction, the role of causality in physics, and the psychology of time and causation. Matt’s research has been published in philosophy journals such as the British Journal for the Philosophy of Science and Synthese, and he is currently writing a book on the philosophy and physics of time direction. Matt received his PhD in Philosophy from the University of Bristol, and has held postdoctoral research fellowships at the University of Sydney and University of Queensland. Graeme A. Forbes is an Honorary Senior Lecturer in Philosophy at the University of Kent. He defends a dynamic view of time (The Growing-Block view) according to which change of time has no spatial equivalent and the past is settled while the future is open. His work is mainly in metaphysics, but encompasses topics from pragmatist philosophical methodology, though epistemology, philosophy of language, and philosophy of mind, to ethics.

xii

THE CONTRIBUTORS

Kasia M. Jaszczolt (D.Phil. Oxon, PhD Cantab, MAE) is a Professor of Linguistics and Philosophy of Language at the University of Cambridge and Professorial Fellow of Newnham College, Cambridge. She published extensively on topics in semantics, pragmatics, and philosophy of language. Her current interests are representation of time in language and developing a conceptualist-compositional theory of meaning in discourse. Her authored books include Semantics, Pragmatics, Philosophy: A Journey through Meaning (2023, CUP), Meaning in Linguistic Interaction (2016, OUP), Representing Time (2009, OUP), Default Semantics (2005, OUP), Semantics and Pragmatics (2002, Longman) and Discourse, Beliefs and Intentions (1999, Elsevier). She co-authored and co-edited twelve volumes, including The Cambridge Handbook of Pragmatics (2012, CUP). M. Joshua Mozersky is a Professor of Philosophy at Queen’s University in Kingston, Ontario. His research interests are in the philosophy of science, philosophy of language, metaphysics, and political philosophy. His essays have appeared, among other places, in Philosophical Studies, Synthese, International Studies in the Philosophy of Science, The Oxford Handbook of Time, and The Blackwell Companion to Time. His book, Time, Language, and Ontology, was published by Oxford University Press in 2015. Anna Piata is a postdoctoral researcher at the University of Neuchˆatel (Switzerland) and an adjunct lecturer at the University of Athens (Greece). She holds a PhD in Linguistics (University of Athens) and an MPhil in Linguistics (University of Cambridge). Her publications include articles in journals such as Pragmatics & Cognition, Journal of Pragmatics, Metaphor & Symbol, and Internet Pragmatics. She is the author of the monograph The Poetics of Time: Metaphors and Blends in Language and Literature (2018) and co-editor of the volume Time Representations in the Perspective of Human Creativity (2022). Her research interests are in cognitive semantics, pragmatics, and stylistics. Simon Prosser is Reader in Philosophy at the University of St Andrews. He is co-editor (with François Recanati) of Immunity to Error through Misidentification: New Essays (Cambridge University Press, 2012), and author of Experiencing Time (Oxford University Press, 2016). He has published articles on a variety of topics in philosophy of mind, philosophy of perception, and metaphysics, including consciousness, perspectival thought, temporal experience, mental files, and emergent properties. Much of his current work concerns the first-person perspective. Louis de Saussure is Professor of Linguistics at the University of Neuchˆatel (Switzerland). His main research areas are tense, aspect and modality, lexical semantics and pragmatics, post-Gricean pragmatics and persuasion in discourse. After completing a PhD at the University of Geneva, he lectured on philosophy of language at the University of Texas at Austin and pursued research at University College London as a visiting scholar. He further taught at various institutions including the École des Hautes Études en Sciences Sociales in Paris and the Universities of Fribourg, Lugano and Athens. He is a founding member of the Cognitive Science Centre at the University of Neuchˆatel and of the research network Beyond Meaning dedicated to expressivity and emotion in language use.

THE CONTRIBUTORS

xiii

Giuliano Torrengo is an Associate Professor at the Department of Philosophy ‘Piero Martinetti’ of the University of Milan, the founder and coordinator of the Centre for Philosophy of Time, and research fellow at the Autonomous University of Barcelona. He is the PI of the project CHRONOS: Rethinking and Communicating Time of the Spanish Ministry of Science and Innovation. Torrengo’s current focus of interest is the connection between temporal experience and the fundamental features of temporal reality.

1 Introduction Metaphysical time, human time, and time in language Kasia M. Jaszczolt

1.1 Time tout court and the dimensions of inquiry Understanding the human concept(s) of time straddles different disciplinary boundaries. This is a book for linguists and for philosophers—with the proviso that it aims at attracting both categories of readers (if indeed a boundary can be drawn) to both categories of papers. The aim is motivated by the big question of the characteristics of what we humans understand as the concept of time and feel as the passage of time. ‘Our time’, human time, is the time in thoughts, and as such, time in our epistemic attitudes such as belief (that time passes), knowledge (that death is inevitable), or fear (that I am going to be late). It is also the time of our feelings and sensations. It is the time that we (think we) experience, and as such it is the Big Unknown that, at one end, touches upon the real time of spacetime pursued through the laws of physics on the micro-level of human reality, and, at the other, time in language and discourse, pursued by linguists on the macro-level of social reality or, more traditionally (and formally) also in a bubble of an abstract construct of a language system. Both relations shed light on it, but they both still hide mysteries and secrets. They also obfuscate it by adding a dimension on which we have to calibrate what exactly we talk about when we talk about ‘time’. This volume touches upon them both, combining insights into the concept of time and the feeling of the passage of time (if indeed this is what it is— passage of time—but read on) with insights into (i) how best to conceive of real time (through questions to do with its nature, as static or dynamic), as well as research into (ii) using temporal expressions in natural language, or indeed explicitly speaking about time. In other words, windows on human time open from what we know and think we know (that is, have well-supported theories of ) about ourselves and our comprehension of the universe (here philosophical aspects come in) and what we know and think we know about Kasia M. Jaszczolt, Introduction. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Kasia M. Jaszczolt (2023). DOI: 10.1093/oso/9780192896445.003.0001

2

KASIA M. JASZCZOLT

our activities and behaviour in it—for the purpose of this volume, especially linguistic behaviour. Pertinent questions are ample and can be arranged along several different dimensions, such as the following. ‘↔’ stands for bi-directional travel (although some directions will be more well-trodden and more feasible than others). [1] the feeling, sensation, experience of time passing ↔ the concept of time; [2] speaking about time ↔ thinking about time; [3] thinking in time ↔ the concept of time; [4] the awareness of human time ↔ the awareness of real time of spacetime; [5] the properties of human time ↔ the properties of spacetime; [6] temporal thoughts ↔ modal thoughts; [7] temporal expressions ↔ expressions of modality and aspect; [8] consciousness ↔ human time; [9] human space ↔ human time; [10] feeling of time passing ↔ emotions, to name a few seminal ones. I have deliberately arranged this selection of dimensions in an order that does not attempt to distinguish between philosophical, linguistic, and psychological questions in that the very raison d’être of this collection of papers is to demonstrate the advantage of thinking outside such categories—not merely outside theories within fields but also really, truly, and freely as an organized whole, crossing disciplinary boundaries. Such aspects and dimensions sum up to an investigation into the reality of time tout court, normally separately pursued within the metaphysical, cognitive, and linguistic domain—with the input from microphysics at one end, and human sciences including poetics at the other. One of the basic questions is where time is to be found. On the level of the physical laws of the universe, the answer appears to be: as dimensions of spacetime. But it is still debated whether spacetime exists on the fundamental level of reality, or, rather is itself emergent, as has been suggested in theories of quantum gravity (see e.g. Rovelli 2018). And if it is emergent, one has to ask what kind of metaphysical emergence it is—that is, how time, as the emergent feature, relates to the feature from which it emerges: in addition to the assumed co-temporal material dependence, how ontologically and causally autonomous it has to be (see e.g. Wilson 2021, 2022 on strong and

INTRODUCTION

3

weak dependence and types of dependence; see also Baron 2019). As such, time need not be thought of as reducible to this lower level: emergence does not mean reduction (but see also Jaszczolt 2020 on the explanatory role of reductionism). Next, once we move to the ontologically autonomous level of organic life, and especially beings with conscious awareness, we can legitimately conceptualize this autonomy of time as fundamentality of time and as such spacetime. But this is only the first hurdle; the next one is the step from real time so conceived, time of the static universe, to the dynamic, flowing human time. Here, again, giving ourselves the option of moving back and forth along the relevant dimension (dimension 5) creates the much-needed conceptual frame in that, on the one hand, conceptualization of the universe specific to humans allows us specific insights into the symmetrical universe, obfuscating others, and barring yet others (dimension 5 above, forward direction; see e.g. Price 1996), while on the other, the understanding of the micro-laws of physics allows us a much better insight into the emergent reality of complex systems like ourselves (dimension 5, backward direction; see e.g. Ismael 2016). Here is where, assuming real time doesn’t flow but events are ordered as relatively earlier or later (that, is McTaggart’s (1908) B-series), or assuming the symmetrical universe and as such McTaggart’s C-series of time that not only doesn’t flow but does not have a direction either (see Farr, this volume), we begin asking the question about the flow: how it is that we feel it (if indeed we do) and where this feeling comes from. The grounding is provided by the concepts of consciousness, personal identity, and self-awareness, in that the passage of time comes with the enduring self. Now, we either experience time as if it were passing (phenomenal passage illusionism, see e.g. Torrengo 2017a, this volume; Prosser 2012, 2013a, 2016) or we only think we do (phenomenal passage eliminativism; see e.g. Hoerl 2014) in that we may only believe that it passes (phenomenal passage cognitivism; see Farr, this volume, Sections 8.4.1 and 8.4.4). Once we address the feelings, we also address emotions (e.g. Piata 2018, this volume; Prosser, this volume). Then we can move to the human concept of time itself (e.g. Jaszczolt, this volume) and to the most important sources of knowledge about this concept, that is its linguistic realizations in different languages, cultures (Caudal, this volume; Everett, this volume; Jaszczolt, this volume), functions, and levels of linguistic analysis (Saussure, this volume; Caudal, this volume; Piata, this volume). Note that the very repetitions of the contributions per theme testify to the methodology adopted here (and in the workshop from which this collection originated) of a multidimensional approach, where the bunches of adopted dimensions differ from chapter to chapter but substantially overlap. (I will risk

4

KASIA M. JASZCZOLT

here a programmatic thought that should such a project be developed into a principal research programme, the insights into time tout court would be fast forthcoming and exciting indeed.)

1.2 A roadmap to the collection In view of these assumptions and methodology, the volume allows for drawing multiple roadmaps between the chapters, whereby links can be established according to the above (and more) dimensions. I will suggest some sample roadmaps here but, to reiterate, the very raison d’être of this compilation is to draw attention to the availability of many novel paths through research on time—real time and human time, here with special emphasis on evidence from temporality in language and discourse. We begin the inquiry with the importance of the ego and in particular with the perspective it imposes on eventualities, that is with egocentric mental states. In Chapter 2 (‘Tense and emotion’), Simon Prosser takes on board the well-rehearsed ‘thank goodness’ argument by which, originally, Arthur Prior attempted to rescue the view that time flows—that is, McTaggart’s A-theory, that the past, present, and future are all real (also known as ‘tensism’, that is, tensed view on reality—not to be confused with the use of the term ‘tense’ in ‘grammatical tense’). The gist of Prior’s defence of tensism is that thinking of a, say, painful dental procedure as being in the past comes with different attitudes and emotions than thinking of it as, say, being on Monday at 10am. But, as Prosser argues, the scenario can also be used in conjunction with the B-theoretic outlook: time as such doesn’t flow but thoughts about the past, present, and future are egocentric—bound to the thinker who assesses the relation between an eventuality and themselves as, say, being in the past. Then, the explanation for the emotional attitude of relief comes from the evolutionary utility of such reactions. In our roadmap, this gives us several possible routes to take: we can probe deeper into the subjectivity and emotional attitudes to time (Piata, Torrengo); continue on the topic of tenseless reality (Torrengo, Farr, Mozersky— again, in variable orders); pursue the first-person perspective on time (Jaszczolt); or go into the expression of past-tense reference in natural language (Saussure, Forbes) and through it to cross-linguistic insights into temporal reference in discourse (Caudal, Everett). Some routes are more discernible than others but pursuing all of them makes sense. Here we choose to follow subjectivity and emotions in the poetics of time. In Chapter 3 (‘An exploration into construals of subjective time in poetry’), Anna Piata sheds new

INTRODUCTION

5

light on the human understanding of time by investigating how temporal experience, and in particular the order and duration of eventualities and the passage of time, are captured in poetry. Having presented some landmarks in the phenomenology of time, she moves to the mental representations of these aspects of temporality, finding that the utility of time for the subjects makes a great difference to conceptualization, as reflected in the utilized (often metaphorical and open-ended) expressions. Such a functionalist analysis is traced back to the idea of embodied cognition in cognitive semantics that locates the representation of states and events, including their temporality and emotional attitudes to it, in the human body—arguably, making the ‘subjective’ less subjective through such interdisciplinary insights. At the next crossroads, again, we are relatively free to choose a path. In addition to continuing straight ahead to the subjectivity of temporal experience but stepping from the level of poetics to metaphysics (e.g. Torrengo, or back to Prosser and then forward), one inviting and suitably challenging path is that of the semantics of temporal expressions—the question of meaning and truth as its explanans. In Chapter 4 (‘The 2D past’), Graeme A. Forbes looks into the metaphysical significance of grammatical tense and aspect, pointing out that focusing on the somewhat neglected role of aspect (that is, situation-internal time) sheds light on the question ‘When am I?’ (meaning asking about one’s objective temporal location), in that, whether the situation is, say, ongoing or completed can make a difference when assessed from the perspective of the relative future. He makes use of the concept of two-dimensional (2D) semantics, also using it to confront stubborn cases where the truth value changes a posteriori, say, when all circumstances of a past eventuality are revealed. We can now choose to move ahead with the question of semantic and metaphysical utility of grammatical categories employed for expressing temporal reference, also staying close to the topic of perspective-taking focused on in Forbes’ paper. In Chapter 5 (‘Counterfactuality as pragmatic inference in perspectival readings of Past Conditional utterances with modal verbs: Evidence from French’), Louis de Saussure looks at the relative scope of modality and tense markers in epistemic modal constructions (‘John might have been surprised.’ ), pointing out that their French counterparts (‘Pierre aurait dû être surpris.’ ) give rise to a counterfactual reading, thereby suggesting that the relative scope is not MODAL>PAST, like for the English constructions, but rather PAST>MODAL. He then inquiries into the possibility of a compositional explanation that would reveal the source of the pastness of the epistemic attitude in French, as contrasted with the present-time location in English. He offers a linguistic-pragmatic solution consisting of a perspectival shift that allows the

6

KASIA M. JASZCZOLT

potentiality of the situation to be seen from a point of view in the past. In other words, the temporality of the modal pertains to that perspectival time rather than to the time of speech. This solution from perspectival shift makes us think of our roadmap not only with reference to the previous stop (Forbes) but also with reference to Prosser’s chapter, suggesting a possibility of yet another promising conceptual route to understanding human time. Counterfactuality, however, need not be conceived of as a clear-cut category. Whether we now pursue a path of cross-linguistic inquiry into representing time, or to counterfactuality itself, or indeed to aspect and even emotions, we find on our path Chapter 6, ‘Avertive/frustrative markers in Australian languages: Blurring the boundaries between aspectuo-temporal and modal meanings’, in which Patrick Caudal gives an insight into the understanding of the flow of time by offering a typological account of a specific category of modal/temporal markers in Australian languages—grouped as non-PamaNyungan—not a genetic group but rather a made-up category, so to speak, negatively defined, spread over most of the territory of mainland Australia, and of particular semantic interest due to their particular morphological complexity and polyfunctionality of categories. The category he looks into is that of grammatical avertive and frustrative markers (often called ‘proximative’), that is foregrounded (grammaticalized) expressions of expectations and disappointments (‘He nearly left’; ‘He tried to leave’; ‘He came in vain’). As such, these refer to negative past events, past modals (such as deontic or volitional), and past imperfective markers (prospective and progressive). The compositional structure of such meanings is revealed through a range of identified patterns. The richness of information they convey, including the case of a mistaken take on an eventuality and an event not leading to desired results, suggests a complex construal of events, associated with a social construal of time where actual and counterfactual events form a continuum. We can now continue with cross-linguistic variation in expressing temporal reference (Everett), or, in a more unorthodox way, ask about the properties of the past and the future, the comprehension of which such linguistic expressions reveal. Our roadmap continues with the latter, moving to the ontological underpinnings of the semantics of sentences with future-time reference. In Chapter 7 (‘On modelling the future’), Joshua Mozersky addresses epistemic, semantic, and ontological indeterminacy of the future, focusing on the dependency of the semantic on the ontological. On an ontological model of branching future, utterances about the future are neither true nor false simpliciter but rather are true on some timelines and false on others—becoming true when time moves on along one of the branches, so to speak, causing

INTRODUCTION

7

attrition of the others at each consecutive ‘present’. Needless to say, this is where the problem starts. He argues that branch attrition does not follow from a model of the determinate past. This leads him to the conclusion that semantics of temporality has to be founded on static time of B-theory, with determinate past, present, and future, in which the passage associated with the way we think and speak is accounted for in a separate way. Here he proposes an explanation from a theory of four-dimensional spacetime—a model using ‘temporally oriented vector field’ that determines an absolute future for each spacetime point. All in all, he concludes, we have to look at time both from the external (as above) and the internal perspective, where the latter comes with retaining thoughts in memory, while ‘sliding’, so to speak, along the timeline of our human, dynamic time: what is future will be present and then past. And as such, determinate or not, it will remain unknown for some time (no pun intended). This interdisciplinary landmark on the roadmap can take us back to the truth-conditional semantics of temporal expressions (Forbes), to other B-theoretic solutions to experienced time flow (Torrengo), or further into the question of directionality of time. We take the latter path. In Chapter 8 (‘Perceiving direction in directionless time’), Matt Farr defends the view that time has no direction (providing in the process a useful overview of what directionality of time can mean on different philosophical accounts) and looks at the compatibility of this C-theoretic picture (McTaggart 1908) with the experience of time flow; although symmetrical time has been well supported by theories of the universe in micro-physics, it does not fit well with our human experience of time. His argument is twofold: showing that the perceived directionality does not pose a problem for such adirectional reality of time, and showing that Ctheory can in fact explain human experience of directional, dynamic reality. Next, he looks at a number of ways in which the experience of time passage has been explained (or explained away) and concludes that one can pick and choose between them (as he tentatively did), leaving timeless reality intact. Moving on from the somewhat underappreciated C-theory to the standard dispute between flow/no flow (that is, A-theory/B-theory), explaining away the passage of time now continues in Chapter 9 (‘Temporal transparency and the flow of time’). Giuliano Torrengo discusses the relation between the temporal properties of experience (such as duration) and temporal properties of the objects of this experience—say, the duration of an event. The relation can go both ways: (i) experience can inherit the temporal structure of an event, or (ii) experience can project temporal structure onto an event. The real question appears when we ask about awareness. For example, according to the

8

KASIA M. JASZCZOLT

latter direction (projection), an unfolding performance of a sonata has a certain duration because our listening to it has this duration. But in order for this to work we would have to be aware of this duration of listening. The first (inheritance) direction suffers from an analogous missing link. Pointing out this missing link adds a new argument to his overall programme of repudiating temporal transparency, or, put simply, denying that we are aware of, say, the duration of a piece of music prior to being aware of the analogous property of our listening to it. All in all, we feel that time is passing but this is not an experience of the passage of time but of something else, ‘phenomenologically modified’ to feel that way. It all boils down to the awareness of the way in which a performance of a sonata is presented to our senses. And this way feels like moving—or somehow ‘flowy’. But it is not really time that we represent. Now, we can either incorporate that landmark on a route back through Prosser and Mozersky, or, in the spirit of the innovative methodology and cross-disciplinary cross-feeding, move on to arguments from linguistic semantics that time is really ‘something else’. (Yes, literally, not just colloquially.) On the latter path, we move on to Chapter 10 (Kasia Jaszczolt, ‘Does human time really flow? Metaindexicality, metarepresentation, and basic concepts’) that brings together the metaphysics of time (timeM ) and linguistic evidence on the conceptualization of time (timeL ), in order to address the question of why human time (or, epistemological time, timeE ) appears to flow. Following emergentist views, indexicality of time is taken to be dependent on the indexicality of the first-person perspective—a construal that I call the ‘metaindexicality’ of time. This explains the apparent flow of timeE and the apparent incompatibility with timeM . A further argument for possible compatibility comes from the fact that timeM as understood in modern physics percolates to common knowledge but only as semi-understood, semi-propositional (Sperber), representational beliefs. This semi-propositional character, paired with the indexicality of the ego-perspective, explains the apparent flow, as well as the apparent variable rate of the flow. Next, I move to the follow-up question in what sense time is dynamic, arguing, with the help of cross-linguistic evidence, that apparently dynamic timeE is essentially static; it is a complex concept that decomposes into conceptual building blocks that are (epistemic) modal in character. On the level of these building blocks, time does not flow; it only flows on the level of their culture- and language-specific combinations. Finally, I propose how this dynamicity can be formalized in contextualist semantics. Here we can revisit philosophical arguments against the time flow from the previous landmarks, or follow a linguistic route and move back to modality

INTRODUCTION

9

as an explanans for temporal reference. Or we can move forward on the path of linguistic diversity to Chapter 11 (‘Underspecified temporal semantics in Pirahã: Compositional transparency and semiotic inference’), in which Daniel Everett analyses the grammar and semantics of temporality in Pirahã, an indigenous language spoken in the Amazon. Pirahã has aspectual suffixes, such as for example telic, continuative or frustrative (see the summary of Chapter 6 above) but not grammatical tense. Neither does it have markers that locate an eventuality unambiguously in the past or in the future. In addition, it lacks perfect tenses and the meaning associated with them. This makes Everett rethink the universality of the Reichenbachian (Reichenbach 1947) logic of tenses that utilizes speech time, event time, and reference time as correlates, in that the correlate of reference time appears to be absent. Next, he suggests that culture permeates semantics and affects it to an extent greater than formal features of the language system do. He talks about ‘degrees of fit’ between morphosyntactic structures, meanings, and cultures, which calls for adopting a more practical, common-sense, or inference-based idea of compositionality, whereby meanings are compositional not on the level of syntactic structures but on the level of culture-infused interpretations. (In our roadmap, a path back to Chapter 10 and contextualist Default Semantics employed there would be pertinent.) Historically, Everett finds conceptual roots of such contextualinference-driven compositionality in Peirce’s semeiotics. Languages, he claims, can be classified on the scale of their compositional transparency. But conceptual transparency is not a matter of variation: we infer meaning because our species developed the ability to infer from signs—from all kinds of signs that present themselves as sources of information. (Note that another possible route back to Prosser’s evolutionary explanation of emotions becomes evident here.) All this pertains to temporal location, where different cultures come foregrounding different temporal concepts. Cultural and linguistic diversity lead to the question as to whether there indeed is the human time—something we discern as a ‘human way’ of thinking about time. Perhaps the diversity and the lack of direct correlates between languages point towards such a degree of relativity (or at least correlation) that it overshadows what could usefully be thought of as universal? Across the landmarks of this roadmap, the human time stands for inner experience (of time or something else), with its semblance of pastness, ‘nowness’, and futurity, as well as subjectively assessed duration. Philosophers still inquire into what it is that we actually experience and how it translates into beliefs about temporal location and duration, offering a multitude of theories (some of them helpfully reviewed in what follows). But it is also linguistic diversity that allows us to get

10

KASIA M. JASZCZOLT

a glimpse of what really matters about time, and as such an insight into its properties. Since (note: ‘since’ rather than ‘if ’ for the purpose of this volume) time passage is the human take on static real time that only emerges on the level on which it is of some use, it is of paramount importance to search for what this ‘take’ really is and what it depends on. And search we do, in different disciplines, just not often enough in a bold cross-disciplinary and inter-disciplinary way. To reiterate, many other roadmaps can be drawn, utilizing the choice of dimensions [1]–[10] from Section 1.1 and adding others. It is hoped that this ‘informed combinatorial’, so to speak, approach will help with discovering such novel, interdisciplinary (or at least cross-disciplinary if recalibration is more of an issue) slants on the Big Questions to do with human time. So, follow my roadmap, or draw your own, don’t skip philosophers or linguists, read them both, and enjoy the experience—the time spent on it will contract if you do!

2 Tense and emotion Simon Prosser

2.1 Introduction If the B-theory of time is true, then why do we say ‘thank goodness’, and feel relief, when a traumatic event is over, and in the past, but experience feelings of apprehension when it is in the future? I shall suggest at least the outline of an answer. In doing so, I shall describe the special kinds of egocentric mental states that one must be in, in order for certain kinds of emotional reactions to an event to be rational, or appropriate. This will also add to a certain picture of the kinds of egocentric mental states that are ‘essential’ in order for one to be able to interact with the world around one, and which characterize the firstperson perspective. Arthur Prior (1959), who originally put forward the ‘thank goodness’ challenge, held that only a ‘tensed’ metaphysics, in which there are mind-independent, objective facts about which times and events are past, present, or future, could explain how such a difference in attitudes could be justified. For many years this was regarded as an important challenge to the ‘tenseless’, or ‘B-theory’ metaphysics, according to which, although there is a real time series, and hence an objective ordering of times and events, there are no times on the series that are objectively past, present, or future, and there is no such thing as the passage of time. For the B-theorist, time is in many ways similar to space; the world is extended in time, much as it is extended in space. Just as objects in different locations in space coexist on an equal footing, so, according to the B-theory, do events located at different times. Changes take place in time, but consist only in there being different states of affairs located at different times.¹ According to Prior’s preferred metaphysics, usually known these days as the ‘A-theory’ or ‘tensed’ theory, and often regarded as the ‘common sense’ view of

¹ For a classic defence of the B-theory, and discussion of A-theoretic alternatives, see Mellor (1981, 1998). See also Farr (this volume) and Mozersky (this volume).

Simon Prosser, Tense and emotion. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Simon Prosser (2023). DOI: 10.1093/oso/9780192896445.003.0002

12

SIMON PROSSER

time, the time series itself changes as time passes. On this view, when I dread a forthcoming dental appointment, I have reason to do so because it is simply a fact about the world—a fact about reality as a whole—that the appointment is in the future (the future—not just my future). And, similarly, after the appointment, I have reason to feel relief and say ‘thank goodness that’s over’ because time has passed and reality has now changed, and the dental appointment has gone from being in the future to being in the past. It is this change in reality itself that justifies the change in my attitude to the event. Different realities justify different attitudes. The challenge for the B-theorist, by contrast, is to explain why it can be appropriate to take different attitudes at different times to the very same event, given that, according to the B-theory, nothing about the time series itself ever changes. One would, it seems, be taking conflicting attitudes towards the very same reality. As a number of B-theorists have noted, on closer inspection, it is not entirely clear that the A-theorist is really better off, for it is not immediately clear why a traumatic event being in the objective past or future should explain the appropriateness of taking different attitudes to that event, let alone why one should take the specific attitudes of dread followed by relief.² Nevertheless, whatever problems may face the A-theory, the challenge to the B-theory remains; an explanation is needed before the theory can be regarded as fully satisfactory. More recent discussions by B-theorists have assumed that the phenomenon to which Prior drew attention was just an example of the more general phenomenon of the essential indexical (Perry 1979), according to which the thoughts that we express using indexical terms like ‘I’, ‘here’, and ‘now’ play essential psychological roles in bringing about actions (for these purposes, an indexical term is any linguistic expression whose semantic value varies systematically with context). For example, I may believe all day that SP is in danger, but it is only when I come to believe that I am in danger (where my token of ‘I’ refers to SP) that I take evasive action. Or, I may believe that St Andrews would be a good place to dig for gold, but it is only when I come to believe that here would be a good place to dig for gold that I start digging. Finally, to borrow a temporal example from Perry, I may believe all day that the meeting is at 3pm, but it is only when I come to think of 3pm as now that I rush to the meeting. On the face of it, the thought that the meeting starts at 3pm, and the thought that the meeting starts now (where that thought occurs at 3pm), both represent the meeting as happening at the same time; but it is only when ² See for example Garrett (1988).

TENSE AND EMOTION

13

the time of the meeting is thought of in a special way, as now, that the thought leads to action. The similarity, in this respect, between ‘I’, ‘here’, and ‘now’ has suggested to many B-theorists that Prior’s example tells us nothing specifically about time. I agree, but there is more to say. More recently, however, scepticism has been expressed by some philosophers about the doctrine of the essential indexical (in particular see Millikan 1990; Cappelen and Dever 2013; Magidor 2015). According to them, the fact that the change in my behaviour occurs only when I go from thinking of the time of the meeting as ‘3pm’ to thinking of it as ‘now’ shows only that this is a case of the familiar phenomenon of referential opacity. A similar change in my behaviour may be brought about by a change from thinking of a person as ‘Norma Jeane Mortenson’ to thinking of them as ‘Marilyn Monroe’. Much of the focus of this discussion has been on actions, though Cappelen and Dever give broader arguments to the effect that there is nothing special or fundamental about the first-person perspective (which is commonly associated with the use of terms like ‘I’, ‘here’, and ‘now’). I disagree with this broader claim. I have argued in previous work that although it is true that indexicality per se has no special role, there are associated egocentric mental states that do indeed have an essential role in action (Prosser 2015). In this chapter, I shall focus on emotions rather than actions, though I shall note some obvious connections between them. Emotions have received very little attention in the debate over essential indexicals. By using Prior’s ‘thank goodness’ case as my central example, I shall explain why certain emotional reactions to an event are appropriate only when the event is thought of as being past, present, or future, where these terms are understood in a manner compatible with the B-theory. Thinking in terms of past, present, or future is part of the broader phenomenon of adopting the first-person perspective, and adopting this perspective is essential to a variety of emotional reactions; these reactions are not appropriate when the same states of affairs are thought of tenselessly or, as we might equivalently say, from the third-person point of view. My aim is therefore to do two things at once: firstly, I shall add to the proper understanding of Prior’s ‘thank goodness’ puzzle from a B-theoretic point of view, and secondly, I shall provide a new kind of response to scepticism about the essential psychological role of the first-person perspective.

2.2 Indexical pronouns and egocentric predicates I should start by saying something about what I take to be the relation between the indexical pronoun ‘now’ and predicates like ‘past’, ‘present’, and ‘future’.

14

SIMON PROSSER

In my view, the philosophical discussion of the first-person perspective has focused rather too much on words like ‘I’, ‘here’, and ‘now’, and has not paid sufficient attention to expressions like ‘near’, ‘far’, and ‘to the left’, as well as ‘past’, ‘present’, and ‘future’, all of which I shall refer to as egocentric predicative terms, or egocentric predicates for short. It is easy to find examples showing that thoughts expressible using egocentric predicates play the same ‘essential’ role for action as those expressible using ‘I’, ‘here’, or ‘now’ (and, as we shall see, the same applies to certain emotional reactions). Perhaps it is sometimes assumed that when one thinks that there is danger nearby, what one really thinks is that there is danger near to me, and that it is the unstated indexical ‘me’ that is essential for action. But one cannot make the same claim in the temporal case; an event that is in the past cannot literally be understood just as standing in some relation to me, for I may have also existed before the event, and therefore stand (or stood) in many different temporal relations to it. At any rate, the B-theorist must say this, since the B-theorist must regard a persisting person as a temporally extended entity.³ I shall suggest, in what follows, that there is in fact a crucial role for the relevant temporal ‘stage’ of a person; but not as the referent of ‘I’. The focus on indexical pronouns (‘I’, ‘here’, ‘now’) has, I think, created an impression that the ‘essential indexical’ phenomenon is a matter of thinking of a person, place, or time, in a special way, perhaps under some kind of special indexical mode of presentation (whatever that would mean). I think this is a mistake, and has led to much confusion. In my view, the explanation for the apparently special role of these indexical terms lies not in the indexicality itself, but in the fact that a use of any such term implies that the speaker believes, or perhaps presupposes, that they stand in a certain relation to the reference. Under normal circumstances, a speaker, S, who uses ‘here’, when speaking about location l, believes that l is the place in which the token of ‘here’ was uttered. But, arguably, S also takes L to be the place at which S is located (albeit perhaps only implicitly, as explained below). Similarly, when S uses ‘now’ when speaking about a time, t, S believes that t is the time at which the token of ‘now’ was uttered. But, arguably, S also believes, at the time of utterance, that t is present. This seems hard to deny—no one could coherently

³ I do think, in fact, that we often think of ourselves as though we were wholly located at one time, and that we ‘move’ through time, as time ‘passes’. This probably explains the degree to which it can seem to make sense to say that an event was in my future, but is now in my past. But this does not make literal sense, given that one exists at more than one time by virtue of persisting.

TENSE AND EMOTION

15

refer to a time t as ‘now’ while believing that t was past or future. I think that something similar is true for ‘I’, though I shall not go into the details of this here.⁴ Egocentric predicates are used precisely to describe, or think about, relations in which one stands to places, times, or persons. So, I am suggesting that whenever a speaker makes a sincere assertion using an egocentric indexical such as ‘I’, ‘here’, or ‘now’, the speaker also has a corresponding belief involving an egocentric predicate. In what follows I shall explain, for the temporal case, exactly why temporal egocentric predicates, which I take to include ‘past’, ‘present’, and ‘future’, have a special role, connected with the first-person perspective, that is essential for emotional reactions to events (and also, arguably, for actions).⁵ If this explanation is correct, then no further explanation in terms of special modes of presentation associated with indexical pronouns is needed.

2.3 Truth conditions for thoughts about the past When one thinks ‘thank goodness that’s over!’, what, according to the B-theorist, does one thank goodness for? That is to say, what are the truth conditions for ‘x is over’? Although there are some subtle differences between ‘x is over’ and ‘x is past’, I shall gloss over these, for I do not think they affect the key claims that I wish to make.⁶ Now, traditional versions of the B-theory divide into two camps concerning the truth conditions for ‘x is past’. The date theory holds that words like ‘past’, ‘present’, and ‘future’, when applied to a time t, concern the temporal relation between t and the time of utterance (which should be identical to t for an utterance of ‘present’, but not for ‘past’ or ‘future’). The token-reflexive theory holds instead that the relevant relation is between t and the uttered token of ‘past’, ‘present’, or ‘future’.⁷ ⁴ The different types of belief associated with, and indicated by, the use of an indexical term are referred to as indicated linguistic beliefs and indicated egocentric beliefs in Prosser (2015: 214–215), where more detail is given. The corresponding set of relations in the personal case is a far less obvious one, and we have no reason to have words to describe such relations. See Prosser (2015: 216–217) for details of these -relations. Very briefly, -relations are certain relations in which one stands to oneself but not to other people—most plausibly, relations that involve a combination of monitoring and control. ⁵ In Prosser (2015) I showed, via a regress argument, that mental states involving egocentric predicates are essential for actions. But, despite the title of the article (‘Why Are Indexicals Essential?’), there is a sense in which the regress argument does not fully explain why they are essential. In what follows I try to fill this gap. ⁶ For discussion of the differences, however, see Hoerl (2013a). ⁷ See Dyke (2002) for defence of the token-reflexive theory relative to the date theory. See Le Poidevin (1998) for a much more nuanced discussion than I have given of the different versions of both A-theoretic and B-theoretic truth conditions.

16

SIMON PROSSER

If one of those theories provided the whole story, then one would thank goodness for the relation between the traumatic event and either a date or an utterance token. But, as Prior pointed out, neither is plausible: One says, e.g. ‘Thank goodness that’s over!’, and not only is this, when said, quite clear without any date appended, but it says something which it is impossible that any use of a tenseless copula with a date should convey. It certainly doesn’t mean the same as, e.g. ‘Thank goodness the date of the conclusion of that thing is Friday, June 15, 1954’, even if it be said then. (Nor, for that matter, does it mean ‘Thank goodness the conclusion of that thing is contemporaneous with this utterance’. Why should anyone thank goodness for that?) (Prior 1959: 17)

Here is a way to see Prior’s point: according to the B-theory, the fact that the traumatic event stands in a certain temporal relation to the utterance, or to the time of utterance, is an eternal fact. It does not change, and is just as much a fact before the traumatic event as it is afterwards. So it is unclear why someone should have any reason to feel relief about it, and if there were any such reason, then one would have just as good a reason to feel relief before the event as after it. Yet this is not the case. Of course, before the event, one might not be aware that such an utterance ever occurs. But if, before the traumatic event, an oracle were to tell one, with perfect reliability, that an utterance of ‘thank goodness that’s over’ would occur after the event, it is hard to see how one’s feelings would be rendered any different. One would look upon the forthcoming event with just the same feeling of trepidation. Moreover, if there is a reason to feel relief at a given time, then that reason exists regardless of whether any utterance occurs at all. Neither would it help to think of the utterance demonstratively, as ‘that utterance’. Suppose that one were equipped with a crystal ball that enabled one to witness future events. The events would not look as though they were in the future—it is hard to imagine how such a thing could look. As D. H. Mellor (1998: 16) once pointed out, every perceived event appears to us to be happening in the present. Nevertheless, if one could witness the utterance of ‘thank goodness that’s over’ through one’s crystal ball, then one could refer to it demonstratively, and one might come to believe that ‘that utterance occurs after the conclusion of the traumatic event’. Somewhat less fancifully, one might see a film of a past utterance, perhaps running long enough to see both the traumatic event and the utterance. Again, one could refer to the utterance

TENSE AND EMOTION

17

demonstratively. In either case, Prior’s point remains unchanged; one would still have no reason to feel relief. A hint at what is wrong can be found by reflecting on the fact that the proposed truth conditions can be thought about equally easily by anyone, even though not everyone has equal reason to feel relief. If I have a traumatic dental appointment tomorrow, then there is reason for me to feel trepidation now, and there will be reason for me to feel relief afterwards. But, unless you have a truly exceptional concern for my wellbeing, your knowledge of my dental appointment gives you no reason for you to feel as I do. Yet you are just as capable as I am of understanding the temporal relations between the relevant events and the relevant times or utterances. What we need is an explanation of how there can be a reason for me to feel a certain way, which is not also a reason for you to feel that way.

2.4 Half of the answer: The evolutionary story Here is an obvious reason for me, but not you, to dread the forthcoming traumatic event, and feel relief when it is over: the event affects me, but not you. Our emotions evolved, presumably, to influence our behaviour in such a way as to raise our chances of survival. Feelings of dread or fear, for example, can influence one to act in such a way as to prevent a harmful event from occurring at all. If I foresee a possible future in which I am attacked by the predator that I see before me, and I fear this, then I move away, to avoid the attack. It must be me, not you, who feels the fear, in order for me, rather than you, to be motivated to act in the relevant way.⁸ This insight can help us with the temporal case. According to the stage theory of persistence (Sider 1996, 2001; Hawley 2001), a persisting person consists of a series of temporal ‘stages’—momentary time-slices, each of which we could think of as being a kind of very short-lived person (the theory also applies to persisting physical objects in much the same way). If we think of the personstages before and after the traumatic event as distinct persons, then Prior’s puzzle comes down to this: how can there be a reason for the earlier personstage, but not the later person-stage, to fear the traumatic event? And how can there be a reason for the later person-stage, but not the earlier person-stage, to feel relief ?

⁸ There are of course cases in which it is adaptive for one’s emotions to motivate actions that protect someone else. Nevertheless, it is the one who acts who must feel the emotion.

18

SIMON PROSSER

When the question is posed in this way, an explanation, corresponding to the one described above for the interpersonal case, suggests itself. There are different reasons for emotions for the earlier and later person-stages, because they stand in different relations to the traumatic event. In particular, the earlier person stage is causally upstream from the possible traumatic event, and therefore can potentially influence or prevent it, whereas the later person-stage cannot.⁹ James Maclaurin and Heather Dyke (2002) have suggested a solution very much along these lines, albeit not explicitly described in terms of the stage theory. They suggest that it is adaptive, in evolutionary terms, for a creature to have different emotional reactions to an actual or possible event depending on the temporal relation between the creature and the event at the time of the emotion. They give further details, for example suggesting that there is a similar evolutionary explanation of our differing emotional attitudes to events in the near future and the more distant future. The rough idea is that a possible event located in the distant future is one that is hard to influence now, and there will be plenty of time to influence it later. So, even if the event will be traumatic if it occurs, there is reason to feel only a relatively mild trepidation now. Stronger emotions, at this point, would just waste time and energy. When the event is to occur soon, however, then there is reason to act straight away if it is to be prevented, and that is why it is adaptive to feel stronger emotions about an event in the nearer future.¹⁰ No doubt there is room for debate over the details, but I think the broad kind of account that Maclaurin and Dyke suggest is prima facie plausible, and I shall assume that some account of this kind is correct. But this still leaves us with two questions: what are the truth conditions for thoughts and utterances about the past, present, and future, and why is it that only the relevant personstage is disposed to feel the relevant emotions when thinking thoughts with these truth conditions? The first question is easy to answer; the second is more subtle. I shall deal with the first in the remainder of this section, before moving to the second in the next section. ⁹ This, I take it, is the objective situation that grounds the adaptiveness of our past and future attitudes. But, as Bordini and Torrengo (2022) point out, one’s psychological grasp of what it is for an event to be in the future may be essentially bound up with the notion that a time that is future is one that is going to be present, at which point one will find oneself existing simultaneously with it. As Bordini and Torrengo suggest, this may relate to the psychological sense (albeit not the objective reality) of time passing. See also Prosser (2016, section 7.1), on the connection between an event being thought of as future, and it being thought of as ‘approaching’ (or of oneself as approaching the event). ¹⁰ See Maclaurin and Dyke (2002), and also Suhler and Callender (2012), for further details of arguments of this kind. See also Sullivan 2018, however, for dissent on the rationality of ‘time bias’ in general. I shall assume in what follows that evolutionary considerations do at least explain why we have temporally biased attitudes, even if there are arguments that it is not in one’s self-interest to do so (which is not to say that I accept all of Sullivan’s arguments).

TENSE AND EMOTION

19

The first question can, I think, be answered just by considering the states of affairs that have to occur in order for it to be appropriate for a given personstage to experience a given emotion. Consider a person, S, who has a person stage, St , existing at time t, and an event, e, occurring at some specific time, to which it is appropriate for St to be disposed to produce emotional reaction E. According to the kind of account suggested by Maclaurin and Dyke, what ultimately makes it appropriate for St to produce reaction E is that St stands in some specific temporal relation, R, to the event e. For example, if R is the ‘past’ relation, and e is a traumatic event, then St stands in the relation R to e that makes relief appropriate. So the state of affairs for which St appropriately thinks ‘thank goodness e is over’, and feels relief, is the state of affairs that e is earlier than St . Consequently this relational state of affairs, which we can write as ‘R(St , e)’, is the truth condition for the thought that produces the emotional reaction in St . I am not necessarily suggesting that R(St , e) is the truth condition for St ’s use of the word ‘past’, however. When I utter the word ‘past’, as part of a complete sentence, and you understand me, perhaps the understanding of the utterance that we have in common is captured by the token-reflexive truth conditions (i.e. that ‘x is past’ is true if, and only if, the token of ‘past’ is produced at a later time than x). I shall remain neutral on this, for it is possible that the personstage that produces an utterance will serve to capture the truth conditions just as well as the token utterance thus produced.¹¹ If we are to correctly understand the significance of thoughts about the past, present, and future, however, then I do think it is necessary to accept that the states of affairs being thought about, at least in basic cases in which one simply thinks that a particular event is in the past, present, or future, concern the relation between the event and the person-stage that is entertaining the thought. So far, so good, but we have not yet fully solved the problem raised by Prior’s challenge. The state of affairs R(St , e) is, after all, just as eternal and unchanging as those concerning the relations between e and the time of utterance, or the token utterance. Why should anyone thank goodness for R(St , e)? There is no reason for anyone other than St to thank goodness that St is located later in time than e. Even S’s earlier temporal stages have no reason to thank goodness for that. So why should the situation be any different for St ? Normally, the identity ¹¹ This is a technical issue about which, I admit, more needs to be said than I can say at present. It relates to the question of what it takes for participants in a conversation to presuppose a common ‘now’. On the face of it, insofar as a token utterance must be produced by a speaker at a time, and hence by a person-stage, the token and the person-stage that produced it should give rise to extensionally equivalent truth conditions. There might, however, be trickier cases, such as the much-discussed answering machine messages saying things like ‘I am not here now’ (Predelli 1998).

20

SIMON PROSSER

of the thinker of a thought does not make any difference to the psychological importance of the thought. I have suggested, above, that the problem cannot be happily solved by suggesting that, in addition to thinking that R(St , e), St also thinks ‘I am St ’. I am not identical with my stages—I exist, or existed, at many different times. It should also be added that it seems implausible that thoughts of this level of sophistication should be necessary for the feeling of relief to be appropriate. There could be creatures who did not have ‘I’ thoughts (that is, thoughts involving an ‘I’ concept), let alone thoughts about personstages, but were nevertheless capable of understanding that a traumatic event was in the past, and felt relief as a result?¹²

2.5 Egocentric mental states The puzzle we face is to explain how it is possible that, for some state of affairs, A, when one specific person-stage, St , thinks about A, St has reason to react with certain emotions, yet no one else who thinks about A has any such reason. In order for this to be possible, there must be some way of thinking of A that produces the relevant emotional reaction, and only St can think of A in that way. I have suggested that the state of affairs in question is R(St , e), i.e. personstage St stands in relation R to event e, where R is a temporal relation such as earlier or later (perhaps to some specific degree).¹³ The answer is found in a phenomenon that I have elsewhere called firstperson redundancy (Prosser 2015). This is an epistemological phenomenon, though it is closely related to the more frequently discussed fact that a subject, S, can think about S’s relations to places and times monadically, that is to say, by using monadic predicates such as near, far, to the left, and also (according to the proposals above) past, present, and future, when dealing with what is

¹² Similar points have been made by many others. See for example Recanati’s (2007) distinction between the implicit and explicit de se, and Millikan’s (1990) discussion of people who do not use indexicals. There is another reason for my reluctance to accept a reduction of all egocentricity to the occurrence of thoughts involving ‘I’. It is that this would still leave us with the problem of explaining what is special about ‘I’ thoughts. I don’t know of any satisfactory way to explain this; yet, if the account that I propose below is correct, then everything has already been explained. ¹³ For the purposes of this chapter it will do no harm to think of R as simply a temporal relation. But in fact I think the relation that is relevant to the subject’s psychology is a more subtle one. Consider two creatures, one of whom has internal processes and outward behaviour that all occur twice as quickly as the other. I think that to the former creature, there is a sense in which the same events would seem twice as far into the past or future as to the other. This suggests a sense in which the degree of subjective pastness or futurity of an event is a matter of how it relates to the thinker’s capacities for actions, including mental actions, and that subjective pastness and futurity are causal–functional relations. See Prosser (2016: chapter 4) for a full development of this line of thought.

TENSE AND EMOTION

21

in fact a two-place relation.¹⁴ No claim about contents being monadic plays any part in what follows, however. First-person redundancy is an epistemic phenomenon, and is compatible with the relevant predicates being two-place. First-person redundancy occurs in certain cases in which a subject, S, judges that a relation obtains between S and something else (such as a place or time), and trades on the fact that S is one of the relata in the relation to take a kind of epistemic shortcut. Normally, in order to judge that an n-place predicate applies to some state of affairs, one needs n pieces of information. By a ‘piece of information’, I mean a parameter of some kind—typically something that can be represented by a number, or an order series of numbers (as one might use a triplet of numbers to represent spatial location or direction). In many cases, this requires attending to n objects in order to acquire information from them. So, for example, to judge that A is taller than B, one normally needs two parameters, representing the heights of A and B. Sometimes, however, when S is one of the relata in the relation, S is able to judge that the relation holds by acquiring only one parameter, or attending to only one object. Consider, for example, some of the multiple ways in which it is possible to judge the spatial distance between oneself and an object. In order for someone else to judge that S is at distance d from object O, they would need two parameters, corresponding to the locations of S and O. But suppose, for example, that S were equipped with a lens with adjustable focus, such as a camera lens with a focus scale. By twisting the focussing ring until the object was in focus, S could read off the distance to the object just by looking at the resulting angle of the focussing ring (as shown by the scale). S would therefore be able to judge the distance from S to O by obtaining just that single parameter. As a matter of fact, this loosely corresponds to the phenomenon of visual accommodation, one of several ways in which the human visual system can judge depth. One’s brain can—within a limited range of a few metres, and with limited accuracy—judge

¹⁴ The fact that we can use n-place predicates in dealing with use n+1-place relations was first discussed at length by John Perry (1986). In such cases, there are what Perry calls unarticulated constituents in the content of speech (and perhaps thought). Many others have made claims about the apparently monadic form of some of our mental representations, most commonly in relation to spatial experience, though sometimes for other cases (see for example John Campbell’s (1994) notion of causal indexicality). First-person redundancy is also closely related to what François Recanati (2007) has called the implicit de se. It should be stressed, however, that first-person redundancy is an epistemic property, not a semantic-syntactic one. It also should not be conflated with the notion of identification-freedom, which Recanati (2007) associates with the implicit de se, and to which he appeals in his explanation of immunity to error through misidentification (IEM). A judgement is identification-free if the epistemic grounding of the judgement does not rely upon the truth of an identity statement (such as ‘I am NN’), which may or may not be explicitly entertained. I do not agree that the implicit de se itself explains IEM in quite the way suggested in Recanati (2007), though I do agree that there are some important connections.

22

SIMON PROSSER

the distance to a perceived object just by registering the amount by which the eyes must be focussed to bring the object into focus.¹⁵ There are several other phenomena, such as stereopsis and convergence, as well as contextual information about perceived textures and so on, that combine to provide information to the brain about distances of perceived objects. As a result, when one looks at an object one can judge how far away it is— the distance between oneself and the object—without needing to compare the location of the object with one’s own location. One does not even need to perceive oneself to make such a judgement. Whereas, with the camera lens, one would have to make the judgement that one was located in the same place as the lens—a judgement that someone else could have made—in the case where one makes the judgement using one’s own visual system, no such identity step is needed (one is the measuring apparatus, so to speak). This is the phenomenon of first-person redundancy—so called, because attention to oneself, or even any representation of oneself, would be epistemically redundant in the epistemic grounds for the judgement (the judgement would have just the same justification without it).¹⁶ Our judgements about the past and future can also exhibit first-person redundancy. Consider the ordinary ways in which one can judge the time interval between the births of Plato and Aristotle. In order to discover this, one would need to know two things: the different dates on which Plato and Aristotle were born. But when one is one of the relata in a temporal relation, one can trade on this fact to take the kind of shortcut described above. Suppose one wished to know the size of the time interval between some past event and the temporal stage of oneself that is making the judgement. A loose analogy with the camera lens would be a stopwatch—one could start the stopwatch at the time of the event, and then learn the size of the time interval between the event and the person-stage making the judgement just by reading the figure from the stopwatch. As with the camera lens, this would require making the judgement that the perceived state of the stopwatch was simultaneous with the relevant person-stage, and someone else could make that judgement. But there are ways of making the judgement that involve no such intermediate step. One can, for example, think about a remembered event, and make a direct judgement

¹⁵ For details of the relevant research, see Heinemann, Tulving, and Nachmias (1959); Leibowitz and Moore (1966); Ku¨nnapas (1968); Wallach and Floor (1971). ¹⁶ While I have coined the phrase ‘first-person redundancy’ to emphasize the epistemological, rather than semantic, phenomenon, I am certainly not the first person to have noticed the fact that one can make such judgements without perceiving or attending to oneself. But the full epistemological significance of this phenomenon does seem to have generally been missed.

TENSE AND EMOTION

23

about how long ago the event occurred, without having to compare temporal locations. One cannot always do this. Sometimes, when one remembers an event, especially if the event is in the distant past, one can only figure out how long ago it occurred by thinking about what else was happening at around the same time, figuring out the date, and so on. But over shorter timescales one can make the judgement more directly. One can, for example, hear a sound, and for a certain period of time one can consult one’s memory of the sound, and one simply has a sense of how long ago it occurred. One does not do this by comparing the time at which the event occurred with the time at which the judging person-stage exists. One need only pay attention to the felt sense of how long ago the event occurred.¹⁷ Something comparable can perhaps be said about some of our judgements about the future. One cannot perceive the future (barring crystal balls and the like, which, as noted above, would not present the future as future anyway). But one can anticipate certain future events, including the unfolding of one’s own actions. When one does so, one has a sense of their degree of futurity. The ‘feeling’ of having dental surgery starting in ten minutes is quite different from the feeling of it starting in ten seconds. But one need not represent a future event, and its degree of futurity, by explicit representing both the event and one’s own current temporal stage. One need only think about the event, and its felt degree of futurity. In all cases, I shall speak of such judgements, where one less parameter is needed because the subject is trading on being one of the relata in a relational state of affairs, as being made first-person redundantly. Only the relevant person-stage—the one who stands in the relation in question—can judge, firstperson redundantly, that the relation obtains. So, although anyone might be in a position to judge that R(St , e), it is only St who can do so first-person redundantly. Only St can stand in that particular epistemic relation to the state of affairs R(St , e). So, with regard to the temporal case of interest, there is a way for St to judge that a certain event is past, present, or future, that is only available to St . Ways of judging are associated with ways of thinking. Consider the standard, Frege-inspired way to distinguish different thoughts for a given subject at a given time. Thoughts T1 and T2 are distinct, for a subject S, if, and only if, ¹⁷ Just as with the case of visual judgements of distance from oneself, there has been much empirical research on the mechanisms that allow one to make these temporal judgements. And, as with vision, it seems likely that there may be a variety of different mechanisms at work in different contexts. See Grondin (2010, especially pp. 567–570), for a useful brief survey, and see Grondin (2020: 1–71) for a more in-depth introduction.

24

SIMON PROSSER

it would be rational for S to accept T1 while denying, or remaining agnostic about, T2 (or vice versa). It will be helpful to revisit the reason for this. We can think of a subject as being rational if, and only if, the subject’s view of the world is coherent. A rational subject should not hold the world to be a way that it is logically impossible for a world to be. So, to use a standard example, no world could be one in which Hesperus is bright and Hesperus is not bright (‘Hesperus’ is an ancient name for the planet Venus). So, a rational subject should not hold that this is the case. Given that ‘Phosphorus’ is also a name for the planet Venus, there is no logically possible world in which Hesperus is bright and Phosphorus is not bright. It does not, however, follow that it would be irrational for a subject to hold that Hesperus is bright and Phosphorus is not bright. This is because the epistemic conditions under which the subject would judge that Hesperus is bright may differ from the epistemic conditions under which the subject would judge that Phosphorus is bright, and there may be possible worlds in which those epistemic conditions would lead to judgements about different objects. Suppose—again following the standard example—that the subject judges that something is ‘Hesperus’ when it is the brightest object in the evening sky (the ‘evening star’), and judges that something is ‘Phosphorus’ when it is the brightest object in the morning sky (the ‘morning star’). In the actual world, the brightest object in the evening sky is the brightest object in the morning sky—it is just the same object seen at a different time. But there are possible worlds in which different objects would be seen in the evening and morning skies. So, the subject who believes that Hesperus is bright and Phosphorus is not bright is implicitly taking the actual world to be of this latter kind. It is a way that the world could have been. So, the subject is mistaken, but not irrational. This is what we normally have in mind when we say that ‘Hesperus is bright’ and ‘Phosphorus is bright’ express different thoughts. So, wherever there are different ways for a subject to judge that a given state of affairs obtains, there may be different corresponding thoughts. Consider, now, what this tells us about cases of first-person redundancy. Where there is first-person redundancy, there is a type of epistemic access to a specific state of affairs that is only available to one specific person-stage (though other personstages may have the same type of access to states of affairs involving a different person-stage, but the same relation). This differs from many other types of epistemic access to states of affairs. If I can judge that Hesperus is bright by virtue of judging that the evening star is bright, for example, then so can you. The latter kind of case involves epistemic access to a state of affairs from the third-person perspective—a perspective that can be shared by many different person-stages. But where there is first-person redundancy, the subject’s

TENSE AND EMOTION

25

epistemic access to the state of affairs is from the first-person perspective, and no one else can share that same perspective on the very same state of affairs. It follows that, at time t, a subject, S, may have epistemic access to the relational state of affairs R(St , e) from both the first-person perspective and also from the third-person perspective. Given the difference in epistemic access, it should be conceivable, from the subject’s point of view, that the state of affairs thought about in one way obtains, but the state of affairs thought about in the other way does not. This suggests a difference in thought. So there is a particular species of ‘egocentric’ thought about states of affairs of the form R(St , e) that is available only to St .

2.6 The inferential isolation of the egocentric So far, we have seen that when someone thinks of an event, e, as past, present, or future, they can do so egocentrically, and their thought is thus distinct from any thought that they could entertain non-egocentrically, such as the thought that e is earlier than St . It is important to notice that this difference in thoughts is not like the difference between ordinary third-person thoughts, such as the difference between thoughts about Hesperus and Phosphorus. If this were the only difference between tensed and untensed thoughts then there would be nothing to say against Cappelen and Dever’s (2013) claim that such differences amount to nothing more than standard cases of substitutivity failure, showing nothing special about the first-person perspective, or about tense. The difference between egocentric and non-egocentric ways of thinking runs deeper. Egocentric thoughts are inferentially isolated from nonegocentric thoughts. Suppose S believes that Hesperus is bright, but does not infer from this that Phosphorus is bright, even though, unbeknown to S, Hesperus is Phosphorus. If S judges that something is Hesperus just when it is encountered as the evening star, and judges that something is Phosphorus just when it is encountered as the morning star, this leaves open the possibility that Hesperus is not Phosphorus, and so the inference does not go through. But suppose S’s knowledge were supplemented by further pieces of non-egocentric knowledge. Eventually, the inferential gap would be closed, as it would become apparent that the brightest object seen in the evening sky was also visible, and was the brightest object, in the morning sky. So this further third-person knowledge would eventually put S in a position to know that the conditions of application for ‘Phosphorus’ were satisfied by the bright object in the evening sky. Arguably, there are no epistemic gaps between non-egocentric thoughts that cannot be closed in this way.

26

SIMON PROSSER

But it is never possible to infer an egocentric thought from non-egocentric thoughts. One person’s inference is the same as another’s—the identity of the person making the inference makes no difference to what can be inferred from what. But only St can think about R(St , e) egocentrically, for only St can be in a position to make the judgement first-person redundantly. For any true nonegocentric thought, T, adding non-egocentric knowledge can always put one in an epistemic condition in which it would be appropriate for one to judge that T was true. But adding non-egocentric knowledge cannot, in itself, bring about the conditions in which it would be appropriate to judge an egocentric thought to be true.¹⁸ So egocentric thoughts form a special class, inferentially isolated from non-egocentric thoughts. This alone already shows that there is something special about the first-person perspective.

2.7 Self-location Let us take stock. I suggested that in order for the B-theorist to address Prior’s challenge, it must be explained how St ’s thinking that the traumatic event is over, or past, can put St in a position where it is appropriate for St to feel relief, yet no other person, and no other person-stage of S, can be in that same situation, and have the same reason. We have now seen that, as a consequence of the phenomenon of first-person redundancy, although others can think of the state of affairs R(St , e), there is a particular way of thinking of R(St , e) that is only available to St . Moreover, this kind of ‘egocentric’ thought is inferentially isolated from non-egocentric thoughts. The fact that one has to be St in order judge R(St , e) first-person redundantly shows that this kind of thinking is self-locating. If St thinks of e egocentrically as past, and does so correctly, then it follows that St stands in the ‘later than’ relation to e. If e is a traumatic event, it follows that it is appropriate for St to feel relief that e is over. So, combined with the kind of evolutionary considerations to which attention was drawn by Maclaurin and Dyke (2002), we have the full answer to Prior’s challenge. John Perry (1977) first used the phrase ‘self-locating’ when writing about indexicals. Setting aside unusual contexts such as Predelli’s (1998) answering machine, Kaplanian ‘monsters’ (Kaplan 1989), and so on, one can only correctly speak about a certain person as ‘I’ if one is that person, one can only ¹⁸ Someone could infer from other, non-egocentric knowledge, that the conditions were in place for St to judge, first-person redundantly, that R(St , e). But to know this about a St ’s thought is not the same as being able to think that thought.

TENSE AND EMOTION

27

correctly speak about a certain place as ‘here’ if one is at that location, and one can only correctly speak about a certain time as ‘now’ if one is ‘at’ that time (which, in this chapter, I have taken as equivalent to being identical with the relevant person-stage existing at that time). But I think the phrase ‘selflocation’ is better understood in terms of first-person redundancy. The fact that one can only use indexical linguistic expressions in the relevant contexts does nothing to explain how one has to think in order for one’s thoughts to have the relevant self-locating feature. It doesn’t explain the epistemology. Knowing the truth-conditions for indexical linguistic expressions does not help; it always results in Prior’s challenge (‘why should anyone thank goodness for that?’).

2.8 Godlike actions, godlike emotions I shall close with a brief discussion of the scope of the claims that I am making. I have claimed that thoughts belonging to a certain class—egocentric thoughts, characterized by the feature of first-person redundancy—are necessary in order for certain kinds of emotional reactions to past, present, or future events to be appropriate. While I have not emphasized the point here, I also claim that such thoughts are necessary for action.¹⁹ Perhaps this latter claim is immediately quite plausible, given that actions are often motivated by emotions. But is egocentricity of this kind necessary for emotional reactions in all kinds of beings, or just those that have certain contingent features possessed by humans? If these claims apply in all cases, then they may be of deeper philosophical interest, in terms of our understanding of the very nature of emotional reactions (and actions). The idea that the above claims apply only to humans, or to creatures similar to humans, might be motivated by the claim that there could be godlike beings who took a view of space and time as a whole, and reacted to the world, and acted upon it, without having to think of themselves as located within it (even if they were in fact so located). They would not think of places as here, of times as now, or of events as being past, present, or future. A godlike action would be one in which the godlike agent simply thought of a state of affairs at a given place and time (e.g. ‘rain in St Andrews on 2 April 2022’), and in doing so was able to bring about that state of affairs without any need for egocentric thought. We can even imagine the state of affairs being brought about by the spatiotemporally located movements of the agent, so long as we thought of ¹⁹ Again, for full arguments see Prosser (2015).

28

SIMON PROSSER

the details of the execution of the action as being delegated to non-conscious systems that controlled the movements of the body. Opponents of the claim that the first-person perspective is universally ‘essential’ are likely to want to defend the possibility of godlike actions and emotions. But I think the temporal case raises immediate problems for both.²⁰ Consider, first, godlike actions. A godlike being could only decide the overall state of the world over a period of time. They could not interact with the world (perhaps their bodies might do so, if they acted on the world via their bodies, but they would be oblivious to this). The very notion of person-level interaction requires that one choose how to act in ways that depend on the state of the world at the time of action. Consider, for example, a conversation between two people, consisting of their taking turns at producing utterances over a period of time. What one says at any given time depends on what has been said before. Suppose it is now the turn of the godlike creature to speak. It seems they must have a representation of the conversation up to that point, and then decide what to say next. But this requires that the world be represented in a way that acknowledges that the sequences of utterances have happened, and cannot now be changed, and that the only thing to be decided is what to say next. But in that case, an egocentric, tensed representation has been presupposed. Godlike emotions strike me as even more problematic. What would it take for there to be a being who experienced emotions at the appropriate times, yet did not think about those times as being past, present, or future? Perhaps someone might claim that this could happen, because there could be a being who was simply wired up to automatically experience the right emotion at the right time. They would not think of the dental appointment as future, but would nevertheless automatically feel dread when it was in fact so; and they would not think of it as past afterwards, yet would automatically feel relief at that time. It is unclear what the purpose of such emotions would be, in a godlike being. But there is a deeper problem: The proposal treats emotions as nothing but feelings that occur at particular times, as though they were no different from toothaches. But emotions are not like that; they are intentional states, directed at states of affairs. One doesn’t just dread, or feel relief: one dreads the dental appointment, and one thanks goodness that the dental appointment is over. I take this to be a necessary feature of genuine emotions. Consequently, one cannot feel relief that the dental appointment is over without thinking

²⁰ See also Prosser (2015: 221–223) for some other objections.

TENSE AND EMOTION

29

of it as over. It seems to me that scepticism about the role of the first-person perspective, with its focus mainly on actions, has neglected this crucial fact about our emotional reactions to the past, present, and future.²¹

2.9 Conclusions I have argued that the emotional reactions produced by thoughts about the past, present, and future come about in part because of their evolutionary utility, as suggested by Maclaurin and Dyke. But in order to answer Prior’s ‘thank goodness’ challenge, we must also explain how there can be a state of affairs for which it makes sense for a person, at a time, to ‘thank goodness’, yet it would not be appropriate for another person, or the same person at a different time, to have the same reaction. I have argued that this can be explained in terms of the egocentricity of ‘tensed’ thoughts about the past, present, and future. Crucially, egocentricity is understood here in terms of the epistemic property of first-person redundancy. This explains why tensed thoughts—that is, temporally egocentric thoughts—are inferentially isolated from non-egocentric (untensed) thoughts. It also explains the sense in which such thoughts are self-locating: it is only possible for a specific person-stage to correctly think of a time or event as past, and doing so entails that the person-stage stands in the relation to the event that makes the relevant emotional reaction the appropriate one for that person-stage (and no other) to have.²²

²¹ It has sometimes been suggested that it is an essential part of an emotion that its content contains an evaluation of the object of the emotion. The object of dread may be represented as dangerous, for example. See Tye (2008) for a theory along these lines. Evaluations tend to involve relations to the subject. If I judge that something is dangerous, for example, then the state of affairs with which I am really concerned is that the object presents danger to me (it might not be dangerous to someone else— someone stronger, or faster, for example). This might suggest an intriguing possibility that emotional states themselves are egocentric, and essentially involve first-person redundancy. I shall leave this as a speculation; the claim of this chapter is only that egocentric thoughts about the past, present, or future are an essential enabling condition for emotional reactions (rather than the stronger but far more speculative claim that egocentricity is partially constitutive of the emotional reactions themselves). ²² I am grateful to Kasia Jaszczolt and an anonymous referee for helpful comments on a written draft of this chapter, and for feedback from audiences in St Andrews, Leeds, Edinburgh, and Cambridge. No new data were created during the research.

3 An exploration into construals of subjective time in poetry Anna Piata

3.1 Introduction For philosophers, the study of time revolves around the metaphysical question: is time real? For linguists, the question when studying time is a different one: how is time encoded and expressed in natural languages? And psychologists are concerned with yet another question: how is time experienced and represented in the mind? Different questions elicit, as expected, different answers, and the picture is far from being conclusive (for an overview see Jaszczolt and de Saussure 2013). While drawing on philosophy, psychology, and linguistics, this chapter sets forth to address a rather unconventional question: can poetics illuminate our understanding of human time? Its starting point is the assumption that our expression of time, however creative, taps into our experience of time and reflects the different ways in which experienced time is represented, namely the human concept of time.¹ In other words, in the present work I consider poetry as a window into the conceptual representations of time and, more specifically, of felt temporal experience. Whether this can illuminate our understanding of time per se is a contentious matter (cf. Jaszczolt 2020). This approach presents itself with opportunities and challenges, which I aim to pursue in this chapter. The way we speak about time is awash with what Prosser calls ‘obscure metaphors’ (2013b: 315), that is, everyday expressions like ‘time flies’, ‘those sad days are behind’, ‘we are approaching the holiday season’, etc. Prosser admits that metaphors of this sort are the only available way with which we can ¹ The present work builds on the premise that we perceive mental images (that is, representations) rather than the things themselves. I take the phenomenal character of temporal experience as a building block for the conceptual representations of time, and analyse the language data as evidence for the structure of the human concept of time (see Section 3.4). The debate on representationalism vs non-representationalism vis-à-vis time is beyond the scope of this chapter. On debates concerning representationalism see Torrengo, this volume.

Anna Piata, An exploration into construals of subjective time in poetry. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Anna Piata (2023). DOI: 10.1093/oso/9780192896445.003.0003

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

31

talk about time. In fact, this assumption underlies a robust, and still growing, body of research in cognitive linguistics, suggesting that most of our language and thought about time is metaphorically structured in terms of space and motion (see Clark 1973; Traugott 1978; Lakoff and Johnson 1980, 1999; Alverson 1994; Radden 2003; Moore 2006, 2014; Bender and Beller 2014; cf. Galton 2011). This cross-domain mapping from space/motion onto time surfaces in language in the form of everyday metaphorical expressions as the ones mentioned above (see Section 3.4.2). In short, language serves as a reflection of the metaphorical patterns with which the concept of time is organized in the mind. The alleged obscurity of temporal expression is at the core of this chapter: its focus is on how time is figuratively represented in the discourse domain of poetry. Poetics is expected to mirror the perplexity and diversity with which time is experienced by humans. The obscurity of metaphor is thus taken as a luminous signal for the phenomenology of time,² conceived in terms of the different ways in which things appear to us as temporal or in which we experience time, thus serving as building blocks for the concept of time. The relevance of experience for language and meaning-making is central to the cognitive linguistic view of meaning: as Talmy (2000: 4) puts it, cognitive semantics can be seen as ‘a branch of phenomenology’ in the sense that its main object of study is qualitative mental phenomena as these exist in awareness (Zlatev 2016). Our everyday experience of time often suggests a deviation from the uniform measurement of time by the clock and even a breach of the social norms that regulate temporality. This phenomenon comes up with different labels in the time-related literature, such as ‘psychological time’ (Fraisse 1963; Block 1990); ‘lived time’ (Minkowski 1970); ‘psychic time’ (Traugott 1978), and ‘subjective time’ (Flaherty 1999; Arstila and Lloyd 2014a; Tho¨nes and Stocker 2019). These terms are derived from scholars with different disciplinary scopes (ranging from social and cognitive psychology to linguistics and cognitive science) and are aimed at describing the experience of time across an array of different contexts: altered states of consciousness, mental disorders, as well as everyday situations that are emotionally coloured. Nevertheless, they share a common assumption that our concept of time is derived from a phenomenologically real experience. In this chapter I opt for ‘subjective time’ as an umbrella term that captures the representations of experienced time across the board (see Section 3.3). ² By referring to the ‘phenomenology of time’ I do not wish to subscribe to a particular philosophical tradition. I use the term quite loosely as an umbrella for the different ways in which the phenomenal character of time manifests itself in human experience.

32

ANNA PIATA

It is worth noting that while examining construals of subjective time I am interested in how these are represented both at the level of language (in the form of figurative expressions of time) and with regards to mental representations of time. This is a legitimate endeavour within cognitive linguistic accounts of metaphor, which consider metaphorical linguistic expressions as a corollary of cross-domain mappings established in the cognitive system. But the present work is not concerned specifically with metaphors of time, but rather looks at figurative expressions of any sort. Moreover, it aims to embrace the phenomenological experience of time as a layer (or a building block, for that matter) for our concept of time. One therefore may question how legitimate it is to bring together the phenomenological experience of time, the human concept of time, and the linguistic expression of it. In this respect, Jaszczolt (2020: 1874) has convincingly argued for a ‘unified analysis of the three domains’. The present work dovetails with such a methodological move in the study of time, while it also differs from it in important ways. Jaszczolt examines temporal reference in natural languages (‘timeL ’ in her terminology) with a view to shedding new light on the concept that is being expressed. At the same time, she admittedly leaves aside the experience of the flow of time, as well as the concept of the flow itself. It is this particular dimension that is at the core of the present study together with two other ways in which humans experience time, namely temporal order and duration. I will return to these in Section 3.4. Still, a question is pending as to how such an endeavour is methodologically feasible. This includes two aspects: which texts to examine (and how to track them down), and how to identify, and analyse, time. For the first part, data were collected from English PoetiCog, a collection of poetic texts written originally in the English language (across geographical latitudes and regional, or other, varieties). Searches were not focused on specific authors, eras, or artistic movements. The corpus is accessible through the Sketch Engine software, which enables targeted searches on the basis of particular key words (or combinations of them). For the purposes of this study, the keywords included lexemes that belong to the semantic field of time, including time, past, present, future, and calendric terms (such as moment, day, century, etc.). This rendered a sum of results that had to be skimmed manually so as to discard any irrelevant instances. The selected poems are presented and analysed in Section 3.4. But in order to navigate through the language data that pertain to the experience of time, a conceptual framework was needed that would reflect mental representations of time in a way that is phenomenologically realistic.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

33

For this purpose, I draw on the conceptual framework proposed by Tho¨nes and Stocker (2019), according to which the experience of time is represented in three main conceptual structures: temporal processing; the passage of time; and duration. In the light of this conceptual framework, ten examples were selected from the corpus data, each one representing one of these schemas (with the passage of time clearly prevailing over the others). Although far from being uncontroversial, this framework qualifies as suitable for the following reasons: it is interdisciplinary, drawing on evidence mostly from cognitive psychology but also from philosophy and cognitive linguistics; it is up to date, based on most recent evidence in the relevant disciplines; and it is parsimonious. Certainly, the very nature of time posits a great challenge to anyone intending to parse it into distinct, meaningful categories, even more so when this endeavour is specifically focused on representations of our experience of time—diverse and distorted as it may be. I will come back to these issues in Section 3.3. One last issue remains open, related to what this approach entails for understanding linguistic meaning. The time expressions I examine are figurative, yet metaphor alone does not suffice to explain how meaning is derived. This seems to posit a challenge vis-à-vis meaning-making. Thus, in Section 3.5, I pursue the argument that the processing of these expressions is likely to involve a mental simulation of temporal experience in line with the tenets of embodied cognition. If on the right track, this idea opens a new path in the study of embodiment, time, and meaning construction. The following section is aimed to offer an overview, however sketchy, of existing work on the phenomenology of time in philosophy before introducing in some detail the framework followed in this study, and the poetic texts to be examined.

3.2 On the phenomenology of time: A brief overview In his book The Order of Time, physicist Carlo Rovelli (2018: 213) contends that ‘in the elementary grammar of the world’ there is no space or time. Nevertheless, for humans both space and time are two fundamental aspects of experience, which are also intertwined. As Rovelli (2018: 208) puts it elsewhere, time is ‘the form in which we… interact with the world: it is the source of our identity’. Here Rovelli invokes time’s foundational role for human psychology. Although this may sound more than obvious nowadays, the trajectory has been long, featuring quite diverse perspectives on the matter. Here I will focus,

34

ANNA PIATA

albeit briefly, on some philosophical perspectives that marked this trajectory, presented from the perspective of a non-philosopher. Kant was the first to take subjectivity into consideration in the account of time in his philosophy of transcendental idealism (1838). For Kant, time is no longer conceived as an external, objective reality. It is considered an inherent mental prerequisite for humans to interact with the world, a necessary representation of the human mind that emerges from the conscious experience of subjectivity. Kant locates time, along with space and causality, onto an intermediate level between objectivity and subjectivity. Although this is a major advancement against the ‘objectivist’ view of time sustained back then, Kant still adheres to a linear, universal, and homogeneous conception of time. For him, time is ideal; therefore, it is not an empirical question but a metaphysical issue. It is only within the phenomenological paradigm that time shifted to a representation of the perceiving subject that is grounded in experience. Phenomenology can be described as any attempt to describe things as they appear (phainomena) without metaphysical and theoretical speculations of any sort. Its object is ‘the study of human experience and of the ways things present themselves to us in and through such experience’ (Sokolowski 2000: 2). Edmund Husserl, one of the founders of the phenomenological movement, introduced the notion of ‘internal time’ (1964 [2019]), according to which time is inextricably associated with human consciousness. For Husserl, time is considered primarily as a chain of events from the recent past to the immediate future, which amount to a sense of the present; what he calls the ‘living present’. It is rather the essential structure of consciousness that makes possible any unified perception of an object across subjective moments. Time is now viewed as a dimension of our being rather than as an object of our knowledge. As Merleau-Ponty (1962: 146) puts it: ‘I am not in space and time, nor do I conceive space and time; I belong to them, my body combines with them and includes them.’ In this account, the problem of time is intertwined with embodied subjectivity; it is a temporal flux of bodily consciousness. Merleau Ponty’s view was indirectly influenced by yet another existential interpretation of the phenomenology of time, fleshed out by Heidegger with his notion of Da-sein (‘Being in-the-world’). Heidegger defines time as ‘the horizon of every understanding and interpretation of being’ (1996 [1927]: 17), which is grounded in the sense of finitude that the conscious ‘Being in-the-world’ has. Heidegger’s account frames time and its correlation with subjectivity in death, conceived as the knowledge of its anticipated (and inevitable) occurrence.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

35

Last, but not least, two more accounts of the phenomenology of time need to be mentioned, each one focusing on a different aspect of temporality: duration, and instant. The former was famously proposed by Bergson (1960 [1889]) and sees time as a continuous but heterogeneous experience of duration (durée in his terminology), also conceived in terms of the human consciousness. The latter belongs to Bachelard (1994 [1932], 1993 [1936]) and conceptualizes time as a segmented and discontinuous experience of instants. Finally, the subjectivity of time can only happen against the backdrop of socalled ‘social time’ (Elias 1992). Time is a social construct formulated out of the need for the organization of social life and is thus established as a normative principle. The shared representations of time entertained by clock-based time are both derived by, and dictate, society. Whether subjective time is subsumed under social time or vice versa is a question that goes beyond the scope of this chapter.

3.3 A conceptual framework for the mental representations of time As already mentioned, for the purposes of this chapter I draw on the conceptual framework of time proposed by Tho¨nes and Stocker (2019). In their terminology, subjective time refers to ‘the conscious or unconscious mental representation of physical temporal information’ (ibid.: 114), which thus serves as an umbrella term for different aspects in the representation of time. Following Tho¨nes and Stocker, in this chapter I opt for the term ‘subjective time’ in order to refer to the different, and diverse, representations of our experience of time.³ The framework proposed by Tho¨nes and Stocker (2019) distinguishes between three mental representations of time, which may often co-exist or complement one another in experience but are conceptually distinct. The first one is that of processing basic temporal information, such as the order of events and simultaneity, and is referred to as temporal processing. The second one corresponds to the passage of time, which includes variation in time perception (as in, e.g., ‘Time flies’), as well as the distinction between the past, the ³ Some scholars in social psychology (Flaherty 1999) and cognitive linguistics (Fauconnier and Turner 2008) define subjective time as the opposite of objective time, which is attuned to clock measurements and is therefore uniform across experiencers and latitudes. Subjective time, therefore, hinges on deviation from the norms and expectations of objective time. In contrast, Wearden et al. (2014: 288) use the term ‘subjective duration’ in a more general sense that encompasses ‘all psychological judgments and events’ (emphasis in the original).

36

ANNA PIATA

present, and the future. The last one relates to duration, namely how much time has elapsed during an interval.⁴ While temporal processing corresponds to basic and largely unconscious processes, the passage of time and duration are more consciously processed, require higher order cognitive processes (such as attention and memory) and are more closely related to what is generally considered as time perception. Time perception, in the form of duration and passage, constitutes the bulk of most psychological research on time (for an overview see Wearden et al. 2014 and references therein). At this point, it needs to be noted that terminology is far from being consistent, and no consensus has been reached among scholars and even across disciplines as to how to categorize the different mental representations of time. For instance, in his account of the fundamental aspects of temporal experience (so-called ‘elementary time experiences’) Po¨ppel (1978) considers order and simultaneity as distinct concepts rather than integrating them into a single category. Experience of non-simultaneity, he argues, is not the same as experience of temporal order because when two events occur close to one another we may perceive them as successive without being able to say which one occurred first. Another contentious issue relates to the concept of duration. Wearden et al. (2014) define duration (in particular, ‘subjective duration’) in terms of judgements as to how much time has elapsed, as well as how fast time is perceived to pass. Thus, in this account the term ‘duration’ conflates two aspects in the perception of time which Tho¨nes and Stocker categorize under different labels. Likewise, Flaherty (1999) introduces the term ‘protracted duration’ in order to refer to the sense of time as passing slowly, rather than an estimate of the lapse of time (as in Tho¨nes and Stocker). Evidence, therefore, is far from being conclusive, perhaps reflecting the extent to which mental representations of time are multifarious and at once subtle. In this study, for methodological reasons I will follow the categorization proposed by Tho¨nes and Stocker (2019), although I acknowledge that the issue is far more complex.

3.4 Poetics as a window into the phenomenology of time In this section, each category of time representations—temporal processing, the passage of time, and duration—is examined in the light of language data ⁴ An anonymous reviewer has pointed out that the directedness of time, namely the ‘arrow of time’, is missing from Tho¨nes and Stocker’s framework. I am not sure, however, if this would qualify as a distinct mental representation of time in Tho¨nes and Stocker’s framework or it could be subsumed, in a rather reductionist way, under the passage of time. I consider this to be tangential to the scope of this chapter and leave it open for future research.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

37

from poetry. While drawing on the extant evidence regarding the various manifestations of the concept of time, the aim of this section is to make a case for how these are exploited by poets, often in non-conventional ways, to the effect that they reflect the phenomenological experience of time.

3.4.1 Temporal processing For Tho¨nes and Stocker (2019) temporal processing refers to a basic sensory capacity to detect relations of succession or simultaneity. Being automatic and primarily unconscious, temporal processing is considered as forming the basis for the more complex representations of time passing and duration.⁵ In experimental terms, temporal processing can be identified through tasks of temporal-order judgement and judgement of simultaneity. The stimuli involved in these experiments correspond to temporal information in the range of milliseconds. Unlike the passage of time and the perception of duration, the processing of temporal order and simultaneity does not rest upon cognitive resources, and this is why it is done automatically, but, as the authors admit, when temporal processing includes events some cognitive processes (e.g., attention) may be recruited and thus the mental representation of order (or simultaneity) becomes conscious. The poetic extracts that follow are far from being typical examples of temporal processing as conceived in conceptual and experimental terms. Instead, they illustrate how times and events are ordered in subjective experience. Consider first the poem ‘An Aged Man Remembers April’ by Raymond Fischer:⁶ (1) Do I remember April? Yes and no – Depending on the meaning of remember. It marks a season and it still comes on between the chill of March and warmth of May. If that’s remembering, perhaps I do. Raymond Fischer, ‘An Aged Man Remembers April’ (my emphasis) Here temporal order relates to the succession of months in the calendar: when asked to recall April, the poetic voice is able to say that it comes after March ⁵ One could counterargue that no temporal processing can occur without a primary conception of duration in the form of distinct units of stimuli, however small or short they may be. This issue goes beyond the scope of this chapter. ⁶ The poem ‘An Aged Man Remembers April’ by Raymond Fischer originally appeared in the April 1984 issue of Poetry. I would like to thank the magazine’s editors for granting me permission to reprint the poem in the present work.

38

ANNA PIATA

and before May. The succession of the months is also reflected in nature, as suggested by the ‘chill’ (in March) and the ‘warmth’ (in May); April must be somewhere between the two, not cold any more but not really warm either. However, this is only the meaning that is explicitly communicated. The opening lines of the poem implicate that remembering April may be more than that: ‘Do I remember April? Yes and no –/ Depending on the meaning of remember’. The poetic voice can only recall that April ‘marks a season’, located between March and May. But the if-clause in the poem’s last verse (‘If that’s remembering, perhaps I do’) invites the inference that remembering April may go beyond a transitional season and a unit in the calendar. Presumably April is also (if not more importantly) about the memories it is filled with. Likewise, it signals the advent of the spring with the rebirth of nature and new, rejuvenating beginnings, coloured with a sense of beauty, happiness, and optimism. The reluctance (or incapability?) of the poetic voice to recall April most likely suggests a painful memory, one that is better to repulse.⁷ If anything, the question is left open for the readers to decipher and answer, deriving their own inferences. The two-layered meaning of remembering in the poem—explicit and implicit—reflects a distinction between objective time (observed in nature and engraved on the calendar) and its subjective reckoning: what April means to the poetic voice is far richer and more complex than yet another season in the cyclic year. Clearly, this is significantly different from the notion of ‘temporal processing’ as studied by psychologists, viz. a relation of order or simultaneity between events of short duration that happen closely to one another. Fischer’s poem is about temporal order but not in the sense of distinguishing between successive (or simultaneous) stimuli in the present; it is about the order of events in the natural world, reflected also in shared artefacts for time-telling. But Fischer’s poem is also about memory and the perplexities of temporal experience; even if the order of events in the world is standard and regular, how they are felt (and remembered) may not be. In essence, what the poem illustrates is that there is a discrepancy between the temporal order found in subjective memory and that of objective temporality. The perceiving subject is fully aware of this discrepancy but can only retrieve temporal order in nature and in the calendar; what has happened has faded away.

⁷ This echoes the opening lines in T. S. Eliot’s ‘The Waste Land’: ‘April is the cruelest month, breeding / Lilacs out of the dead land, mixing / Memory and desire, stirring / Dull roots with spring rain.’

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

39

A breach between the objective order of events and subjective experience is also suggested in the opening lines of T. S. Eliot’s ‘Burnt Norton’, the first of his well-known ‘Four Quartets’:⁸ (2) Time present and time past Are both perhaps present in time future, And time future contained in time past. If all time is eternally present All time is unredeemable. What might have been is an abstraction Remaining a perpetual possibility Only in a world of speculation. What might have been and what has been Point to one end, which is always present. Footfalls echo in the memory Down the passage which we did not take Towards the door we never opened Into the rose-garden. My words echo Thus, in your mind. T. S. Eliot, ‘Burnt Norton’ (my emphasis) These lines are about the past, the present, and the future, suggesting that they are all fused into an inseparable whole. Having said that, it appears that this poetic extract evokes a disruption vis-à-vis temporal order: recall that we conceive of the past as what has happened before the present moment and the future as what will follow after. Eliot’s lines challenge the boundaries between past, present, and future, as well as the standard sequence of earlier vs later events. Whether the experience of an event as past or present is more fundamental than the experience of order (or vice versa) is a contentious matter, reflecting the discussions in the domain of time metaphysics that date back to McTaggart’s (1908) distinction between the A-series of tensed time (i.e., events that are past, present, or future) and the B-series of tenseless time (i.e., events that are earlier or later than other events, or simultaneous with them). In what follows, I wish to suggest that Eliot’s verse can be better understood in terms of William James’ (1950 [1890]) notion of ‘specious present:⁹ ‘If all time is eternally present / All time is unredeemable’. Although the term ‘specious present’ was coined by the psychologist E. R. Clay, it was James who ⁸ I would like to thank Faber & Faber for granting me permission to reprint extracts from T. S. Eliot’s poetry in the present work. ⁹ See Andersen (2014) for an overview and evaluation of the notion of ‘specious present’.

40

ANNA PIATA

defined it as ‘the prototype of all conceived times…, the short duration of which we are immediately and incessantly sensible’ (ibid.: 632). The phenomenal present essentially amounts to the feeling of nowness, which extends from a few seconds to less than a minute, thus comprising both an earlier and a later part. In contemporary terms, this can be best described as a perceptual gestalt. Unlike other perceptual stimuli, time cannot be perceived directly through the senses and therefore needs to be inferred and reconstructed in the human mind.¹⁰ Of course our experience of the present is constantly changing: this may explain why the present and the past are ‘perhaps’ present in the future. Meanwhile, every small single event in the present is past by the time we perceive it: this may justify in what ways future time can be ‘contained in time past’. As Eliot writes in the remainder of the poem: ‘What might have been and what has been / point to one end, which is always present.’ These lines elicit an interpretation that clearly echoes James’ phenomenology of the specious present. If anything, Eliot’s poem challenges, if not breaches, the order of past, present, and future as these are conceptually and socially represented: it rather portrays how they are perceived by the self. What Eliot writes later in the poem fairly speaks in favour of such a view of time: ‘Time past and time future / Allow but a little consciousness’. As Wittmann (2014: 513) asserts, ‘evidence for the discreteness of temporal processing does not necessarily entail the existence of discreteness in the subjective experience of time’. Indeed, the two poems discussed in this section illustrate that the subjective experience of temporal order may not be fully discrete, although it rests upon discrete constructs of temporal processing: before May and after March (in Fischer’s poem), and the past, present, and future (in T. S. Eliot’s poem). Thus, in the two poems temporal order is represented in a rather distorted way that reflects inner experience rather than align with the discreteness of temporal processing assumed in cognition. While temporal processing is often considered less complex compared to other mental processes in time perception, it is likely, as Lloyd and Arstila (2014: 659) suggest, that ‘simultaneity and order are no less multifarious than intervals and durations’. The two poems above attest to such a conjecture. In the next section, I move to poems that exemplify how the passage of time is represented in the context of poetry.

¹⁰ In his analysis of the phenomenal present, Wittmann (2014: 513) considers the possibility that events may be perceived at present, and experiences may be restricted to the present, but the present itself is not experienced.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

41

3.4.2 The passage of time The passage of time comprises two aspects of temporal representation, which Tho¨nes and Stocker (2019) consider to be interwoven: the localization of the past, the present, and the future in the mental timeline, and the speed of time passing. The speed of time passing is often perceived as the most prominent manifestation of time perception, commonly found in the experience of everyday life as when feeling that time drags while waiting for an important medical result or that it flies when engaged in an amusing activity. These types of temporal experience, together with the sense of time as coming to a halt (for instance, when in shock), constitute what has been considered as subjective time par excellence (Flaherty 1999; see also Jaszczolt, this volume). According to Tho¨nes and Stocker (2019), speed presupposes a spatial distance that is to be traversed; this suggests that the mental timeline is necessary for conceiving at what speed time passes. As the authors themselves admit, this approach is debatable and remains to be substantiated through further experimental investigation. (Note that the notion of duration is also spatially represented in terms of extension; see Section 3.4.3.) In cognitive linguistics, it has been established that the spatialization of time comes in two variants: the Ego-moving metaphor of time, and the Timemoving metaphor (Lakoff and Johnson 1999: 141–147). While the former conceptualizes times and events as destinations towards which Ego is moving (e.g., ‘We are approaching the holiday season’), the latter represents times and events as moving towards, and past, Ego who remains static (e.g., ‘The Easter recess is approaching’). Regardless of which entity is moving (Ego or Time), in Western languages (including English) the spatial structure of time employs the front/back axis for the representation of the future and the past: future events are located ahead of the experiencer and past events behind, while the present is always co-located with Ego.¹¹ In the cognitive linguistic literature, this is referred to as the Time Orientation Metaphor (ibid.: 139– 140). Although universal, this metaphor is also subject to significant cultural variation, with some languages featuring the opposite pattern.¹² In linguistic (and poetic) terms, this category introduces the spatial metaphors for time that have been established by cognitive linguists. A case in ¹¹ It is worth noting that temporal co-speech gestures make use not only of the sagittal axis of front/back but also, and perhaps more pervasively, the lateral axis from-left-to-right (Casasanto and Jasmin 2012). The latter manifests itself also in the graphic representation of timelines but is not attested in any natural languages (Radden 2003). This dissociation between language and other modalities gives rise to much more complex and versatile mental representations of time (Casasanto 2016). ¹² This is the case in Aymara (Núñez and Sweetser 2006) and Vietnamese (Sullivan and Bui 2016).

42

ANNA PIATA

point is found in the following lines from Owen Meredith’s poem ‘The Dead Pope’: (3) By the mere faint thought of it, well I wish Such a heaven on earth were hardly amiss; And I hold it no crime to set it in rhyme That I think a man might pass his time In company worse than this. But, however we pass Time, he passes still, Passing away whatever the pastime, And, whether we use him well or ill, Some day he gives us the slip for the last time. Even a Pope must finish his fill, And follow his time, be it feast time or fast time. As it happened with this same Pope. No doubt What sleep was his after that last bout, When he could not wake! so they laid him out. ‘He is gone,’ they said, ‘where there’s no returning. Of the college who is the next to come?’ Then they set the bells tolling, the tapers burning, And bore him up into Peter’s dome. Owen Meredith, ‘The Dead Pope’ (my emphasis) These lines instantiate what Lakoff and Turner (1989: 70–72) call composition of conceptual metaphors. In their account, composition corresponds to combining two (or more) conceptual metaphors, and is considered as one of the mechanisms with which poets and authors can reach metaphorical creativity. In (3), there is composition of an Ego- and a Time-moving metaphor in one single line: ‘however we pass Time, he passes still’. Still, and contrary to Lakoff and Turner’s claims, this linguistic expression does not come across as particularly creative: the use of the verb pass to talk about the passage of time is anything but novel or striking. To say that time passes and we pass time sounds like a triviality. In fact, the poem assumes a conceptualization of all-powerful time. Creativity is about how metaphors are combined in particular ways in order to impart new inferences, which are not conventionally derived from the relevant metaphorical expressions. This is also the case in (3). The conventional metaphors of ‘Time’ and pass are further combined with a representation of time as a resource in life (‘whether we use him well or ill’), a rather conventional metaphor too. But all these representations of time (diverse and at once

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

43

complementary) are subsumed under his overarching power: ‘Some day he gives us the slip for the last time.’ This personification metaphor results in an enemy construal of time; we are racing against time, or time is running after us and ultimately no one will escape his power, not even the Pope. The inference is rather straightforward: the passage of time will eventually lead to death and this holds indiscriminately for all humans. In figurative terms, this is derived from a composition of conventional metaphors of time together with a novel personification of time as giving us the slip for the last time. The passage of time is thus embedded in a broader construal that only validates its ultimate power over humans. The next poem also tackles how time impacts humans. The whole poem, entitled ‘The Paradox of Time’, is cited below: (4) Time goes, you say? Ah no! Alas, Time stays, we go; Or else, were this not so, What need to chain the hours, For Youth were always ours? Time goes, you say? ah no! Ours is the eyes’ deceit Of men whose flying feet Lead through some landscape low; We pass, and think we see The earth’s fixed surface flee: Alas, Time stays, we go! Once in the days of old, Your locks were curling gold, And mine had shamed the crow. Now, in the self-same stage, We’ve reached the silver age; Time goes, you say? ah no! Once, when my voice was strong, I filled the woods with song To praise your ‘rose’ and ‘snow’; My bird, that sang, is dead; Where are your roses fled? Alas, Time stays, we go!

44

ANNA PIATA

See, in what traversed ways, What backward Fate delays The hopes we used to know; Where are our old desires? Ah, where those vanished fires? Time goes, you say? ah no! How far, how far, O Sweet, The past behind our feet Lies in the even-glow! Now, on the forward way, Let us fold hands, and pray; Alas, Time stays, we go! Henry Austin Dobson, ‘The Paradox of Time’ (my emphasis) This poem exploits another mechanism of metaphorical creativity: questioning. According to Lakoff and Turner (1989: 69–70), questioning is about challenging the conceptual patterns predicted by a particular metaphor. Indeed, this is the case in the poem. It negates the passage of time (‘Time goes, you say? ah no!’) and reverses the pattern of the Time-moving metaphor: it is us that will go, Time will stay (‘A las, Time stays, we go!’). These expressions are repeated, whether separately or jointly, throughout the poem, thus serving as the poem’s main theme while iconically suggesting recurrence in the experience of time. The poem thus resonates with another aspect of our experience: time will keep going even when we are no longer here (note that here ‘go’ evokes the DEATH IS DEPARTURE metaphor). This conceptualization alludes to time as an infinite flow that exists independently of the experiencer. As its title suggests, the whole poem is about the paradox of time. In linguistic terms, this translates into the verbal paradox that is evoked through the time metaphor: ‘A las, Time stays, we go!’. An iconic link is established between the poem’s meaning and the linguistic means used to this end; language itself iconically reflects the meaning conveyed. The paradox essentially juxtaposes Time with ‘we’, while it exploits the polysemy of go as meaning either ‘move to a different location’ or ‘die’. Similarly to the previous poem, in this juxtaposition Time is the one who has the upper hand. The verbal means used by the poet communicate also how he feels about it. While ‘alas’ explicitly suggests regret, the verbal paradox renders the poet’s otherwise painful remark witty and somewhat whimsical. Another way for exploiting a conceptual metaphor is through extension (Lakoff and Turner 1989: 67). By extending a metaphor a poet or author

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

45

activates an aspect of the source domain that does not typically surface in language use. Such a case is exemplified in a poem by Tanay Pathak, titled ‘Time Flows’, in which time is conceptualized in a fairly common way as a flow, a variant of the well-known Time-moving metaphor (Lakoff and Johnson 1999: 144–145). Flowing denotes a particular motion event, namely that of a fluid in a current or stream. Unlike other types of motion, a flow is steady and continuous and therefore qualifies for the conceptualization of time: our commonplace understanding of time is that of a continuous sequence of distinct events. As suggested in the poem, the flow of time is what enables our ability to feel, perceive, and act in the world. But what is novel and original in the poem is that it introduces a deictically grounded recipient, i.e., ‘you’, as an agent that attempts to stop and block the flow of time, a task that is utterly unattainable: there is no way for humans to impede or control the passage of time. Here, the metaphor does more than merely extending the pre-existing mapping of a flowing substance onto passing time. It rather evokes our own understanding of time as a force over which we have no control. Any attempt to take control over time, whether by seducing or by destroying it, is evidently failed, as suggested by a sequence of verbs that are paired together through rhyme: ‘turn/burn’, ‘harm/charm’, and ‘kill/fill’. The fact that humans are powerless in the face of time infiltrates the whole poem as each and every effort to resist time is repeatedly accompanied with an explicit admission that, no matter how much you try, in the end you can’t escape time; time will keep flowing no matter. It is worth noting, finally, that the textual pattern of parallelism that structures the poem (with the repetition of the time expressions, as well as through rhyme) iconically reflects recurrence and continuity similarly to the passage of time. Once again, iconicity is at work only to portray how the passage of time is felt and understood by the human subjects. So far, the poems discussed in this section illustrate the powerfulness of time (or, to put it differently, the powerlessness of humans against time). In the following extract, cited in (5) below, the representation of time shifts to its felt experience by two lovers. While portraying a scene between the two lovers, the poem sketches an emotional atmosphere rather than providing an accurate description of their encounter. It is in this context that time is said to pass in a way that goes beyond the ordinary: (5) Two lovers, here at the corner, by the steeple, Two lovers blow together like music blowing: And the crowd dissolves about them like a sea. Recurring waves of sound break vaguely about them, They drift from wall to wall, from tree to tree. ‘Well, am I late?’ Upward they look and laugh,

46

ANNA PIATA

They look at the great clock’s golden hands, They laugh and talk, not knowing what they say: Only, their words like music seem to play; And seeming to walk, they tread strange sarabands. (…) ‘One white rose . . . or is it pink, to-day?’ They pause and smile, not caring what they say, If only they may talk. The crowd flows past them like dividing waters. Dreaming they stand, dreaming they walk. ‘Pink, – to-day!’ – Face turns to dream-bright face, Green leaves rise round them, sunshine settles upon them, Water, in drops of silver, falls from the rose. She smiles at a face that smiles through leaves from the mirror. She breathes the fragrance; her dark eyes close… Time is dissolved, it blows like a little dust: Time, like a flurry of rain, Patters and passes, starring the window-pane.¹³ Conrad Aiken, ‘The House of Dust’ (my emphasis) In this poem, the passage of time is represented in a highly figurative way through a host of metaphors and similes: it is ‘dissolved’ and resembles ‘a little dust’ and ‘a flurry of rain’ while blowing, pattering and passing on a windowpane. These lines evoke rich imagery in the representation of time, unlike the previous examples (at least in the time expressions that are the focus of this analysis). What all these expressions suggest is that time is evanescent similarly to something that is being dissolved, the dust that blows in the air or the rain that patters and passes over the window-pane. Clearly, the two lovers do not follow the ‘ordinary’ time of the world; instead they are subsumed into their own sense of time passing. In psychological terms, this is not surprising. A compelling body of evidence on the psychology of time suggests that emotions can bring upon a distortion in the regular passage of time; this phenomenon is referred to as the ‘time-emotion paradox’ (Droit-Volet and Gil 2009; see Droit-Volet 2014 for an overview). In the poem above, positive emotions such as arousal, happiness, and excitement are evoked. These emotions are likely to induce a sense of time ¹³ The House of Dust by Conrad Aiken, copyright ©1920 by Conrad Aiken. Used by permission of Brandt & Hochman Literary Agents, Inc.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

47

as passing more quickly (see, e.g., Wearden et al. 2014: 301). This does not seem to be the case in (5), unless one takes the verbs used to denote the passage of time as implicitly suggesting rapidity (‘dissolved’, ‘blows’, ‘patters and passes’). What is rather straightforward in the poem is that the two lovers disregard the time-keeping system: ‘They look at the great clock’s golden hands, / They laugh and talk, not knowing what they say: / Only, their words like music seem to play’. In short, subjective experience alongside its emotional intensity overrides timing mechanisms and other temporal norms. With regard to the figurative expressions of time in (5), it is worth noting the use of the verb pass. Here, pass does not denote the passage of time, as is usually the case, but the flow of the rain on the window-pane. Once again, the poet exploits a pre-existing pattern (here the Time-moving metaphor) and, in this case, elaborates on it. Elaboration, according to Lakoff and Turner (1989: 67– 69), is the mechanism with which a poet or an author transforms a conceptual metaphor by modifying one of its aspects; in the case of time, in particular, this can take the form of a motion verb that is not conventionally used when speaking about time. What makes pass a creative choice here is the fact that it is evoked in relation to rain rather than as a collocate of time. Finally, it is also worth noting that the other verbs used for the representation of time (blow and dissolve) appear early on, at the outset of the scene: ‘Two lovers blow together like music blowing: / And the crowd dissolves about them like a sea’. Used in this way, blowing and dissolving connote togetherness for the two lovers and detachment from the outer world—what may be the case when one is engrossed in love and desire. In a poem dedicated to the French poet Auguste Villiers de l’Isle-Adam, a well-known representative of nineteenth-century symbolism, Aldous Huxley offers an obituary (‘Villiers de l’Isle-Adam’) pronounced by a younger poet to one of his poetic ancestors. In Huxley’s ode to Villiers, time does not pass regularly for him because his work and art exceed the boundaries of his existence. The passage of time for Villiers is disentangled from the ‘quotidian beats’ of ordinary, shared temporality. But unlike the previous examples, this is not so for psychological reasons but rather because of the uniqueness of Villiers’ art that will render him immortal. The poem illustrates how subjective time (the passage of time in Villier’s life as perceived post hoc by Huxley) arises against the backdrop of a shared norm: subjective and objective time are juxtaposed only to suggest that the former wins over the latter. In the cognitive linguistic literature, this translates to a conceptual integration network, also known as a blend, that results from integrating two time-related inputs: the former is that of experienced

48

ANNA PIATA

events, and the latter refers to clock time (Fauconnier and Turner 2008). The first input contains events as they are felt by the experiencer and are therefore conceptualized in terms of experienced motion (that is, motion as it is experienced rather than in its technical sense); e.g., ‘The lecture went by quickly.’ As everyday experience suggests, events can be felt differently across experiencers, therefore for another attendant the same lecture can be experienced as regular or slow. Crucially, events are typically measured through time-telling devices and are thus reckoned in terms of minutes, hours, days, weeks, etc. The latter are contained in the second input, clock time, which is uniform and shared among all experiencers (itself a blend of natural time and the mechanical clock). It is only when experienced events from the first input are integrated with calendric units in the second input that the concept of time arises. The emergent blend of time can derive its topology from either input and thus yield a conceptualization of subjective or objective time. In the case of Huxley’s poem, the long and slow rhythm in which time passed for Villiers clearly overrides the topology of clock time. This is what makes Villiers, according to the poet, immortal. In the last poem to be examined in this section, the focus shifts to the conceptualization of the past and the future. This poem, an extract from T. S. Eliot’s ‘The Dry Salvages’ (the third of his ‘Four Quartets’), manifests an elaborate and somewhat paradoxical construal of time passing, including also the past and the future. The relevant excerpt is cited in (6) below: (6) And under the oppression of the silent fog The tolling bell Measures time not our time, rung by the unhurried Ground swell, a time Older than the time of chronometers, older Than time counted by anxious worried women Lying awake, calculating the future, Trying to unweave, unwind, unravel And piece together the past and the future, Between midnight and dawn, when the past is all deception, The future futureless, before the morning watch When time stops and time is never ending; And the ground swell, that is and was from the beginning, Clangs The bell. T. S. Eliot, ‘The Dry Salvages’ (my emphasis)

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

49

The full import of Eliot’s verse is hard to grasp, since it blends many different aspects of temporality, such as the past and the future, the midnight and the dawn, and of course time itself—more precisely, times. What obscures meaning even more in the poem is the verbal paradox that evokes the representation of time: ‘When time stops and time is never ending’. How is it possible for anything to stop and at once never end? In our commonplace knowledge of the world, stopping entails the end of an action. The passage of time is construed in a way that is anything but straightforward, let alone ordinary. The first lines of this extract introduce the device that is said to tell the time: a tolling bell. This reminds us of death, as tolling bells are associated with a specific funeral sounding, slow and with short gaps between strikes, serving to mark that someone has passed away. As a symbol for the church, the tolling bell also triggers religious associations. The fact that time-keeping is ensured through a tolling bell (rather than a clock) is already novel. What is even more striking is that the tolling bell does not measure ‘our time’ but ‘a time older than the time of chronometers, older / Than time counted by anxious worried women’. A distinction is thus established between human time and universal time, the crucial difference being that the former is finite (‘stops’) while the latter is infinite (‘never ending’). The apparent paradox is thus resolved: there is no single time but a duality, which, as Eliot notes later in the poem, is a divine (i.e., not human) mission: ‘But to apprehend / The point of intersection of the timeless / With time, is an occupation for the saint.’ One last remark needs to be made with regard to the past and the future, as they are both construed in the poem as unreal or non-existent: the past is said to be ‘all deception’ and the future appears as ‘futureless’. This echoes Eliot’s allusion to the ‘specious present’ argued in the analysis of (2); if the past and the future do not exist, what remains is only the present. The fact that the past and the future are said to ‘unweave, unwind, unravel / And piece together’ invites an additional inference, namely that while being dispersed or even disunited parts the past and the future need to be brought back together. As a fundamental dimension in human experience, the passage of time is perhaps one of the major existential issues humans are confronted with. It is this dimension that is most closely associated with the ravages of time and ultimately death, thus reminding us how powerless we are against the passage of time. It is also the dimension that is subject to variation under the influence of contextual variables, most notably emotions. The passage of time is essentially about the relationship between time, the self, and the world. In this sense, it may not be surprising that it is a recurrent theme in poetry and one that exemplifies the phenomenology of time at its best.

50

ANNA PIATA

3.4.3 Duration Duration is the last aspect in the conceptual framework of time outlined by Tho¨nes and Stocker (2019). It is defined as temporal extension, which relates to how much time has elapsed since the onset of an event until its completion, with the event ranging from several hundreds of milliseconds to a few minutes. While it is closely associated to the passage of time (since they both rely on the perception of change), experimental evidence suggests that duration judgements and ratings of the passage of time are not correlated (Wearden 2015; Droit-Volet and Wearden 2016). As Wearden et al. (2014: 298) note, ‘[film] clips judged as more exciting almost always receive faster passage of time judgements than boring ones, although effects on duration judgements are unpredictable’. When one says that time flies, it is not clear whether the corresponding time estimate will be longer (because there is ‘more time’ in the event) or shorter (just like journey time is shortened when going faster). In sum, duration qualifies as a distinct temporal dimension in the framework of Tho¨nes and Stocker (2019). Generally, duration judgements can be either retrospective (when they involve events that have already happened and are retrieved from memory) or prospective (when the events to be judged are anticipated). In either case, time estimation exploits cognitive processes, namely memory for retrospective judgements and attention for prospective ones. Contextual changes, whether environmental or internal (emotional), may also be at work, affecting the estimation of an interval (see also Prosser, this volume). The first example to examine is an excerpt from the opening of John Keats’ nineteenth-century epic poem ‘Hyperion’, which recounts how Hyperion, one of the Titans, was overthrown by the Olympian Apollo, the god of sun, music and poetry. In this excerpt, Thea, Hyperion’s wife, finds Saturn (ruler of the Titans) alone and utterly dejected, and addresses him with the following words: (7) ‘Saturn, look up! – though wherefore, poor old King? I have no comfort for thee, no not one: I cannot say, “O wherefore sleepest thou?” For heaven is parted from thee, and the earth Knows thee not, thus afflicted, for a God; And ocean too, with all its solemn noise, Has from thy sceptre pass’d; and all the air Is emptied of thine hoary majesty.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

51

Thy thunder, conscious of the new command, Rumbles reluctant o’er our fallen house; And thy sharp lightning in unpractis’d hands Scorches and burns our once serene domain. O aching time! O moments big as years! All as ye pass swell out the monstrous truth, And press it so upon our weary griefs That unbelief has not a space to breathe. Saturn, sleep on: – O thoughtless, why did I Thus violate thy slumbrous solitude? Why should I ope thy melancholy eyes? Saturn, sleep on! while at thy feet I weep.’ John Keats, ‘Hyperion’ (my emphasis) In these lines, time manifests itself in the phrase ‘O aching time! O moments big as years!’. Although, strictly speaking, it is not metaphorical, the linguistic pattern used for this purpose is still entirely figurative, comprising simile (indicated with ‘as’) and hyperbole (in the form of exaggeration). It is commonplace that moments and years differ vastly in terms of the intervals they contain, with the former being greatly shorter than the latter. When moments are said to be ‘big as years’, the result is a construal of elongated duration, with the interval being felt as massively longer than it is in reality. It is worth noting that the modifier ‘big’ may also predicate magnitude and significance, alongside protracted duration. This linguistic pattern for expressing time is not widely found in everyday language except for a few formulaic expressions, e.g., ‘It feels like yesterday when I was a graduate student at Cambridge’ and ‘The moments that followed the earthquake were felt like centuries’. Nevertheless, as I have shown elsewhere (Piata 2018, 2019), it is one of the various figurative ways with which subjective time (whether compressed or protracted) is talked about in poetry. Interestingly, it does not hinge on the metaphors that pervade time conceptualization and, in linguistic terms, it comes up in different forms and expressions while it arises in emotionally loaded contexts. A case in point is also Keats’ poem. In the excerpt above it is not accidental that time is modified as ‘aching’. As noted previously, Saturn (who is the receiver of Thea’s words in the excerpt) is now fallen and replaced by the Olympian gods, hence awash with feelings of loneliness, sadness and even despair. In cognitive psychology, it is has been shown that time estimates can be distorted by emotions, and each emotion

52

ANNA PIATA

seems to affect temporal judgements in specific ways (see Droit-Volet 2013 for an overview). Sad feelings, in particular, are associated with a slowing down in time perception, which leads to longer time estimates (Droit-Volet 2014), as well as to a sense of time as passing more slowly (Wearden et al. 2014).¹⁴ This can explain also Thea’s construal of moments as years in Keats’ poem. Note that Thea seems to sympathize with his emotional distress, as suggested throughout the excerpt: ‘our weary griefs’, ‘thy slumbrous solitude’, and ‘thy melancholy eyes’. Duration is also evoked in a poem by Marion Strobel, titled ‘L’Envoi’, albeit in a way that clearly differs from the previous example. In Strobel’s poem, the focus is on moments that appear to figuratively extend (‘reach and touch’) to the hours. This results in a construal of a protracted time span that reminds us of James’ ‘specious present’ discussed in Section 3.4.1. The fact that this is said to happen in a gentle manner connotes a positively coloured context. This time expression is repeated three times in the poem, thus implicating that this is a recurrent event. But the protracted, yet bounded, duration of moments as hours is juxtaposed to another interval, which appears at the end of the poem and suggests unbounded time, i.e., infinite duration; time has no beginning and no end, says Strobel, and in doing so she alludes to the infinity of natural time, as suggested by a scene of flowers’ blooming ‘again’, and the poet herself collecting a bunch of fresh jonquils ‘every day’. This further emphasizes recurrence, while conveying a sense of hope, optimism and even consolation. In sum, in Strobel’s poem duration is two-fold, construed as both protracted and unbounded. The two layers are conjoined to yield a more elaborate conceptualization of duration, in which the present times are embedded in the circular and never ending time in nature. Finally, the fact that the poem is addressed to a friend (the reader?), as suggested in the last line, renders this an intersubjective experience, one that is (or should) be shared. The data examined in this section illustrate the assumption that time is ‘multifaceted, inconsistent, and variable’ (Lloyd and Arstila 2014: 658). Like temporal processing, duration in poetry is represented in ways that differ in important ways from the routine ways such phenomena are studied in psychological research. For instance, in the case of duration, research has been focused on retrospective vs prospective time judgements and the role played by time-specific mechanisms, such as the so-called internal clock, the level of ¹⁴ A similar correlation has been attested also in depression patients as a result of physiological differences in time-keeping mechanisms (see Gil and Droit-Volet 2009 and references therein). Recently, El Refaie (2019) has demonstrated how such a distorted perception of time is metaphorically represented in graphic memoirs of depression.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

53

information processing, contextual changes, etc. (see, e.g., Block 1982; Zakay and Block 2004). Instead, the examples discussed here suggest time intervals as they are felt in the present under the influence of contextual cues, whether affective or discursive. Finally, similarly to temporal processing, duration is significantly less pronounced in poetry—at least in the small corpus examined here. This is not entirely surprising; as already noted in the previous section, it is the passage of time that mostly lends itself to rumination, contemplation, and emotional involvement, all of which favour its prominence in poetry. There may be one additional factor that further motivates creativity in the expression of time passing vis-à-vis temporal order and duration, and this relates to language itself; the linguistic patterns used to talk about the passage of time, which are to a large extent metaphorical, are easily amenable to creative exploitation. Clearly, this is only a conjecture that awaits to be further explored in future research.

3.5 From embodied meaning to embodied temporality (and back) The conceptualization of time has been the object of extensive and thorough research within cognitive linguistics, serving as a classic case study for metaphor theory. This endeavour has been focused primarily on establishing the metaphors that underlie the conceptualization of time through cross-linguistic and experimental studies (for an overview see Núñez and Cooperrider 2013). It is more recently that cognitive linguists have addressed the conceptualization of time as felt in human experience (Fauconnier and Turner 2008; see also Piata 2019). Still, the aim has been to refine existing models of representing the metaphoricity of time. In this section, I wish to tackle some broader implications that follow from the data analysis above, albeit in an admittedly preliminary manner and in hopes that my attempt will initiate more discussion in the future. In linguistic terms, the poems examined here largely conform to the predictions of cognitive linguists, especially in the case of the passage of time. The Time-moving metaphor seems to underpin the creative construals of time passing that manifest themselves in the relevant poems through mechanisms such as elaboration, extension, composition, and questioning. Likewise, the division between the objective time of the clock and the subjective time of the human subject attests to the tenets of Conceptual integration theory, which

54

ANNA PIATA

assumes a more elaborate network of mental spaces and blends for the conceptual structure of time. Still, however, much is left unanswered as to how meaning is actually derived while processing these temporal construals, which clearly differ from the ones found in everyday language. In what follows, I wish to pursue the argument that meaning-making rests upon a process of ‘mental simulation’ as laid out in Bergen’s embodied semantics (2005, 2012). According to Bergen (2005), simulation is essential to understanding literal and figurative language processing. Based on previous action and perception, mental simulation involves our ability to imagine entities, actions, and sensations that we have not experienced before. Such perceptual, motor, and even affective content of past embodied experience is activated for understanding utterances, not only literal ones (as in, e.g., ‘to grasp a bottle’) but also, and perhaps more importantly, non-literal ones. For instance, understanding the metaphor ‘to grasp an idea’ is constrained by our embodied experience of grasping, even if this event is abstract and therefore physically impossible to perform (for a detailed analysis see Gibbs 2006). Simulation therefore enables meaning construction, including any inferences derived. This approach views meaning as embodied, in line with the assumption that our cognitive system at large operates on the basis of embodied experience, including mental processes such as perception, memory, mental imagery, problem-solving, concepts, and also language comprehension (Gibbs and Macedo 2010). Embodiment is assumed to be at the core of language understanding in the sense that meaning depends on previous embodied experience in the actual world, which the language user is able to retrieve in order to create a qualitatively similar mental experience and thus produce a meaningful interpretation.¹⁵ To return to the poems examined in Section 3.4, we can legitimately assume that they are also based on the internal recreation of bodily experience, related to physical motion (e.g., flowing), sensory perception (e.g., the sound of the tolling bell) but also to time itself (e.g., no beginning and end). For example, let us consider again the metaphor ‘Time stays, we go’. If understanding language that denotes action recruits the same cognitive resources needed to actually perform this action, then this time metaphor may activate the motor system in a way that reflects the meaning of stay and go (and the contrast between the two). Nevertheless, this cannot do full justice to the metaphor as a whole, let alone its paradoxical meaning, unless it also takes into account time (which ¹⁵ It is worth pointing out that in cognitive linguistics all meaning is considered to be grounded in, and motivated by, embodied experience. However, embodied semantics takes embodiment a step further as guiding all language processing through and through.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

55

here relates to the incessant flow of time in the world, perceived as invariant when existence ceases). If this idea is on the right track, then understanding time expressions (even more so in the case of figurative expressions like the ones analysed in the previous section) may also activate the embodied experience of time. Research in cognitive psychology has shown how temporality is largely embodied, with the physical body serving ‘as the field on which time works and as a frame for its representation’ (Lloyd and Arstila 2014: 661). Such a view is couched within the broader framework of grounded cognition, also known as embodied cognition, which assumes, as noted above, that cognition is situated in simulations, action, and bodily states. Likewise, it is suggested that time is grounded in mechanisms for sensory processing and the motor control of action.¹⁶ For instance, Droit-Volet and Gil (2009) argue that temporal judgements result from changes in bodily states, as well as from sensory-motor states experienced or reactivated during interaction with the environment. This also includes emotions, since facial expressions of emotion have been found to trigger the embodied simulation of the corresponding emotions and thus affect time judgements. In a similar vein, Wittmann (2014) points to the role of circadian physiological changes (related to light, temperature, etc.) in encoding the passage of time and the duration of external events. He also considers the notion of the ‘present moment’ (akin to James’ ‘specious present’) as formulating a gestalt of perceptual moments, or snapshots of experience. Not accidentally, he uses the term ‘embodied’ to describe his view of time, both as a continuous flow and as the phenomenal present. Finally, DroitVolet (2014: 495) presents evidence suggesting that depicted body postures of different kinds (running vs standing) may trigger a time judgement that is consistent with the simulation of the relevant posture. It thus transpires that the motor system is activated for time judgements in a way that resonates with the embodied processing of word meaning related to performing an action. Is it possible that the mental simulation of linguistic input activates also a simulation of temporal experience? This is a question that follows naturally from the preceding data analysis. At present, no conclusive answer can be given as to how embodied time may apply to language understanding, which aspects of it may be involved, etc. For instance, a verse in which the moments are said to reach and touch the hours in a gentle manner is likely to trigger the simulation ¹⁶ This discussion leaves out the variety of brain areas that have been proposed as correlates of temporality, although these also reflect embodied mechanisms for time perception (for an overview see Lloyd and Arstila 2014: 660). Needless to say, there is no consensus as to where and how temporal information is processed in the brain.

56

ANNA PIATA

of an elongated interval. Similarly, saying that ‘Time stays, we go’ may evoke the sensory-motor activation that correlates with the sense of us moving past a static entity (or observer). Still, it is less clear how expressions rich in imagery, like ‘Time is dissolved, it blows like a little dust…’, may impart the embodied experience of time in its own right, unless we take these verbs to hint at a rapid passage of time and thus induce a mental simulation of the relevant scene. In other words, the imagery that is evoked in these expressions is likely to trigger sensory experience, which, although not directly associated with time, may be nevertheless recruited, i.e., dust blowing and dissolving instantaneously. However tentative, such a suggestion may open a new window into the intricate relationship between embodiment, time, and language. As Wittmann (2014) contends, action, emotion, and time all meet in the body—for embodied semantics, this conjunction includes also meaning-making. How they may be conjoined is an issue open for further investigation.

3.6 Concluding remarks In his monograph The Poet as Phenomenologist, Luke Fischer (a poet himself ) asserts that ‘the art of poetry can address certain philosophical problems more adequately than philosophy itself ’ (2015: xi). Although I would not go thus far, in this chapter I have ventured into an account of phenomenological (‘subjective’) time as represented in poetry. The merits and the limitations of such an endeavour are yet to be assessed. The poems examined in this chapter illustrate different aspects of how time is subjectively perceived, experienced, and contemplated. It would not be an exaggeration to say, in line with Arstila and Lloyd (2014b: 319), that ‘subjective time could well be replaced by subjective times’ (emphasis in the original). The perceiving subject is in the limelight, trying to weave together the order of events and to put into words how time passing is felt or how much time has elapsed. The result is akin to what Rovelli (2018: 218) calls a ‘naïve image’ of time. Such a view, as Rovelli himself admits, may be suitable for everyday life but not for understanding the world, as the temporal structure of the universe differs from our perception. From the standpoint of analytic philosophy, Prosser contends that ‘[u]nless some satisfactory response can be given we must take it that whatever the subjective nature of experience, it does not constitute an awareness of time passing’ (2013b: 162). Then, what insights can be gained from studying the phenomenal character of human time through the lens of poetry? In tackling this question, I will not be concerned with the

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

57

opportunities and challenges for linguistics, which I tentatively addressed in the previous section. By focusing on poetry and the construals of subjective time found therein, the present work adds, I argue, a novel ecological account to the study of the human concept of time. There is no doubt that the construals of time afforded in poetry are often idiosyncratic and at times even paradoxical (a case in point is T. S. Eliot’s poetry). But this is precisely the ecological layer that I believe poetry can import to the study of human time: the one that goes beyond the ordinary. The construals of human time found in poetry tap into the experience of time (private as it may be) and its representations (shared and established in the human mind) only to transform them in original, figurative ways. The need to study psychological time in realistic settings, outside the lab, has long been noted by Wearden et al. (2014), who thus implemented a web-based questionnaire to track down the participants’ felt passage of time. Likewise, in order to study subjective time Flaherty (1999) relied on autobiographical narratives of compressed and protracted time, both attesting to the phenomenological experience of time passing. I consider the approach followed in the present study as complementing this line of inquiry that seeks so-called ‘ecological realism’ (Lloyd and Arstila 2014: 662) in the study of human time unlike the experimental settings in which time was traditionally studied. Poetic musings can be seen as offering complementary (at times converging) evidence to this line of research. Notice that a view of time construed in subjective terms is by no means an exclusive property of the poetic domain. A robust body of research on time perception has pointed to a host of contextual factors that modulate the perception of time, including psychological states (e.g., arousal, boredom, sadness), specific conditions (such as drug consumption), or the type of activity in which the perceiving subject is involved, e.g. being alone or with friends. As I have argued elsewhere (Piata 2019), poetry lends itself to affective and other contextual triggers in the conceptualization of time, as also suggested in the present study; for instance, when meeting with a significant other and being subsumed by love (in Aiken’s poem) or when struggling with distress and solitude (the case of Hyperion in Keats’ poem). Finally, a construal of subjective time can only be a figurative, hyperbolic way to praise the uniqueness of poetry that overrides the boundaries of the poet’s life (i.e., Villiers de l’Isle-Adam). It is therefore motivated as a stylistic device rather than being grounded in a psychological state; this is yet another contextual trigger in time conceptualization, albeit of discursive nature.

58

ANNA PIATA

Clearly, this approach is not without limitations. As Lloyd and Arstila (2014: 657) note, there is ‘a sharp methodological divide’ between the phenomenology and the psychology of time. While the former illuminates temporality as embedded in the world and in overall experience (comprising also the remembrance of the past and the anticipation of the future), the latter represents the explicit temporality of quantized progression that arises against the backdrop of the clock. Can the two be somehow reconciled in methodological terms? This remains unclear and open to debate. Another open question of theoretical import relates to whether time is real or an illusion: ‘the nature of experience in fact provides no reason whatsoever for believing that time passes’ (Prosser 2013b: 315). Along these lines, the experience of the passage of time is considered an illusion, and the A-theory (that inheres in the experience of time passing) proves itself to be implausible (see, e.g., Prosser 2013c, i.a.). However, the present work does not favour B-theory either, since even when time is discussed in terms of order (namely, relations of anteriority and posteriority) the human subject is enmeshed and thus does not enable to disentangle the order of events in their own right (see also Farr, this volume). As already noted in the Introduction, the scope of this chapter subscribes to Jaszczolt’s (2020: 1874) view of the concept of time as a ‘human imprint’ of real time, with the two being ‘perfectly commensurate’ at the level of conceptual building blocks. However, in the context of the present work the study of human time is focused on aspects that do not pertain to temporal reference and is also infiltrated through the lens of embodied cognition. This is hoped to further encourage a dialogue between different analytic frameworks (and even disciplinary perspectives) in the study of human time and beyond. Finally, there is one more limitation to the present work, related to its methodological scope. In poetry (but also in ordinary language too), time can be expressed without being explicitly mentioned or designated; it can only be inferred on the basis of the linguistic input alone. Consider, for instance, the following lines by Emily Dickinson, which directly challenge our understanding of crumbling as instantaneous: ‘Crumbling is not an instant’s Act / A fundamental pause / Dilapidation’s processes / Are organized Decays’.¹⁷ In ¹⁷ The Poems of Emily Dickinson: Variorum Edition, edited by Ralph W. Franklin, Cambridge, Mass.: The Belknap Press of Harvard University Press, Copyright © 1998 by the President and Fellows of Harvard College. Copyright © 1951, 1955 by the President and Fellows of Harvard College. Copyright © renewed 1979, 1983 by the President and Fellows of Harvard College. Copyright © 1914, 1918, 1919, 1924, 1929, 1930, 1932, 1935, 1937, 1942 by Martha Dickinson Bianchi. Copyright © 1952, 1957, 1958, 1963, 1965 by Mary L. Hampson.

AN E XPLORATION INTO CONSTRUALS OF SUBJECTIVE TIME IN POETRY

59

the present study, the occurrence of the lexeme time as a criterion for the collection of the data analysed was necessitated for methodological reasons, as it would facilitate the trackability of time expressions in the corpus. But it had an additional asset too, since it would ensure that the linguistic expressions thus identified would evoke the concept of time, or, for that matter, its mental constructs of some sort. However, I believe that it would be interesting to also investigate constuals of time (whether poetic or not) in the absence of explicit mention to time. The issue of time recollection in memory is yet another aspect worthy of investigation. I leave these topics for future research. In conclusion, the present work has fostered an integrated approach to the understanding of time, bringing together linguistics, poetics, philosophy, phenomenology, and psychology. By drawing, quite originally, on evidence from poetry it has aimed to contribute to the existing, and prolific, crossdisciplinary inquiry into time and its representation(s).

4 The 2D past Graeme A. Forbes

Listen. Time passes. Things change. Old certainties fade, and tomorrow is a new day. We talk, frequently, of differences between past, present, and future. We talk of these as not merely different locations, but as changing.¹ What previously we anticipated as a future possibility becomes a present predicament, and then a past life-lesson, before finally being forgotten as ancient history. Suppose we take that talk as reflecting a mind-independent reality; how do we make sense of time passing? A number of issues converge here. I’m primarily interested in how we deal with cases where what we used to say is no longer true. More specifically, I’m interested in cases where we accept three things: a) we think that what’s true about the past is true because of how the past is b) we accept that how the past is isn’t the same as how the past was when it was present c) we think that there’s an important sense in which what makes things past is that they are now settled. Even to frame the puzzles I have in mind, quite a lot of scene-setting is required. In Sections 4.1–4.2 I will introduce some of the tense logic, and related discussions needed to do that scene setting. Then in Section 4.3 I will set up the specific puzzle: the ‘When am I?’ problem.² This is a puzzle about how we distinguish the objective present from the objective past if we treat them as merely temporal locations. In Forrest 2004 and Forbes 2016, a solution is put forward to the ‘When am I?’ problem; a distinction between states and events, on the one hand, and activities and processes on the other. In Section 4.4, I will explore the relation between activities/processes and verb aspect. My aim is to show that the ‘When am I?’ problem arises because it takes ¹ Leininger (2018) has an excellent discussion of the issues here. ² This problem has been put forward independently by Bourne (2002, 2006) and Braddon-Mitchell (2004)

Graeme A. Forbes, The 2D past. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Graeme A. Forbes (2023). DOI: 10.1093/oso/9780192896445.003.0004

THE 2D PAST

61

tense metaphysically seriously, but doesn’t take aspect as having metaphysical implications. The solution of invoking processes and activities takes aspect as seriously as tense. Having shown the relation between processes and verb aspect in solving the puzzle, we are still left with the more general problem. Activities and processes that used to be taking place no longer are, and we need to explain the relationship between completed processes (associated with perfective aspect) and ongoing processes (associated with progressive aspect). To do this, I will exploit a two-dimensional (2D) framework in Section 4.5.³ This 2D framework doesn’t just help us get a handle on the ‘When am I?’ problem, but also helps with some other cases where what used to be true no longer is. In Section 4.6 I will discuss cases where facts about the past seem to have changed retrospectively, and in Section 4.7 I will consider cases, building on Forbes and Wildman (2022), where meanings of words have changed over time. Both these cases involve changes to statements apparently about the past (changes which appear to have been ruled out by the proposed association between the past and the perfective). The 2D framework allows us to make sense of these without accepting that the past has intrinsically changed qua past. Thus verb aspect and a 2D approach to understanding how it changes can provide a successful model for understanding how what’s true can change over time, without accepting that the past is intrinsically unsettled.

4.1 Tense logic One is always hesitant to say that there is a standard position an in area of philosophy; the tense logic associated with A.N. Prior (1967) has certainly been influential, however. McArthur (1976) gives an overview of this approach to tense logic that reflects how it is often thought about in philosophy of time. I won’t quite follow McArthur completely, but I will summarize the approach, as it concerns us here. Those familiar with tense logic as it appears in contemporary metaphysics may wish to skip this section. We start with a series of statements. We can split these into what McArthur calls definite and indefinite statements; definite statements describe things at a particular time, or describe them atemporally, while indefinite statements don’t specify a time. I will make a slightly different distinction between tensed ³ See chapters 19 and 20 of Savage, Ebbers, and Martin (2019) for a discussion of recent twodimensional approached in philosophy.

62

GRAEME A . FORBES

and tenseless statements. I will call McArthur’s definite statements ‘tenseless’. So, (1) counts as a tenseless statement because it specifies a (relatively) definite time. (1)

Krakatoa erupts on 26 August 1883

Similarly, (2) is tenseless because it is atemporal; the question of when it is true simply doesn’t arise. (2)

2 is the only even prime.

While (1) is temporal, it is tenseless in the following sense: it does not specify whether the time stated is earlier, is later, or is contemporaneous with the time of assessment of the truth-value. The thought (one that I will push back on later) is that it doesn’t matter when it is assessed, because it will have the same truth-value. These statements rely on dates—they refer to particular times— but do not rely on tenses, which relate those times to a moment treated as present. So, one merely has to index statements to dates, and the need to deal with tense (in this philosophical sense of a relation to the present), at least in one’s metaphysics, subsides. If what we are in the business of doing is giving a precise description of the world, as it is mind-independently, it is tenseless descriptions such as (1) and (2) that we need. Or so the thought goes. Tensed statements, by contrast, are temporal statements that rely on tenses, though they may not specify when they are true. Consider (3)

Krakatoa erupted.

This differs from (1) in two respects. It doesn’t specify what date the explosion happened, but does specify the tense (insofar as it tells us it happened in the past). It does this through the verb aspect; by modifying the verb from ‘erupts’ to ‘erupted’ we establish that the explosion took place earlier than the time at which 3 is being assessed. We can mix dates and tenses together: (4)

Krakatoa erupted on 26 August 1883.

This is a tensed statement, because, although it contains a date, it also relates that event to a date later than it in time. McArthur would think of this as a definite statement, but for my purposes, it’s useful to group it with his indefinite statements, because the tense is more significant than specificity of time. The verb-aspect of ‘erupted’ establishes that this happened earlier than now. (4) is meant to be closely related to (1). (4) adds a piece of information to (1) making it more specific. It restricts the time of assessment to times later than the date

THE 2D PAST

63

given. (4) entails (3). But, on this approach, (4) is also meant to entail (1). This is something I will be objecting to later, and which will drive the argument of this chapter. Before objecting to it, however, we should spend a bit longer understanding why (4) is meant to entail (1), and introducing some of the relevant formal notation. Our statement containing dates and tenses, (4), can be split into a conjunction of a statement without tense—(1)—and a statement with tense (5)

26 August 1883 is earlier than now.

as follows: (6)

Krakatoa erupts on 26 August 1883 and 26 August 1883 is earlier than now.

The conjunction of (1) and (5) entails (1). (6), which is precisely the conjunction of (1) and (5), is meant to be logically equivalent to (4). As such, (4) will entail (1). The move here was to treat 6 as a modification of 1, by applying a tense to it. We can do this using a sentential operator. Let’s introduce three sentential operators. Pp, Np, and Fp, where p stands for a statement and ‘P’ is read ‘It was the case that’, ‘N’ is read ‘It is now the case that’ and ‘F’ is read ‘It will be the case that’. The presumption is that we start from tenseless sentences, or indefinite present tensed sentences with no embedded tense operators, and build up from there. So (3) is therefore logically equivalent to (7)

P(Krakatoa erupts)

And (6) is equivalent to (8)

P(Krakatoa erupts on 26 August 1883).

In the future direction, (9)

Krakatoa will erupt again.

Might be rendered as (10)

F(Krakatoa erupts) and P(Krakatoa erupts)

though one might desire some way to specify that it wasn’t the same ongoing eruption. Bells and whistles can be added to this. Metrics can be added to these operators to specify how long ago and for what intervals Krakatoa erupted, is erupting or will erupt. Two complications in particular are worth mentioning.

64

GRAEME A . FORBES

Firstly that these sentential operators can be iterated, and can be combined with other sentential operators, such as negation, possibility and necessity. So we can say things like (11)

PF(Krakatoa erupts)

to express the thought that Krakatoa was going to erupt, and (12)

◊¬F(Krakatoa erupts)

to express the thought that Krakatoa might not be going to erupt. Secondly, that N tends not to show up much. As McArthur (1976: 6) says, ‘The operator N is superfluous when simply prefixed to a statement… But N is non-superfluous when it occurs in the scope of tense operators.’ For example (13)

PFN (Krakatoa erupts)

which would mean something like that it was the case that Krakatoa would erupt now. (13) is logically equivalent to (14)

NPFN (Krakatoa erupts)

since any N operator is superfluous when not in the scope of another tense operator. The tensed/tenseless distinction becomes trickier to follow when one uses this logic. If we recall (1) this will look logically equivalent to a present tensed statement like: (15)

It is now the case that Krakatoa erupts on 26 August 1883.

One can specify that it is intended to be tenseless rather than tensed. (16)

It is now the case that Krakatoa (tenselessly) erupts on 26 August 1883.

But it becomes hard to see what ‘(tenselessly)’ means here. It is meant to convey, it seems, that it is the statement which is presently true, but not that the date mentioned is present. The problems that flow from mixing dates and tenses will play a significant role in the later argument (Section 4.3). For now we have three tense operators, operating over sentences, that can form interesting combinations with negation and modal operators.

4.2 More sophisticated tense logic, and other complications The picture presented so far is still rather simplistic. In this section, I will outline some complications. Some of these will only play a passing role in the argument, but it is useful to acknowledge these points of interaction with other debates. To begin with complicating the picture, we currently have three

THE 2D PAST

65

tense operators, but standard lists of tenses include rather more than three. Reichenbach (1947) for example gives a larger number: the past perfect the past the conditional the present perfect the present the future perfect the future

(e.g. I had done it) (e.g. I did it) (e.g. I would do it) (e.g. I have done it) (e.g. I do it) (e.g. I will have done it) (e.g. I will do it)

Reichenbach gets these extra tenses by having three elements: a speech act, and event, and a reference point (though, as Prior (1967) points out, one can regard the speech act as a particular reference point). Different tenses can be arrived at by considering reference points at times different from their events. Blackburn and Jørgensen (2016) view Prior and Reichenbach as allies, and propose a synthesis of their two frameworks. Blackburn and Jørgensen (2016) provide a hybrid logic that has propositional symbols, which stand for statements as before (e.g. p, q, r), and also has a nominal (e.g. i, j, k) that ‘names’ a time they are true at. With the propositional symbols and nominal as the atoms, we can then add the operators we’ve introduced already, along with conjunction (∧).⁴ Blackburn and Jørgensen then provide a set of tenses as follows: Tense

Examples

Hybrid Logic

Pluperfect Past Future-in-the-past Perfect Present Prospective Future perfect Future Future-in-the-future

Arthur had run Arthur ran Arthur would run Arthur has run Arthur runs Arthur is going to run Arthur will have run Arthur will run (Latin: abiturus ero)

P(i ∧Pp) P(i ∧p) P(i ∧Fp) Pp p Fp F(i ∧Pp) F(i ∧p) F(i ∧Fp)⁵

This hybrid approach takes these extra tenses to be gained by adding in a temporal location (named by the nominal). I won’t engage extensively with ⁴ They don’t, as a matter of fact, use the N operator in their language, but I don’t think anything turns on this point. ⁵ Adapted from Blackburn and Jørgensen (2016: 3682).

66

GRAEME A . FORBES

this view, but provide it as illustration of attempts to make tense logic more sophisticated than the Priorian picture with which we began. Let’s focus for a moment on the difference between the past and the perfect. If we try to translate the symbols back into English, using ‘Krakatoa erupts’ once more as our proposition, we get: (17)

‘It was the case that (it is timei and Krakatoa erupts)’

and (18)

‘It was the case that Krakatoa erupts.’

respectively. The claims on offer are that 17 is logically equivalent to 3, and 18 is logically equivalent to (19)

‘Krakatoa has erupted.’

These claims, viz. 3 ≡ 17, 18 ≡ 19 are going to be significant later too. In Section 4.3 we shall see that we generate some philosophical puzzles when we mix dates and tenses. For now, I want simply to agree with Comrie (1985:78): ‘however perfect differs from past, it is not in terms of time location’. It is Comrie’s thought here that leaves me unsatisfied with Blackburn and Jørgensen (2016). We store up problems for ourselves if we think of tense as merely playing with temporal locations, if we additionally hold the view that there is a metaphysically significant change that constitutes the passage of time. The issues here are somewhat delicate, however, as there are various different purposes at work. Prior, Reichenbach, and Blackburn and Jørgensen were all logicians; Comrie is a linguist; and I am a metaphysician. Of course Reichenbach and Prior both had metaphysical interests too, but different things are at stake if we are just interested in developing an adequately expressive tense logic, describing distinctions made in natural language, or describing the nature of the reality with the best tools available. My worry here is that when one has a hammer, everything looks like a nail. When one has tense logic, one can think it is the only tool available to explain how we talk about time. It’s clear that the basic Priorian set-up of Section 4.1 can be made to do more by employing hybrid logic, but I fear it might not be best suited for dealing with some particular concerns that come out of a specific debate in metaphysics. On top of complications to Priorian logic, there are also complications around what it is the propositions are about, not merely of when they are true. Whether we are describing events, states, processes/activities, achievements or

THE 2D PAST

67

accomplishments also affects things (see Vendler 1957; Moens and Steedman 1988; Smith 1997; Truswell 2011). States describe things at a particular time (or for a duration where there is no relevant change). Events are occurrences, and can either be very brief (e.g. lightning flashes) or very long (the past can be considered as a very long event). As Steward (2013: 805) argues, we can contrast events, understood as collections of temporal parts à la Quine, (1953[1950]), with processes understood as having a certain duration, but only accidentally. Whereas events are (on this view) just the durations they are, processes are the sorts of things that could have gone on longer, or in a different way. This contrast between events and processes is important. Steward (2015:114) says ‘A process, unlike an event, is capable of change. It can have a property at one time, but a different property at another.’ Processes are changes, but they are changes which can themselves change. Activities, for our purposes, can be thought of as processes carried out by things. Accomplishments and achievements are processes that end in a ‘culmination’ (Truswell, 2011: 58). So getting out of bed is an accomplishment, because there is a process of about half an hour involving checking social media, sitting up, checking e-mails etc. that precedes the culmination: not being in bed. Following Comrie (1976), I will use the term ‘situations’ as neutral between these ontological categories. Whether we are dealing with processes or states makes a big difference to whether verb aspect is relevant. Verb aspect, most crucially for our purposes, concerns ‘the progressive’; whether something is ongoing or completed. So aspect isn’t, like tense, an issue strictly of temporal location, but of whether something is going on in time. The Quinean view that we can treat everything, metaphysically speaking, as events obscures the distinctions between objects, processes, states, events, and so on. Though I wouldn’t attribute this to Quine specifically, the lack of engagement with verb aspect by metaphysicians in the Quinean tradition might be explainable by the fact that everything they wanted to say could be done by appeal to considerations of temporal location. If one assumes that everything is going on at any time at which it exists, for example, the progressive and temporal location go hand in glove. Verb aspect is a linguistic consideration, which relates closely to a particular tradition in philosophy: process philosophy. Process philosophers, such as Vendler (1957) tend to think of processes as being irreducibly dynamic. Though appeals to process are made by those who do not accept this (e.g. Salmon 1997; Dowe 2000), the significance of the category of process has been advanced by many, often using distinctions of verb aspect in Ancient Greek or English to support their position. (e.g. Vendler 1957; Stout 1997; Steward 2013,

68

GRAEME A . FORBES

2015). One of my main contentions will be that if we think tense is a metaphysically significant feature of the world that grammar can acknowledge so too is verb aspect. Both are required to make sense of a dynamic view of time. The considerations of linguists get mixed up with those of metaphysicians here, and it is important to keep in view the various ways they can interact. Quine’s view that everything can be understood in terms of regions of filled spacetime by no means represents the contemporary understanding of events in linguistics. Partee (1984), Moens and Steedman (1988), Dowty (1986), Kamp (1979; 2013 [1999]), and Blackburn and Jørgensen (2016), to name just a few, all give accounts of various linguistic phenomena that combine tense and aspect. But from a metaphysical point of view, they are all taking events to be metaphysically basic. Everything they say is just a matter of indexing utterances to locations in regions of spacetime. None are (explicitly) committed to the dynamism that the tradition of process philosophy stemming from Whitehead (1929), through Vendler (1957) into contemporary discussions. It is this metaphysical commitment to dynamism, and a metaphysical role for aspect that I am interested in combining. My contention will be that we either need to set both tense and aspect to work in our metaphysics, or neither, but problems arise when we have one without the other.⁶ What do I mean by ‘dynamic’? A third complication to the Priorian picture, in addition to refinements to tense logic, and the role of verb aspect, is change. When one is dealing with change in logic, the central issue is whether one is a temporalist or eternalist about propositions (cf. Richard 1981). An eternalist, who very much represents the orthodoxy in metaphysics, is one who thinks that propositions cannot change truth-value. This view dates back to Frege (1956) who thought that propositions have timeless semantic contents that refer to the true or the false. Being timeless, they cannot change reference. Temporalists, by contrast, accept that one and the same proposition can change its truth-value. This doesn’t stop them thinking that there are abstract semantic contents that we grasp when we understand something, it merely stops them thinking that those contents are timeless. Pretty much everyone accepts that it can change from raining to not raining, they just disagree about what the semantic content is when we say it is raining and when we say it is not raining. The use of dates, here, is illustrative. If we cast our minds back to (1), (Krakatoa erupts on 26 August 1883), we’ll note that it has a date. The eternalist believes there is no need for such a proposition to (per impossible) change its truth-value. ⁶ This isn’t a claim that natural languages need to have both tense and aspect.

THE 2D PAST

69

For sure, the utterances that people assert on the Clapham omnibus might be tensed, and cannot be expressed without appeal to a reference point they take to be present. Once we can fill in the detail to that reference point we can generate eternal propositions. Temporalists think that even if you fill in any such details, such propositions are liable to change in truth value. That is to say that tenses will be ineliminable, not just from ordinary language, but from the optimal third-personal objective theory that we—or perhaps a more objective theorizer like God—might use to describe or theorize the world. Are changes to be described as mere variations across a temporal dimension, or are they changes not reducible to mere variation? Dynamic views are committed to the claim that the poker being hot on Monday and cold on Tuesday is different from the Poker being hot at one end and cold at the other, because the latter is a variation across space, and the other is a change not reducible to mere temporal variation. This commits them to temporalism: that the proposition ‘the poker is hot’ changes truth value. What this amounts to, in logical terms, is that our best theories are going to have mixtures of dates and tenses in them, as we saw in (6), rather than reducing the tenses to dates and perspectives from which things are assessed. I’m going to argue, in Section 4.4, that defenders of dynamic views will need a role for verb aspect too. It’s a controversial issue whether dynamism (and, a fortiori, temporalism) is true. I happen to think it is. The problems I’ll be dealing with rely on dynamism and temporalism to get going. So let’s just assume them, for the sake of argument. The final complication I will bring in is ontic indeterminacy. Perhaps some times are ontically indeterminate. That is to say propositions may have indeterminate truth-value, not because there is unclarity in the proposition, or in the knowers, but due to indeterminacy in the world. In particular, propositions about future contingencies may be ontically indeterminate (Barnes and Cameron 2009, 2011; Briggs and Forbes 2012, 2019). For example, (20)

F(Kilimanjaro erupts)

is plausibly ontically indeterminate. There may, for all I know, simply be no fact about whether the dormant volcano will erupt. This isn’t necessarily to deny bivalence. There may be only two truth-values, true and false, it is just metaphysically indeterminate, on this view, which (20) has. We’ll call statements like (20) ‘future contingents’. Though you could have a view that combined metaphysical indeterminacy with eternalism about propositions—MacFarlane (2003) is at least in the spirit of doing so—the combination of temporalism about propositions and future contingents gives you a picture on which propositions can change from being ontically indeterminate

70

GRAEME A . FORBES

to determinate. This might be the case for (20) were some seismic event to make it inevitable that Kilimanjaro will erupt.⁷ We have, then, in addition to a basic tense logic, a number of complications to bear in mind. In summary; tense-logic can get more complicated, different ontological categories can give rise to complications involving verbaspect, and also complications involving whether propositions can change truth-value. Finally, the changes in truth-value may be from indeterminate truth-value to determinate truth-value.

4.3 The ‘When am I?’ problem That was a substantial amount of background, much of which is merely hinting at topics that themselves are at the centre of their own debates. In this section I’ll move more squarely into my own home turf; metaphysics. Imagine we have developed a theory such that • The theory is ontologically committed to non-present times • The theory is indispensably committed to tenses • The truth-values of propositions can change Such theories face a challenge: The ‘When am I?’ problem (Bourne 2002, 2006; Braddon-Mitchell 2004). Imagine Caesar storming Bigbury Camp in 54BC. Suppose, in an unusually reflective moment during the battle, he thinks ‘I am at the unique present moment’. His grounds for thinking this, it seems, are just as good as your grounds for thinking you are present. No claim has yet been made to the effect that Caesar differs in any way from you or me, besides being located at a different time. The past, for all we have said, is just a matter of being located at a time earlier than now. Was not our sentential operator—P—simply a method of restricting the scope of our claims to times earlier than now? (3) ≡ (7), after all. The worry then is that if we accept Caesar differs from you in temporal location, but not in sentience, any claim to knowledge of your own presentness is now suspect. For you have no kind of evidence that Caesar can’t in principle have. But what could you be more certain of your knowledge of, than that you are living in the present? If there is a unique present moment—a changing perspective from which our indispensable tenses are to be objectively ⁷ In Section 4.6 I will be considering cases where this picture is complicated, since the past can be intrinsically determinate, but extrinsically indeterminate.

THE 2D PAST

71

determined—surely you must be at it. But given your parity with Caesar, the chances are you are in the distant past. After all, there are so many more past times than there are present ones.⁸ Purely as a matter of statistics, we should expect you to be long dead. This argument, in Bourne’s hands at least, is meant to force you onto a dilemma: either claim only a single time exists, at which everything is located, or else give up on tenses in our best theory of the Universe (i.e. as required by our metaphysics).⁹ The former horn is Presentism, the latter a rejection of the passage of time as anything other than a temporal analogue of variation across space. This problem has generated various responses, and counter responses (e.g. Forrest 2004; Forbes 2016; Miller 2018; Perovic´ 2019). But a proposed solution is that we deny the parity between Caesar and you, without thereby denying that there is such a person in our theory as Caesar, nor that he is located in 54BC. What we do is deny that (3)≡(7). (3), rather is logically equivalent to (21)

P(Krakatoa erupted)

This move might need some explanation. Our sentential operator, P, still means exactly what it did before. It restricts the scope of our claims to times earlier than now. The difference is that we no longer accept that a past eruption is an erupting-in-the-past. Caesar, by being in the past, is no longer storming. It was the case that he stormed. This move dissolves the problem. For Caesar only has parity with you, if Caesar can think he is present. There is no problem if Caesar thought he was present. He was present. But he is not present. There is no time such that he is present at it. There is a time such that, when it was present, Caesar was thinking at it. We have, at last, all the elements we need. We have tense logic, but we also have the idea that, if you are a temporalist about propositions, and accept that non-present times exist and tense is ineliminable from our best theory of the world, using tense logic to locate things in time isn’t going to be enough. As I will go on to argue, we have some other tools to draw upon that might help us: aspect. Metaphysicians can learn from linguists to have a role of aspect in our best theory of the world, as well as in our analysis of languages. ⁸ I am using ‘times’ as shorthand here. You might think of times as instants, or as partless minimal durations, or as intervals (as is more common in linguistics). Regardless of your account of what makes something a time, there are more past ones than present ones. ⁹ Indeed when I talk about the role of tense in our ‘best theory of the world’, I mean a commitment to dynamism, rather than to an account of how certain natural languages function.

72

GRAEME A . FORBES

4.4 Processes and verb aspect Verb aspect marks, among other things, the progressive. The difference between (7) and (21), I claim, is to be explained using the progressive. The eruption of Krakatoa has long since finished. This finality is not merely a feature of its temporal location, but a consequence of a metaphysical fact: that things (wholly) in the past are no longer in progress. In so arguing I build on previous work by Forrest (2004) and Forbes (2016). The ‘When am I?’ problem, I claim, can be avoided if aspect is metaphysically significant, in addition to tense being metaphysically significant. That is, whether or not a process is ongoing or completed makes a metaphysical difference. Bernard Comrie (1976: 3) says: The difference in French between il lisait and il lut, or in English between he was reading and he read, is not one of tense, since in both cases we have absolute past tense. It is in this sense that we speak of aspect as being distinct from tense, and insist on such oppositions as that between the perfective and imperfective as aspectual, even where the grammatical terminology of individual languages has a tradition of referring to them as tenses.

My claim is that aspect must be treated as metaphysically seriously as tense. If you think tense is reducible to dates and embedded perspective, then likewise aspect should be. If you think tense is ineliminable from our best theory of the world, then likewise aspect should be. What I take to be the novel thought here is that the metaphysical role of aspect is giving us the (objective) change of context as time passes, rather than a mere change of perspective. In this section, then, I will introduce some important aspectual distinctions, and explain what I mean by taking aspect metaphysically seriously. We need, centrally, the distinction between the perfective and imperfective. This distinction, as Comrie (1976) remarks, applies to situations and not to verbs. One can add information to a verb to alter the aspect of a situation described. A situation is treated as perfective if it is treated as ‘blob-like’; that is we treat it as lacking distinct parts. Imperfectivity comes in three forms: the habitual, the progressive, and the stative. The habitual is for habits or dispositions (and consequently has a modal dimension), the progressive is for the ongoing, and the stative is for events treated as ‘punctual’ or point-like. The habitual and the progressive are compatible; one can do something repeatedly, and be doing it over a certain period of time. By contrast, the progressive and

THE 2D PAST

73

the stative are incompatible. One can either treat something as going on over time, or as point-like, but not both. We should be cautious about my claim that aspect is to be taken metaphysically seriously. I don’t mean that we should read a comprehensive metaphysics of aspect into ordinary uses of English. We shouldn’t just look out for certain verb forms, for example. There are many verbs in English that vary, on different occasions of use, between stative and non-stative (Comrie 1976: 36). Whether some certain situations are described with stative or non-stative sentences may also vary, for example, depending on contextual shifts which might mean that we re-describe a situation in different terms. An event might be treated as point like when considering one time-scale (e.g. a lightning flash) but may be of some duration when we consider a different time-scale, (e.g. when we look at slow-motion footage of it). We would not get a consistent metaphysics out of reading aspect directly from how we talk. But we can take seriously the idea that, in addition to tense, there is a question of whether a situation is progressive. Indeed, there is one notable way in which we can relate aspect to tense: the present tense is essentially imperfective (see Comrie 1976: 66). This idea that there is a link between the present and the imperfective seems promising to help with the ‘When am I?’ problem. The ‘When Am I?’ problem arises from thinking of the past tense as merely offering a restriction on the temporal location of a situation. But aspect adds something else. Not only are past situations located earlier than now, but they are no longer ongoing. No processes or activities are taking place at times located in the past. They once were (when those times were present), of course, but they are no longer. There has been a change in what’s going on. This response to the ‘When Am I?’ problem is that offered by Forrest (2004) and Forbes (2016). Processes and activities are restricted to the present, on this view, and the past tense does two things: it suggests a temporal location earlier than now, and asks us to consider it as past. I will discuss this further in Section 4.5. Linguistic aspect introduces a subjective point of view the objectively past can adopt different aspectual guises. But realism about tense introduces an objective present, and consequently, realism about aspect claims that it is objectively the case that only that which is objectively present is objectively imperfective. The relationship between processes and activities and aspect is less than straightforward. One can consider processes and activities as telic or atelic. That is, one can consider them as having a culmination or not. So, a walk to the shops culminated in an achievement: arriving at the shops. Simply

74

GRAEME A . FORBES

walking, however, need have no particular culmination. The former has a telos (the shops) while the latter doesn’t. This means that atelic processes have an interesting feature: the progressive entails a perfective. So if it is true you are walking, it is true you have walked. Not so for telic processes. If it is true that you are walking to the shops, it is not thereby true you have walked to the shops. Indeed, while you may have walked to the shops before, you haven’t accomplished some walk to the shops unless you are no longer walking to the shops (but have arrived). For telic processes, then, the progressive and the perfective are incompatible, but for atelic processes the progressive entails the perfective. We have a picture here in which ongoing things generate completed things (or at least things that can be described with the perfective). But the ongoing things in question always have duration (i.e. they are not ‘punctual’ and they always have parts (i.e. they are not blob-like). For something to be ongoing, some parts have to be completed, but there must also be potential for going on further; for more parts. Indeed, the prospective aspect can be used to indicate an anticipated achievement: (22)

Bill is going to throw himself off a cliff.¹⁰

This can be true, even if it is not certain whether Bill will succeed (so is compatible with ontic indeterminacy). That might be because the prospective often expresses an intention, rather than a prediction.¹¹ But equally it might express a present tendency or trajectory, which could be interrupted. As I argue in Forbes (2016), we can understand this in modal terms: the future is potential, the past actual, and the present is not a moment, but a situation that is ongoing, with some parts that are actual, and some potential for more to come. Aspect ties time and modality together. How does this fare as a response to the ‘When Am I?’ problem? There are a few things to note that create some challenges. One is that the relation between aspect and the ‘When Am I?’ problem seems tenuous. In Section 4.3, we said the solution to the problem was to deny that (3)≡(7) and affirm that (3)≡ (21). (3) and (21) are perfective and (7) is stative. The progressive doesn’t come into it. So why think that claims about processes and activities (characterized by the progressive aspect) are what’s doing the work here?

¹⁰ Example from Comrie (1976: 65). ¹¹ Consider with Anscombe’s example ‘I’m going to be sick’ which could be a prediction or an expression of intention (1963: §2).

THE 2D PAST

75

The problem, as I initially outlined it, was treating tense as if it merely shifted temporal location. My claim was that treating temporal location as metaphysically significant (for example by claiming there is a uniquely present time, but there are also non-present times) requires us to treat aspect as metaphysically significant (e.g. by treating some situations as ongoing and others not). But, as I have tried to illustrate, the situation is rather complex, with many different moving parts. Treating the difference between (3) and (7) as merely one of temporal location is a failure to treat aspect metaphysically seriously. That’s because treating our tenseless statements such as (1) as stative, is already to elide the role of aspect. (1) seems to be in the narrative present, so is grammatically correct, but we don’t get the sense of an eruption as being an atelic process that lasts several days. What happens when we compare different descriptions of the eruption: (23)

Krakatoa was erupting on 26 August 1883.

(24)

P (Krakatoa is erupting on 26 August 1883).

If (23)≡ (24), then there is, located at 26 August 1883, an erupting in progress. But that is exactly what is being denied in the solution to the ‘When Am I?’ problem! If they are not equivalent, our past tense operator seems to be useless for accounting for the role of ‘was’ in (23). Tense, unlike aspect, is a matter of temporal location. Consider, however: (25)

∗It was the case that Krakatoa is erupting on 26 August 1883.

This is just an expansion of 24, but seems infelicitous. Rather we would say: (26)

It was the case that Krakatoa was erupting on 26 August 1883.

The P operator can’t be interpreted merely as restricting our attention to an earlier time, but indicating the pastness of that time. The past progressive indicated that something used to be in progress, not that it is in progress at some earlier time. That tense does both of these things (locates and indicates non-presentness) is, I claim, intimately related to the imperfectiveness that is characteristic of the present. For it is the imperfectiveness of the present, and the generation of completed situations that comprise the (wholly) past that explains the need for tense to do both things, and provides the solution to the ‘When Am I?’ problem. The intimacy of the relation between aspect and tense is perhaps illustrated by the existence of languages in which aspect does much of the work of tense. Yoruba lacks tense but has aspect: perfective refers to past, imperfective to present (Comrie 1976: 82).

76

GRAEME A . FORBES

A second thing to note is the claim that the past contains no activities and states doesn’t immediately solve the problem. After all, ‘live’, ‘believe’, and ‘know’ are each stative verbs, so there is no prima facie barrier on the grounds of aspect to think that the problem cannot be framed in terms of what Caesar knows, without appeal to any progressive verbs. My response to this worry is that these states depend on progressive situations. All sorts of processes have to be taking place for an organism to live, or for a thinker to believe or know. So, the absence of such activity going on in the past undermines the possibility of such states in the past.

4.5 The 2D framework This metaphysical difference between the ongoing present and the completed past can be marked using a technique borrowed from discussions of disagreements about the necessary a posteriori. This solution is to exploit a ‘two-dimensional (2D) framework’ (Jackson 1998; Stalnaker 2004; Chalmers 2006; Savage, Ebbers, and Martin 2019).¹² In the original example of the necessary a posteriori, we are to imagine a group of people on a different planet (‘twin earth’) who speak a very similar language to us English speakers (Twinglish), except that their word ‘water’ picks out a different chemical stuff from our word ‘water’ (see Putnam 1975). On Earth, the watery stuff we call ‘water’ is made of H2 O molecules (plus some impurities). On Twin Earth, the watery stuff is made of some other chemical substance (let’s call it XYZ). But the surface properties of these substances are pretty similar and each plays the role of being the watery stuff for the local residents. On the assumption that we can overcome the confusions this difference may give rise to, we can use a 2D framework to make sense of what’s going on, as in Matrix 1. Matrix 1 According to English According to Twinglish ‘Water is H2 O’ True False ‘Water is XYZ’ False True Effectively, we are here treating ‘water’ as having two meanings, one in English and one in Twinglish. But we needn’t do so. We could treat it as having one meaning (the one in our language), but accept that we can interpret the ¹² As a way of thinking about language, two dimensional approaches have been around for a while, and even applied to tenses (e.g. Vlach 1993), but I’m sticking with the route they made an impact in the field in which I primarily work because I know about it better.

THE 2D PAST

77

Twinglish speaker as meaning to convey something other than what we would mean by those words.¹³ The idea of different linguistic communities makes it very easy to think of the two uses or meanings as on a par. It seems odd to describe speakers of Twinglish as ‘wrong’. They are, at worst, wrong relative to a linguistic community. But it is not hard to imagine a situation where the two perspectives are not on a par. A teacher may be able to interpret what a student means, even while accepting that they are misunderstanding the facts or the meanings of words. The point of the example in Matrix 1 is that we could think it a necessary truth that water is H2 O, but still understand the denial of this. Indeed, imagine someone learning English and who doesn’t yet know that water is H2 O. We can understand how they treat the word ‘water’ differently from us, given their different background commitments. When we play the role of the student, we can come to update our understanding by realizing that the correct understanding is very different from what we had been saying. This framework allows us to contrast different understandings of the world without needing to remain neutral about them. ‘Water is H2 O’, let us stipulate, is true, and ‘Water is XYZ’ is false.¹⁴ But we can understand why, from their point of view, the speakers of Twinglish say ‘Water is XYZ’. So, although we can treat them as different language communities, in fact a 2D framework can be useful for understanding differences within a single language community. Dynamic non-presentist views can exploit a similarity between times considered as past/once present, and possible worlds considered as counterfactual/actual, respectively. This means that things once in progress need not be treated as being in progress somewhere in the past, but merely as having been in progress somewhere in the past in virtue of the fact that past times were necessarily once present. We can treat the utterances of past people as being like ours, but centred on a non-present time. This locution of ‘centring’ comes from discussions of possible worlds. Merely having a set of propositions that describe a certain situation doesn’t give you all the information you need. For you need to be able to locate yourself in that situation. Centring offers a means of doing that. Consider Stalnaker’s (1981) example of a person who fell asleep in a trunk who does not know when now is. They may know what happens at each time, but not what is happening.

¹³ Whereas Jackson (1998) and Chalmers (2006) offer a two-dimensional semantics, Stalnaker (2004) offers a meta-semantic re-interpretation. My approach is thus much Stalnakerian, in that I think we are making sense of how language was used, rather than claiming there are two meanings that rival one another. ¹⁴ Chang (2012) argues that, strictly speaking, water isn’t H2 O.

78

GRAEME A . FORBES

In some sense, centred worlds are just a combination of a description of a possible situation and a centre to that description.¹⁵ The 2D framework, as I pick it up from Stalnaker, comes as part of a project to give an account of semantics in terms of possible worlds. Talk of possible worlds has wide currency in contemporary metaphysics; it has become something of a lingua franca in discussions of time and modality. It is in that spirit as a lingua franca that I bring it in here. I have no commitments to thinking our best theory of semantics will turn out to be a form of possible worlds semantics, nor am I committed to the existence of possible worlds (concretely or abstractly) as fundamental constituents of reality. For me, appeal to possible worlds (and a 2D framework) is a way of modelling situations that allows for fruitful conversations with others in metaphysics. Consider Matrix 2. Here we see that ‘Krakatoa is erupting’ is true according to the world centred on 1883 and false according to the world centred on 2023 (the present at time of writing). Matrix 2 Centred on 1883 Centred on 2023 Krakatoa is erupting True False Krakatoa has erupted True True ‘Krakatoa has erupted’ is true in the world centred on 1883, because eruption is an atelic process; if Krakatoa is erupting it follows that Krakatoa has erupted. Had we used a telic process, this wouldn’t follow. It is also true centred on 2023, although Krakatoa is not erupting at any time. This is because 1883 is such that, were we to centre upon it, Krakatoa would be erupting. That is, the world centred on 2023 is such that 1883 at that world has Krakatoa counterfactually erupting, but no actual, ongoing eruption. For something to be past, it is not merely located at an earlier time, but also to be counterfactual in a particular way. The past eruption is not an ongoing eruption located in the past, but a past eruption that used to be present. There’s an obvious circularity here: to explain what it is to be past by saying it ‘used to be’ present hardly seems illuminating. But given that the alternative was framed merely in terms of location, adding the claim that the past is a particular type of counterfactual (one that used to be present), does add something to our understanding, even if it depends on our already understanding what pastness is to some extent. We have ruled out pastness being a mere matter of temporal location, but instead as involving an orderly change in what time

¹⁵ See Liao (2012) for a discussion of some of the complications associated with centred worlds.

THE 2D PAST

79

we’re centring on. It may well be that this change of time is not susceptible to analysis in non-temporal terms.¹⁶ By exploiting this analogy between the past and the counterfactual, we can see how what was true at 1883 when 1883 was present need not be true at 1883 considered as now past. The eruption of Krakatoa was progressive when it was present (just as counterfactual worlds are actual according to the people at them), but as things actually are, the eruption of Krakatoa is perfective; Krakatoa is not still erupting now. This difference, let me emphasize, is not a mere difference of perspective. It is not that we and denizens of the past say different things about the eruption of Krakatoa, because we have a difference of perspective on it, but that 1883 used to be present, and then there was a change in the Universe that involved a change of aspect with respect to Krakatoa’s eruption. Merely possible people, were they to be actual, would be convinced that they were actual. But merely possible they remain. If the past were present, things would be happening. But past it remains. It is not surprising that what’s true at a time can change depending on when we centre of context of assessment. We saw this in the approaches of Partee (1984), Moens and Steedman (1988), Dowty (1986), Kamp (1979; 2013 [1999]), and Blackburn and Jørgensen (2016). What I take to be the novel thought here is that the metaphysical role of aspect is giving us the (objective) change of context as time passes, rather than a mere change of perspective. In the next couple of sections I want to think about cases where what’s true at a time gets more complicated. In Section 4.6 I’ll consider cases when it seemed what happened in the past changed after the relevant events appear to be over, and in Section 4.7 I’ll consider cases where sentences seem to change from necessarily true to necessarily false (and vice versa). In both cases, this 2D framework will be useful in understanding what has changed.

4.6 Tricky cases 1 Sometimes we make claims about the world that are perfectly felicitous, but we later find ourselves forced to contradict. The change from my saying at t1 ‘I am standing’ and at t2 ‘I am not standing but I was standing’ is not normally thought to count as a contradiction of my previous statement, but as confirmation of it. In this section I’m considering cases where there does seem to be a contradiction, even where we account for the passage of time. ¹⁶ Broad (1959: 766) claims this, for example. As I said in Section 4.2, I am assuming Dynamism for the sake of argument, so we don’t need to settle this question here.

80

GRAEME A . FORBES

Consider: (27)

‘Armstrong won the Tour de France.’

(28)

‘Armstrong never won the Tour de France.’

(27), with the context of assessment centred on 1992, would have been false; centred on July 2012, would have been true; but is actually false. (28), centred on 1992, would have been true; centred on 2012, would have been false; and is actually true. In both cases, we have the same verb aspect; we are dealing with a past achievement. You might think that once it is the case that you have achieved something, the achievement lasts. This won’t hold for comparative claims. Queen Victoria is no longer the longest reigning British monarch, but she was the longest reigning monarch when she died in 1901. It would be odd, however, to say that she was never the longest reigning monarch. But I claim (28) is true; Armstrong never won the Tour de France. Recall that in Section 4.2, we discussed the complication of future indeterminacy. This complication comes up in (27) and (28). Both claims look to be claims about completed achievements in the past. (27), in fact, is also a disguised claim about the future. The relevant achievement (winning the Tour de France) isn’t a matter of simply being the first over the finish line, or being handed a yellow jersey, but is matter of having the title conferred by the organizers of the competition. As J.L. Austin (1962) pointed out, such conferrals need to happen in the right circumstances, by authorized people etc. In practice, this means that if the conferral is later undermined by the discovery that cheating occurred, it is as if the conferral never happened.¹⁷ Armstrong was found to have cheated, and so was retrospectively disqualified from the relevant competitions. Given that (28) is true (you can’t have won a competition you were disqualified from), how can it be the case that (27) used to be true? This seems to be a contradiction! If we treat it is a matter of temporal location alone, it is. Saying that (27) used to be true is equivalent to P(Armstrong won the Tour de France), on such a view, and P(Armstrong won the Tour de France) and ¬P(Armstrong won the Tour de France) are contradictories. But if we remember that (27)

¹⁷ You might not like this example, but there are plenty of others where apparent speech-acts are later ruled not to have been successful. Consider the attempted prorogation of the UK parliament in 2019. A speech act appeared to have taken place, but later, at the Supreme Court, Baroness Hale, in the oral remarks of the judgement in R (on the application of Miller) (Appellant) v. The Prime Minister (Respondent) (2019), said ‘when the Royal Commissioners walked into the House of Lords, it was as if they had walked in with a blank sheet of paper’ (13.30).

THE 2D PAST

81

is not true at any time, but rather there are times such that, when we centre on them, (27) is true at them. (27) is true treated as present, but not true treated as the past of this time, where the committee has enacted the various disqualifications. There is a wrinkle here. Even when we centre on July 2012, should (27) come out as true? After all, if it is dependent on how the future unfolds, according to Barnes and Cameron (2009, 2011) and Briggs and Forbes (2012, 2019), the truth-value should be indeterminate. Strictly speaking this is the correct response; it is not straightforwardly true that anyone has ever achieved anything while it is feasible that some committee could strip them of it later. But in many contexts such feasibilities are remote and not salient. When we assess truth-values we use contextually salient standards of precision and confidence. It is felicitous to claim Cambridge is flat when I compare it to the French Alps, even though it is not perfectly flat. It is felicitous to claim that I won’t win the lottery, even though I have a miniscule chance of finding a winning ticket stuck to my shoe. The Armstrong case shows us how claims about past achievements are vulnerable to revision, provided they depend at least in some small way on future events. Indeed, this concern for how our descriptions of the past must be revised is a commonplace of the philosophy of history. As Arthur Danto (1985: 340) puts it: In effect, so far as the future is open, the past is so as well; and insofar as we cannot tell what events will someday be seen as connected with the past, the past is always going to be differently described.

We can see then, that claims about the world may change; especially changes about the social world that depend on the future whims of committees. And we can see how a 2D framework can allow us to make sense of these changes in a way mere change of temporal location does not. But surely necessary truths are beyond the reach of the whims of committees? Alas no, as I shall argue in the next section.

4.7 Tricky cases 2 Logical relationships between things look as though they should be immune from change. If something is necessarily the case, it follows that it’s not possibly not the case. So merely waiting a few years shouldn’t affect it. Yet many of the necessary truths we know seem to be truths about relations between concepts,

82

GRAEME A . FORBES

and those are as vulnerable to the future whims of committees as anything else. Consider: (29)

‘Cheese is non-vegan.’

At one time this may have been considered an analytic truth. If you thought cheese was suitable for vegans, you had simply failed to understand what cheese is, or what vegans are (or both). Vegan cheese used to be impossible. Now, it seems, that vegan cheese is something you can make at home.¹⁸ We can of course claim that it was never analytically true, for example by being temporal externalists, such as Jackman (2005) and Collins (2006). Temporal externalists claim that the (current) meanings of words can be settled by future decisions. A temporal externalist may claim, then, that the vegan cheese was always a possibility, but people were previously unaware of this. It seems plausible, however, that the decisions could have been made that the stuff known as ‘vegan cheese’ could have been described in other terms, and not counted as a type of cheese at all. Indeed, in various jurisdictions the debate still rages.¹⁹ Denying that it used to be plausibly analytic that cheese was non-vegan, because there was a chance a future decision could be taken to change the use of the word ‘cheese’, has significant consequences for our understanding of meanings. If we allow that those future decisions are metaphysically indeterminate, as we did for decisions about whether to strip someone of a racing title, then what our words now mean is metaphysically indeterminate. While defenders of temporal externalism think of this as a feature, rather than a bug, opponents have been scathing. Keith Donnellan (1983: 103) describes temporal externalism as ‘an outrageously bizarre view of language—that the extensions of one’s terms may be determined by the psychological quirks of some people several centuries hence’. Jessica Brown (2000: 187) complains that it ‘fails to accord with our ordinary linguistic practice’, and that, ‘…we do not regard evidence about future practices as relevant to the meaning or truthvalue of current utterances and thoughts. More fundamentally, we do not defer to future linguistic practice for the correct explication of our concepts.’ In work with Nathan Wildman (2022), I defend the view of meanings as persisting through change. I accept Brown’s view that we can’t defer to future practices in the meanings of our terms now, but take the point that meanings of our terms do seem to depend in some way on decisions taken by scientists and ¹⁸ https://www.bbcgoodfood.com/howto/guide/how-make-vegan-cheese/. ¹⁹ One cannot market cheese as being vegan in the Republic of Ireland, for instance.

THE 2D PAST

83

others to whom we defer in classifying the world. The solution is to have meanings that change as different decisions get made. What words mean depends on how words have been used, and are being used (perhaps with deference to experts and perhaps on the structure of the world we describe with them, if semantic externalism is true).²⁰ We can, once more, use the two-dimensional approach to respond to such meaning change. We can compare centred-worlds; one centred on 1944 (when the word ‘vegan’ was coined), and one centred on 2023, when vegan cheese is available, as in matrix 3. Matrix 3 ‘Cheese is always non-vegan’ said in 1944 ‘Cheese is always non-vegan’ said in 2023

Centred on 1944 Centred on 2023 True False True

False

In this matrix, we can see that when we assess utterances in the past, provided we are using current meanings, we can correct past usages. We can say ‘they thought cheese had to be non-vegan, but they were wrong’. But, equally, we can understand why a speaker at the time would think that a future user of the term was wrong, because, as a matter of necessity (according the linguistic conventions they were working with) cheese is non-vegan. We could have this as a framework simply for comparing differences in language use at different times, i.e. treating different times as merely different locations. But these times are not on a par. One time gives the actual meanings of words, because it gives us the present meanings. The other gives us meanings that words used to have, but no longer do. Note that none of this involves accepting that the intrinsic nature of the past has changed. No time-travel or backwards causation has taken place. It’s just that how we classify the world has changed. This is no more alarming than the fact that the word vegan was coined in 1944, but vegan diets had existed long before then. The sense in which the past is changing here, is not a sense in which we need to worry about the ‘When Am I?’ problem. It’s not the case that things are intrinsically changing in the past, but merely that meanings of words and sentences inevitably involve an element extrinsic to the situation they refer to. While semantic externalists believe that the world contributes to the meaning of certain terms, they must also believe that meanings are fixed in part by ²⁰ Semantic Externalism, as defended by Putnam (1975) is the view that the meanings of our words depend in part on the way the world is (for example it’s chemical or biological structure), and not merely on how we think about it.

84

GRAEME A . FORBES

language use. Semantic internalists will appeal to nothing besides competent speakers’ understanding of a language. As language use changes, meanings, and so truth values, will also change.

4.8 Conclusion After some scene-setting, we encountered the ‘When Am I?’ problem. Given views that accept that the past exists, but is not objectively present, we need account of what distinguishes the past from the present. I have offered an account of verb-aspect as enabling us to distinguish between past and present activities or processes. Present activities or processes are ongoing (and so described with imperfective aspect), while past processes are complete (and so described with perfective aspect). This allows us to move beyond treating the passage of time as a mere change of temporal location or temporal perspective. Having been presented with this role for verb-aspect, we need to make sense of the change of aspect over time. The 2D framework was brought in to do this. The 2D framework can also explain the tricky cases involving cases in which the past seems open (because of a dependence on things still in the future) and with cases of meaning change over time. Though the 2D framework could be used with a view that treats change of time as a mere difference of location, it can also, as I have urged, allow us to treat change of time as a genuine change in the world. The 2D framework can do this, while preserving the claim that the past is intrinsically the same as when it was present and that the passage of time is a process of things being settled.

5 Counterfactuality as pragmatic inference in perspectival readings of Past Conditional utterances with modal verbs Evidence from French Louis de Saussure

5.1 Introduction In Romance languages, modal verbs devoir (‘must’) and pouvoir (‘can’, ‘may’) can combine with all tenses and moods. When they appear with the Past Conditional, under certain conditions that I will attempt to unveil, they give rise to counterfactual reading,¹ i.e. readings where it appears that the eventuality did not happen.² This chapter attempts to explain why—looking at the particular case of French—such readings occur. In a case like (1), for example, modal-Past Conditional (P) gives not-(P) (here Pierre was not surprised) quite straightforwardly: (1) Pierre aurait dû être surpris. Pierre must-PAST COND be surprised. A similar situation exists in English with ‘should’: (2) Pierre should have been surprised. The notion that Pierre wasn’t surprised seems quite obvious in such examples; at such a point that Larrivée and Patard (2019) call the (French) Past ¹ By reading, I mean, roughly, ‘interpretation’; the interpretation of an utterance is based on semantic evidence (what words and sentences mean by their own virtue of coding signification) but also, and perhaps more importantly, on the spontaneous reasonings—inferences—that are performed to achieve an enriched meaning that can be taken to correspond to what the Speaker intends to represent, pragmatically, by means of their utterance in the circumstances. ² I take here ‘counterfactuality’ in a strong sense, meaning that the eventuality was not the case (sometimes labelled ‘foreclosed counterfactuality’). Louis de Saussure, Counterfactuality as pragmatic inference in perspectival readings of Past Conditional utterances with modal verbs. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Louis de Saussure (2023). DOI: 10.1093/oso/9780192896445.003.0005

86

LOUIS DE SAUSSURE

Conditional a ‘marker of counterfactuality’. In what follows, I discuss the case of French, returning with some insights about the English case in the conclusion; obviously, French and English modals have significantly different behaviours, but there are clear pragmatic bridges, I suggest, across the semantic gaps. Notably, that must is not counterfactual itself, just as devoir is not either (see further down Section 5.2). Why does a counterfactual reading arise in (1)? Actually, neither the modal verb nor the Past Conditional in themselves seem to lead to a counterfactual reading by their own virtue, and therefore the causalities at play must be either compositional and semantic, or inferential and pragmatic, or a mixture of both. Since, as I develop further down, the counterfactual readings of modal utterances with the Past Conditional is not a stable effect across contexts, the key factor should be pragmatic, context-dependent; if this is correct, then counterfactual readings of such sequences are not ‘conventionalized’ meanings as Larrivée and Patard (2019) suggest but rather the result of a combination of linguistic and contextual features.³ And indeed, modal sentences with the Past Conditional are intrinsically ambiguous between a hypothetical reading (uncertain) and counterfactual (in the strong sense of negating the occurrence) readings in French. There are several ways to envisage pragmatic explanations to such kinds of phenomena. It looks reasonable, in particular, to assume that not-P arises in such cases as a pragmatic inference that provides a rational motivation to the mere fact that the Speaker mentions that ‘P ought to have occurred or to have been true in the past’. However, the inferential path leading to such a conclusion needs to be elucidated, since no counterfactual inference on this basis alone is clearly available: that P ought to have occurred in the past does not lead logically to the inference that P has not occurred (even quite the contrary). Therefore, there must be other factors at play. In this chapter, I suggest that not-P in these cases is not actually inferred as a conclusion (as would an implicature) but is taken for granted as a piece of background information that gives relevance to the speech act (a ‘discursive presupposition’ in the terms of de Saussure 2013a). It is an element of meaning which is contextually necessary to give relevance, or rational motivation to the speech act itself. These elements, just like presuppositions proper, are occasionally accommodated, and are indeed quite generally so when the speech act in the given circumstances would be incongruous without them. However, the ³ As Verhulst and Declerck (2011) suggest, although in a different perspective, to which I return in the conclusion to this chapter.

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

87

reasons why the counterfactual element of meaning arises, I suggest, have to do with some subjective perspective taking: the utterance represents the viewpoint of a subjective persona making a future prediction at some past moment in time, based on necessities or possibilities, which results in forms of expressive meanings. There are a number of linguistic bases for this inference, since it occurs only in certain semantic and morphological combinations. Modality, past, and the conditional mood all play a role in this puzzle. In the first section, I give ground to the assumption that only root, i.e. non epistemic, modal meanings (concrete possibility or necessity, not abstract possibility) give rise to counterfactual meanings, at least if restricting the notion of ‘epistemic modality’ to what Lyons (1977) calls ‘subjective-epistemic’ modality.⁴ In the following section I discuss the notion of a prediction made in the past and its association with the Past Conditional tense together with counterfactuality as a discursive sort of presupposition. Section 5.4 develops the hypothesis that the Past Conditional activates the representation of the eventuality as if from a particular viewpoint, a meaning I propose to call a ‘subjective perspectival meaning’ of the Past Conditional. In conclusion, I discuss briefly the widespread assumption that the Conditional mood might involve an implicit protasis and take a short look at how French and English modal verbs differ in this respect.

5.2 The issue of epistemic modality in the past Example (1), repeated below, involves a modal verb expressing necessity. As mentioned earlier, it conveys, in ordinary circumstances, the notion that Pierre has not been surprised: (1)

Pierre aurait dû être surpris. Pierre must-PAST COND be surprised. Pierre should have been surprised.

Here is another example, this time with a deontic necessity: (3)

Pierre aurait dû ranger sa chambre. Pierre must-PAST COND tidy up his room. Pierre should have tidied up his room.

From (3) one infers that Pierre’s did not tidy up his room. The naturality of the counterfactual reading of these utterances is reflected in the fact that (1’) ⁴ That is, a notion of possibility expressed as a subjective opinion, not as objectively verifiable.

88

LOUIS DE SAUSSURE

and (3’), where the counterfactual reading is verbalized (without redundancy), are natural whereas (1”) and (3”), where factuality if verbalized, are intuitively much less so (however they are not strictly speaking ungrammatical): (1’)

Pierre aurait dû être surpris, et je suis étonné qu’il ne l’ait pas été. Pierre should have been surprised, and I am astonished that he wasn’t.

(3’)

Pierre aurait dû ranger sa chambre, et je suis contrarié qu’il ne l’ait pas fait. Pierre should have tidied up his room, and I’m annoyed he didn’t.

(1”)

?

(3”)

?

Pierre aurait dû être surpris, et je suis content qu’il l’ait été. Pierre should have been surprised, and I’m happy he was. Pierre aurait dû ranger sa chambre et je suis content qu’il l’ait fait. Pierre should have tidied up his room and I’m happy he did it.

Note in passing that the oddity of (1”) and (3”) should not be taken as evidence that counterfactuality with devoir and the Past Conditional is semantic. It can perfectly be that it is pragmatically unexpected, in most context, to question P and at the same time to assert or presuppose P. Necessity in the past in general does not express counterfactuality; with other past tenses than the Past Conditional, it does indeed not. It appears here with a Composed Past and with an Imperfective Past: (4)

Pierre a dû ranger sa chambre. Pierre must have tidied up his room.

(5)

Pierre devait ranger sa chambre. Pierre must have been tidying up his room / Pierre had to tidy up his room.

(4) expresses either an epistemic modality which leaves doubt open and thus forbids a counterfactual (as well as a factual) reading, or a deontic one, which in turn leads to factuality: Pierre has indeed tidied up his room (as a result of his obligation to do so); (5) is either an assumption about what Pierre might have been doing—thus an epistemic modality—or a description of what he had to do at that time (typically hinting at the reasons for which he was not doing something else). The first step I will take to disentangle the various factors at play in my puzzle will be to observe that only non-epistemic readings of the modal verbs lead to counterfactual inferences with the Past Conditional, and they do so with both necessity and possibility modalities. But while (3) is clearly deontic, the status of (1) deserves some elaboration as it seems to have a flavour of epistemic

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

89

modality. Indeed, its structure, a modal verb, here ‘devoir’ (‘must’), that takes scope over a proposition with the verb ‘être’ (‘be’), is generally assumed to be typical of an epistemic modality (see in particular Kronning 1996). At the same time, if (1) is epistemic then it should be incompatible with both factual (Pierre has been surprised) and counterfactual (Pierre has not been surprised) readings, for the reason that epistemic ‘devoir’ (‘must’) expresses the best available explanation as a necessary inference on the basis of potentially incomplete or uncertain premises available. As a result, that inference necessarily leaves room for some degree of uncertainty, and uncertainty cannot go together with (counter-)factuality, of course. This is directly linked to the fact that the pieces of evidence available to the Speaker about her stipulations are those available at Speech time (regardless of the time of the modality itself ). I endorse here the classical view that epistemic attitudes hold at speech time, regardless of the time of the eventuality upon which they scope (Cinque 1999; Stowell 2004; Mari 2013; Gosselin 2010 inter alia).⁵ In other words, when an utterance manifests an epistemic attitude, typically with a modal verb like English may (epistemic possibility) or must (epistemic necessity), the attitude is communicated to hold at speech time (i.e. at the present moment), even when it is about a past or a future situation. Epistemic modality, in that sense, is indexical, i.e. bound to the temporal circumstances of the utterance.⁶ (1) is not, however, an epistemic necessity about a past situation; this would be rendered by (6) in French, (6′ ) in English, which means that a stipulation is taking place at speech time about the likeliness of a past situation, whereas (1) is obviously different: (6)

Pierre doit avoir été surpris.

(6′ ) Pierre must have been surprised. The same happens with the list of examples below: even though the temporal reference varies, they all express an epistemic stance at speech time; this is

⁵ Several authors in formal semantics have argued against this view, suggesting that modals can be epistemic with a past perspective (Condoravdi 2002; see in particular Rullmann and Matthewson 2018). My own approach shares similarities with these works; notably, I also refer to a notion of perspective (de Saussure 2013b). My view is however that representing a particular perspective is a form of represented thought, thus of echoic use of language. For what concerns me here, I hold, contrarily to these approaches, that when a Speaker displays an utterance with an epistemic modal associated with a non-present perspective, it cannot have an epistemic interpretation, unless a form of echoic reading: admittedly, anything ‘epistemic’ is bound to the mental activity of an individual who raises an assumption. ⁶ The canonical way to account for this is to assume a stable, language-independent, hierarchy of scope: modality, in particular, scopes over temporal indicators, so modality is higher in the hierarchy (Cinque 1999).

90

LOUIS DE SAUSSURE

particularly relevant with French (and Romance more generally) since modal verbs devoir (‘must’) and pouvoir (‘can’/ ‘may’) accept conjugation with all tenses and moods: (7)

Pierre a pu partir plus tôt que prévu. Pierre can- PAST leave earlier than expected. Pierre might have left earlier than expected.

(8)

Pierre a dû aller à la piscine. Pierre must-PAST go to the swimming pool. Pierre must have gone to the swimming pool.

(9)

Pierre pourra partir plus tôt que prévu. Pierre can-FUT leave earlier than planned. Pierre might leave earlier than planned.

(10)

Pierre devra être à la piscine. Pierre must-FUT be at the swimming pool. Pierre will probably be at the swimming pool.

This phenomenon is general: all the tenses are possible, with subtle variations of meaning while preserving the epistemic time at the deictic present in all cases. All combinations in (11) and (12) express presently holding epistemic attitudes of the Speaker about past situations: (11)

Pierre pouvait être / a pu être / peut avoir été surpris. Pierre can-IMPERFECTIVE PAST / can-COMPOSED PAST / can-PRESENT be surprised.

(12)

Pierre devait être / a dû être / doit avoir été surpris. Pierre must-IMPERFECTIVE PAST / must-COMPOSED PAST / must-PRESENT be surprised.

The only obvious exception to this general rule is reported thought or speech in its various forms, in particular Free Indirect Speech, and other deictically shifted or ‘perspectival’ (de Saussure 2013b) utterances (Narrative Present, etc.). The fact that counterfactuality emerges with (1) suggests clearly that despite appearances, (1) is not epistemic, unless it is a case of reported thought

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

91

or speech, or an otherwise deictically shifted utterance.⁷ There is however no indication, nor intuitive reason, that it should be so. It is thus reasonable to assume that (1) does not express epistemic necessity, but another type of necessity. As for the apparent epistemic ‘flavour’, arguably, it is not about epistemicity proper but about the expression of a subjective feeling of surprise occurring when realizing that an unexpected situation has occurred, past as it may be. That counterfactual readings of modal utterances at the Past Conditional emerge only with non-epistemic meanings appears clearly when looking at possibility modals. The French verb ‘pouvoir’ (‘can’/ ‘may’) is enlightening in this respect since it is compatible with all modal meanings of possibility; thus depending upon the selection of the relevant modal meaning in the circumstances, which in turn depends upon the context and the attribution of plausible meaning intentions to the Speaker, pouvoir leads, or not, to a counterfactual interpretation (see also Abusch 2012 on English): (13)

Pierre aurait pu ranger sa chambre. Pierre can-PAST COND tidy up his room. Pierre could/might have tidied up his room.

(14)

Pierre aurait pu venir, puisqu’il est guéri. Pierre can-PAST COND come since he’s recovered. Pierre could/might have come since he’s recovered

If these utterances are interpreted as dynamic possibilities, or as understatements for deontic obligations, they suggest that, respectively, Pierre did not tidy up his room, and that he did not come (counterfactual readings). However, if the interpretation is epistemic, for example if (14) is inserted in a conversation or a narration about the possible past whereabouts of Pierre, or if (13) is about what Pierre might have done at a certain point of time, a counterfactual meaning simply does not arise. This is expected, again, since an epistemic meaning of possibility implies that the situation’s truth is unknown, and that the utterance manifests a stipulation that takes place at the time of speech and resting upon the pieces of evidence currently available. As mentioned earlier, modal verbs with the verb être (‘be’) are expected to convey an epistemic meaning; looking at pouvoir (‘can’/ ’may’) with the Past Conditional, however, alternative interpretations appear easily available,

⁷ See also Rullmann and Matthewson (2018).

92

LOUIS DE SAUSSURE

which show how the meaning of this construction, at least, is contextdependent, and that the assumption that être with a modal verb bears an epistemic modality is too strong, in particular when the Past Conditional enters the scene. That combination rules out dynamic modality but opens at least three different interpretive options. Consider (15): (15)

Pierre aurait pu être à la piscine. Pierre can-PAST COND be at the swimming pool. Pierre could/might have been at the swimming pool.

(15) can read: i) typically, as a statement about where Pierre could have been according to past evidence and conveys the counterfactual meaning. For example, (15) with the meaning i) can be uttered by a Speaker who explains the reasons for which she did something, for example why she went to the swimming pool that day at that hour. With this interpretation comes the counterfactual meaning that Pierre was not at the swimming pool. This is not in my view an epistemic interpretation, since the truth about Pierre not being at the swimming pool is known. However, it is bound to a former mental activity of speculating whether Pierre is at the swimming pool or not.⁸ ii) as an epistemic statement, meaning that the swimming pool is where Pierre might have been, according to evidence currently available. This interpretation is about what is currently stipulated regarding Pierre’s possible past whereabouts. It is epistemic in the clearest sense: it presents a possibility as a thought, i.e. as a personal stipulation of the Speaker according to what she can tell. This interpretation, which cannot raise a counterfactual reading, corresponds to what Papafragou (2006) labels an ‘externally inscrutable’ assumption, i.e. an assumption that is presented as a personal stipulation. iii) as a deontic understatement, expressing disappointment that Pierre did not go to the swimming pool, despite the fact that this was what was expected of him in the social context. In that case, a counterfactual reading emerges immediately: Pierre did not come to the swimming pool. Meaning i), which is the one that interests me most as it is of the same general kind as (1), may look like an epistemic modality. However, it does not express ⁸ A parallel can be drawn here again with the notion of perspective in Condoravdi (2002) but with the difference that the meaning of such utterances in my view is not epistemic, and that anything perspectival in my view is related to represented thought; I will get back to this further down.

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

93

a belief as such, but is a description either of objective possibilities holding at the past moment (which are not epistemic in the strong sense of ‘subjective epistemic’) or of the subjective thinking of an individual at the past moment, but which are not those of the Speaker. In any case, the utterance does not present itself as expressing the personal stipulations of the Speaker at speech time, but an assumption which is not bound to the subjective ventures of the Speaker. This type of modal meaning is accounted for in various ways in the literature and is uneasy to categorize. The time of the evaluation of the possibility is not taking place at speech time (contrary to interpretation ii)), since it is about a possibility that existed in the past as an epistemic assumption holding at that time. Moreover, pragmatically speaking, the possibility cannot be asserted at speech time (it is known that it is false), so that the modality cannot scope over tense. As a result of these two elements, it does not fit with epistemic modality in the narrow, usual sense of the expression of a personal belief as such. Clearly, it is not about a capacity, internal or external, provided to the subject, so it is not a case of root, in particular dynamic, modality either. The remaining option is that it is alethic (Kronning 2001a), i.e. expressing what can be or happen, or what ought to be or happen, according to the conditions of the world and not to the subjective reasoning of the Speaker. Alethic modality is often seen as restricted to material and logical sources but I follow Kronning (2001a) in extending the category to ‘alethic lato sensu’ modality, i.e. in a relaxed sense, which involves other, social, sources (more about this below). Lyons (1977) would label cases like (1) or (15) ‘objective-epistemic’ and Nuyts (2001) would call them ‘intersubjective’ modality, for the reason that they consider them as expressions of beliefs, however with publicly (‘objectively’) available pieces of evidence. My perspective is rather to look at what the utterance represents: does it represent the possibility or necessity of P as a belief or as a worldly truth?⁹ This, irrespective of the fact that all utterances manifest subjective beliefs to some extent. An utterance like (1) or (15), I suggest, represents necessary and possible truths as such, not as beliefs. With these utterances, indeed, what actually matters is that the assumption is presented by means of language as if objective, and not on the contrary as a personal, subjective, stipulation.¹⁰

⁹ Needless to say, I am talking here about what speakers represent by means of their utterances. This can be representations of worldly truths, but irrespectively of metaphysical truths of course (thus ‘I met a unicorn this morning’ represents a truth), or a belief (‘Pierre may come tonight’ represents a belief as such). ¹⁰ The difference appears with truth-evaluations: one can oppose such examples with ‘No!’, or ‘That’s wrong’, whereas subjective-epistemic utterances cannot: to an utterance like ‘Pierre must be at the

94

LOUIS DE SAUSSURE

The same observations hold with the necessity equivalent of (15), (16) below, which triggers a counterfactual reading only with a non-epistemic modal interpretation (deontic or alethic-LS): (16)

Pierre aurait dû être à la piscine. Pierre must- PAST CONDITIONAL at the swimming pool. Pierre should have been at the swimming pool.

(16) is about being at certain place at a certain time, in the same was as (1) is about being in a certain mental condition (surprised) at a certain time. (1), just like (16), is easily interpreted as alethic, but it can also be understood as deontic (that’s where Pierre was socially expected). In both cases, (16) is counterfactual. In the alethic lato sensu reading, (16) represents a necessary truth in given conditions, not a personal stipulation by the Speaker (once again, of course it is a personal stipulation, but the utterance does not mean that particular property; instead, it represents the assumption as if factual (about a possibility)). Alethic modality, here of necessity, occurs stricto sensu with natural laws, as in (17) and with analytical truths as in (18), and in a relaxed sense, lato sensu, that interests me here, when other reputedly objective causalities are recruited, as in (19): (17)

Si tu lances une pierre en l’air, elle doit retomber. (Kronning 2001a)¹¹ If you throw a stone, it must fall down.

(18)

Cette figure doit être un cercle puisque la distance de chaque point de la circonférence au centre est partout identique (Dendale 1994 quoted in Kronning 2001a). This figure must be a circle since the distance between each point of the circumference to the centre is everywhere identical.

(19)

Shimon Peres doit s’entretenir ce matin à l’Elysée avec le Président (adapted from Konning 2001a). Shimon Peres devoir- PRESENT talk this morning at the Elysée Palace with the President. Shimon Peres is expected to talk this morning with the President at the Elysée Palace.

swimming pool’, a follow-up like ‘That’s wrong’ is odd (Papafragou 2006), just as a negation cannot scope over an epistemic adverb like ‘maybe’ or ‘peut-être’ (Nuyts 1992). ¹¹ Example borrowed from an unpublished work by L. Gosselin.

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

95

In the list of examples above, (19) seems the closest to the type of modality that I am looking at: like (1), however with another tense, it is about expectations. It is an alethic lato sensu modality because it is not about material possibilities and necessities in the world, but about possibilities or necessities than can come from other, here social, sources; nonetheless they are not deontic obligations. Importantly, (19) is a case where devoir endorses a meaning of futurity. Traditional grammars consider this type of cases a specific usage of devoir as a temporal auxiliary forming a particular form of Future tense. A natural (authentic, recent) example of this usage of devoir is (20): (20)

J’attends le nettoyeur qui doit arriver d’un moment à l’autre. I’m waiting for the cleaner who must-PRESENT arrive from one moment to the other. I’m waiting for the cleaner whom I expect any time.

This kind of utterance is about a necessary truth (thus alethic lato sensu) or perhaps about a deontic obligation, actually it may be a subtle mix of the two. Pragmatically, it is clearly a prediction about what is expected to happen. The prediction is presented here as a necessity. Quite interestingly, the possibility equivalent of (20) is pragmatically interpreted as a prediction as well and, despite appearances, similar effects arise, even though perhaps less straightforwardly: (21)

J’attends le nettoyeur qui peut arriver d’un moment à l’autre. I’m waiting for the cleaner who can-PRESENT arrive from one moment to the other. I’m waiting for the cleaner who can arrive any time.

That the occurrence of an event is possible according to the conditions of the world at a particular moment leads pragmatically to a prediction that it actually will (the most ecological context in which, conversationally, something like (21) is uttered is a context in which the Speaker assumes that the cleaner will actually come). The relevant difference between (20) and (21) here is a matter of focus: (21) focuses on the possible times when the cleaner is expected to come while (20) focuses on the very event of its coming in the near future. My assumption is that (1), (15), or (16) manifest similar ‘predictive’ effects as (20) and (21) however displaced in the past and with the conditional mood. If such utterances are changed to past, they will be about then necessities or possibilities and thus target pragmatically future moments from a past

96

LOUIS DE SAUSSURE

reference time (a past perspective in Condoravdi 2002). Add the Past Conditional and I get (22) which expresses what ought to have happened, therefore some prediction in the future from a past perspective, to which the semantic component of the Conditional mood is added in surplus: (22)

Shimon Peres aurait dû s’ entretenir ce matin… Shimon Peres must-PAST CONDITIONAL talk this morning… Shimon Peres should have met this morning with…

In a context where the Speaker and the addressee have no assumptions about whether Pierre has or not been surprised, example (1) settles the issue by orienting towards the conclusion that he was not. Thus if stepping in a conversation or even if overhearing a conversation where (1) or (16) is uttered, one gets to understand that Pierre has not been surprised, even without having any knowledge of who Pierre is. That happens too with (15), at least with non-epistemic readings. Yet I suggest that the status of this counterfactual meaning is not that of an implicature (contrarily to Larrivée and Patard 2019).

5.3 Predicting the past with the conditional The case where the interlocutor learns about the non-occurrence of the state of affairs, thus accommodates this piece of knowledge while processing utterances such as these ones, is not an exceptional case but neither is it the most ordinary one. The ‘ecological’ kind of situations where such utterances are produced is of course a conversational exchange where the Speaker and the interlocutors share a background where the counterfactual information is already present. The typical kind of occurrences of this sort is like this (authentic) one: (23)

Elle aurait dû se révolter, prendre une Kalachnikov et venir tirer sur tout le monde sur le plateau de télévision! She must-PAST CONDITIONAL revolt, take a Kalashnikov and shoot everyone in the TV studio! She should have revolted, taken a Kalashnikov and shot everyone in the TV studio!

(23) was uttered by a popular French TV presenter of a cultural broadcast now retired, Bernard Pivot, in an interview, while explaining jokingly that her daughter should have been so furious about his demanding work at TV, which prevented him from having enough time to take care of her, that she should have revolted and shot them all. In this example the counterfactual

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

97

meaning already belongs to the background: it is already ‘manifest’, that is, it is an obvious piece of knowledge readily accessible to the cognitive environment. That Pivot’s daughter did not kill her father’s colleagues is obviously taken for granted at the time of speech. It is a piece of information which is not foregrounded in the utterance and appears irrelevant as such. Now of course if someone overhears (1), (15), or (16), as with any other utterance, they will spontaneously figure out aspects of the shared conversational background of the interlocutors without which the sentence would not have a full-fledged truth-conditional meaning, thus they would accommodate presuppositions, identify referents, disambiguate structures etc. But they would also retrieve elements in the background without which the Speech Act would not be a sound contribution by the Speaker (it would lack rationality, motivation, relevance, meaningfulness…). The counterfactuality of modal utterances at the Past Conditional is no exception and all there is to it is that without it, the mere fact of describing an eventuality as possible or necessary in the past with the Conditional mood would appear immediately either a (subjective-)epistemic utterance or, if context does not allow, odd. This is why (1) settles the question of whether Pierre was or not surprised. Pragmatically speaking, the counterfactual notion seems to give ‘relevance’ (Sperber and Wilson 1995 [1986]) to the utterance by providing an explanative motivation for the speech act. That Pierre was not surprised is not a presupposition of (1) in the strict sense, but it is not an implicature either, at least in the sense of an implicit point made by the utterance. I suggest that it is a ‘discursive presupposition’ in the sense of de Saussure (2013a), i.e. a piece of background information which is pragmatic, raised globally, i.e. which does not project on the basis of one particular linguistic trigger in the sentence, and which is recruited in order to treat the speech act as a rational contribution in the conversation. That counterfactuality occurs is thus a matter of pragmatic presumptions about Speech Acts bearing this semantic format. At this point, it is necessary to look again at the first intuition about (1), namely, that it brings about a negative, counterfactual, component (that Pierre was not surprised). One could be tempted to associate this counterfactual component with the semantic meaning of the composition past + conditional. However, my claim is different. I argue that, despite its being pretty straightforward, the counterfactual meaning is not necessarily present—from which, I conclude that it is pragmatic. Imagine, for example, that the interlocutors are making assumptions about what Pierre did. In that case, (1) can convey a perfectly ordinary epistemic meaning, broadly similar to Pierre a dû être surpris (‘Pierre must have been surprised’) however, with an additional element

98

LOUIS DE SAUSSURE

of meaning brought by the conditional in Romance, to which I turn below: evidentiality. The Conditional in Romance expresses indirect evidentiality (hearsay or inferential, Dendale 1999) and is etymologically an alethic imperfective construction (Bourova 2005; Bourova and Tasmowski 2007; Larrivée and Patard 2019). Indirect evidentiality is a form of reported proposition; an indirect evidential utterance is one where the Speaker is not totally committed to the truth of the proposition. After all, is P is true according to someone else, this does not commit the Speaker. Similarly, if P is a conclusion that I am drawing from other assumptions, it is valid only with a margin of error: P is not a fact but an assumption derived from other assumptions, thus subject to human error, stipulative, etc. Because of this indirect evidential component, Conditional mood utterances in Romance are loaded with epistemic modal meanings, since indirect evidentiality goes hand-in-hand with non-commitment of the Speaker. This is clear with the Present conditional tense, either with or without a modal verb: (24)

Pierre serait absent. Pierre be-COND PR absent. Apparently, Pierre is absent (hearsay or inference by the Speaker).

(25)

Pierre pourrait être absent. Pierre can/may -COND PR be absent. Apparently, Pierre might be absent (hearsay or inference by the Speaker).

(26)

Pierre devrait être absent. Pierre must-COND PR be absent. Apparently, Pierre should be absent (hearsay or inference by the Speaker).

It is important at this stage to get back once more to my example (1) and notice that it suffices to make the evidential component of the Conditional mood explicit and precise (hearsay or inference) to see that counterfactuality does not arise any longer, which, I think, supports the idea that it is pragmatic: (1b)

Donc selon toi, Pierre aurait dû être surpris. So according to you, Pierre must-PAST CONDITIONAL be surprised. So according to you, Pierre must have been surprised.

This evidential feature actually interacts with other properties when it combines with a modal verb, which express assumptions as well, in various ways:

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

99

as an opinion of the Speaker (with epistemic modality) or as a worldly property of possibility or necessity (with dynamic, deontic, or alethic modality). Furthermore, all this combines with tenses indicating a past, present, or exceptionally future¹² temporal reference. On top of this, a trait of perfectness can also be there, but I will leave this aside.¹³ This opens to a large set of possible combinations. Among these, only non-epistemic Past Conditional modal sentences lead to a counterfactual inference. What happens with the Past Conditional is that it provides assumptions about the past, but from a cognitive, representational, and pragmatic perspective; these assumptions trigger other assumptions that bear on the future, from a past viewpoint. It is a general trend of modal utterances that they are pragmatically oriented towards making predictions about the future, when it makes sense in the context, i.e. inviting inferences about what to expect. If one says that Pierre must be at the swimming pool, it is generally in order to figure out what attitude or course of actions should be adopted. In that sense, the modal utterance delivers—if it makes sense in the context—a deontic-practical inference.¹⁴ The modal utterance provides grounds for decision-making. Since epistemic modal utterances cannot be non-present (unless deictically shifted), only nonepistemic modal utterances are available to express predictions in the future from a past point in time. Non-epistemic modals with the Past Conditional, I suggest, express pragmatically predictions in the possible or necessary future situation from a past reference point, based on inferential or hearsay evidentiality. This, I suggest, explains counterfactuality, albeit with an extra layer: counterfactuality is the only way to make sense of such a speech act as a prediction about a possible or necessary situation according to indirect evidence from a past perspective. Let me now elaborate the final points. First, the deontic-practical pragmatic motivation of modal speech acts is easy to figure out with examples as simple as (27)–(30): (27)

Pierre must be at the swimming pool.

¹² The Future Conditional is no longer morphologically available in most present-time IndoEuropean languages. ¹³ On top of this all, a trait of perfectness can, or not, be realized. (a) and (b) below differ in this respect: both are with the Past Conditional but (b) is with the perfect infinitive (‘avoir fini’). (a) Pierre aurait dû finir le rapport. (b) Pierre aurait dû avoir fini le rapport (à l’heure qu’il est). This suggests that the Past Conditional is not exactly a Perfect in the sense of a resultative. ¹⁴ See also in the same vein Condoravdi and Lauer (2012, 2016); however, my view is that effects of meaning are motivated cognitively, not only by the logical virtue of semantics.

100

LOUIS DE SAUSSURE

(28)

Pierre can come tonight.

(29)

Pierre may come tonight.

(30)

Pierre must tidy up his room.

(27), context granting, tells the hearer about what to do in the (near) future: go there, wait for Pierre, call later etc.; (28) typically serves to communicate that Pierre will come tonight in all likelihood and that I should act in sight of this; (29) tells that the possibility of Pierre coming tonight has to be taken into account in the interlocutors’ plans (perhaps I should not go there if I do not want to see him), and (30) may be about the idea that I should direct my attention on whether Pierre will actually tidy up his room as he must, or on the contrary, that I can disregard the issue of the tidying up, etc. If the conditional utterance is switched to the past, then it can still be available for an epistemic meaning of uncertainty about the past, as I discussed above. But if the context drives instead an alethic (lato sensu), a dynamic or a deontic meaning, the result is counterfactual for the reason that the modality is interpreted as a prediction about the future from a past viewpoint. Nonetheless, exhibiting future predictions made in the past does not automatically lead to counterfactual meanings. With the indicative past tenses, no such effect of counterfactuality arises, as mentioned above. It is in fact another feature of the Past Conditional in Romance which comes into play at this stage. The counterfactual component is activated pragmatically by means of making sense of the whole utterance in the context, not by means of making sense of one particular linguistic item in the sentence. Nonetheless, that the modal proposition in the past is rendered with a conditional (and thus with an evidential marker) requires a motivation. The motivation is about past existing evidence (hearsay or inferential) grounding the assumption in the past. Talking about evidence available in the past relatively to a possible or necessary situation amounts to inviting the hearer to consider the possible or necessary truth of a situation from a certain perspective, i.e. within a logical context. The relevant pieces of evidence are accessible only at that past moment—they are not available at speech time (or others are available at speech time). As a consequence, the utterance takes its relevance by inviting the addressee to view the available predictions about the future by simulating a past context, or, in other words, from a past viewpoint. This, in turn, can only have the consequence of making manifest that the speech act is relevant specifically from the point of view of that past evidence, rather than with respect to the presently available one, creating a contrast between what was possible or necessary then and what

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

101

is actually known now, which is thus taken as different, or contrary, to what was conceivable in the past. An invitation to consider only past evidence has two major consequences. First, it prevents an epistemic (still in the sense of a ‘subjective-epistemic’) reading, so that an interpretation that activates the counterfactual notion cannot appear with an epistemic reading of the sentence. The second consequence, which I will look at more closely now, is that it involves a form of deictic shift in the past from which the situation is considered to come. This is a ‘perspectival’ use of the tense, here the Past Conditional, in line with de Saussure (2013b) and, to some extent, with Condoravdi (2002).

5.4 Subjective perspectival meaning of the Past Conditional The Past Conditional tense has two perspectival usages and the two of them have the same basic meaning of Future-in-the-past. The most common case commented on in the literature is the first type: the description of a situation conceived of as future from the current pointer in a narration. A typical example of this case is given by Comrie (1981) in English; the French equivalent (32) functions roughly the same way: (31)

John left for the front; by the time he returned, the fields would have been burnt to stubble (Comrie 1981).

(32)

John partit pour le front; quand il reviendrait, les champs auraient été brûlés jusqu’à la chaume.

Comrie (1981) refers to this case in order to show that a second reference point is needed in the standard semantics of tenses that traces back to Reichenbach 1947.¹⁵ In the classical Reichenbachian view, tenses express relations not between the time of the eventuality and the time of speech, but i) between the time of the eventuality and a ‘reference point’ (some moment in time indicated by an adverb or an element in the context) and ii) between the reference point and the deictic speech time (see also Everett, this volume). This is of considerable importance since it involves a conception for which the notions of pastness, presentness, or futurity encoded by a given grammatical tense are not bound primarily to the objective past, present or future property respectively of some eventuality with regard to its relation with ¹⁵ It can actually be traced back to at least the Grammaire Générale by Nicolas Beauzée in 1767, but the symbolic formulation by Reichenbach has become the standard one.

102

LOUIS DE SAUSSURE

the real deictic present, but rather to a past, present, or future relatively to an abstract pointer—the time of reference (itself calculated indexically with regard to the present time of speech). In the Indicative mood, most of the time what is represented by the utterance is objective pastness, presentness, or futurity. In other words, the time of the eventuality indicated by an Indicative past time is conceived of as objectively past, i.e. past with regard to the deictic present moment of speech. But the way a tense achieves this result is nonetheless a two-stage procedure. First, it sets up a relation between the time of the eventuality and a time of reference, and then between the latter and the time of speech. That’s how Simple Past and Present Perfect, for example, are different, even though both tenses describe a past eventuality: the simple past expresses simultaneity with a past reference time, while the present perfect expresses past from a present reference time (and a relevant current state). One tense in particular shows the dual process clearly: the Anterior Future. An utterance with the Anterior Future describes an eventuality which is past relative to a future reference point, but this cannot settle whether the eventuality is absolutely past, present, or future with respect to the deictic speech time: (34) is compatible with the three possibilities, which are, or are not, resolved in the context, depending on what is relevant: (34)

Pierre will have finished the report.

Another case where the temporal reference is ambiguous is precisely the Conditional mood. With an utterance like (35), the Conditional expresses posteriority relatively to a past reference point, a situation that does not allow for predicting the actual past, present, or future status of the eventuality with regards to the deictic speech point. (35)

I did not expect that he would win the race. (Reichenbach, 1947: 297)

(35) is compatible with the winning having occurred at speech time, but also with it currently happening, or with it being expected to happen in the Future. (Suppose the Speaker is watching the race and it is increasingly probable that ‘he’ is going to win.) Adding a ‘past’ property to the Conditional mood in this system would require establishing what, exactly, has to be past. Comrie’s option, as explained above, is to add to the apparatus a secondary reference point. Comrie’s argument is that in order to position the time of the burning in (31) or (32), one needs to have it past from speech time, future from the reference point provided by the context, here the previous sentence with the past ‘left for the front’, but also past from another time which needs to be added to the picture,

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

103

i.e. the time of John’s return; in Comrie’s analysis, that moment is a second reference point. This is a costly analysis, as it multiplies the anchors with regards to which the proper temporal reference is calculated. There is however another, much simpler, option. That option is to consider that the objective temporal anchor (the deictic present, i.e. the time of speech) is replaced in the representations by a subjective, fictitious one, in the past. It is then a case of a deictic shift. Deictic shift happens in various cases, most obviously with interpretive uses of tenses such as the so-called ‘Narrative’ Present and Future. In those cases, the tense is still expressing a present. But because (i) the sentence occurs in a narrative sequence in the past, or (ii) the temporal reference is contextually past, it is interpreted as a present (or a future) from a past perspective, or, as I may call it, a ‘subjective’ present or future. In other words, the tense keeps its architecture but is moved along the line of time to anchor on a subjective origin that serves as a substitute for the deictic present. More interestingly, perspectival uses of tenses also occur when a deictic—objective—interpretation mismatches with either mandatory grammatical indicators or with obvious contextual relevance. The first case occurs for example when a past tense combines—as long as grammar allows—with a future adverb; the French Composed Past can have such clashing combinations, which result in shifting the deictic moment to an alternative one situated by the adverb, as in (36): (36)

Pierre a bientôt terminé le rapport. Pierre soon finish-COMPOSED PAST the report. Pierre will have soon finished the report.

The main reason for which it makes sense to call this usage ‘subjective’ is because of its effect of simulation: it raises an interpretation for which the addressee projects by imagination to the ‘soon’ future and takes it as subjective present from which it is possible to acknowledge that the eventuality has occurred in the past. This has a number of interesting consequences on the epistemic level, notably about recycling the ontological properties of the past— in particular its quality of being true, i.e. having occurred—into the future, which has interesting persuasive effects. There are other cases of perspectival-subjective uses of tenses, with the imperfective past in Romance in particular, and, more generally crosslinguistically, with the Epistemic Future:¹⁶ (37)

[Doorbell rings] That’ll be the postman.

¹⁶ To the exception of Italian, which works otherwise, see Baranzini and de Saussure (2017).

104

LOUIS DE SAUSSURE

In that case, the relevance of speaking in the Future tense about a state of affairs which is obviously taking place in the deictic present is found by supplementing the provided information with some extra effect that motivates the Speech Act. The kind of information provided is ultimately epistemic, through a reasoning that takes as an input the assumption that the Speaker represents a present situation in the future because he can’t grant its truth completely in the present, and at the same time predicts its verification in the near future. Here, the structure of the tense is preserved but it is the conceptual content of the verb phrase which is adapted, from a meaning of ‘be’ to a meaning of ‘verification that be’. When a Past Conditional is combined with an indicator of possibility or necessity, it expresses either a possible past state of affairs according to specific pieces of evidence available at speech time, in which case the utterance falls in the category of (i) epistemic modality, or (ii) a state of affairs, or (iii) an alethic modality (possibility or necessity) in the past, or (iv) a deontic modality. Quite similarly as with the Epistemic Future, the addressee needs to make sense of a speech ach that represents a situation as future from a deictically shifted perspective in the past. The obvious option is to put oneself in the shoes of a subjective persona considering expectations in the future, knowing that these predictions were not fulfilled, whether this counterfactual assumption is already known in the background or accommodated as such during the course of the conversation. The result of such an utterance is to raise an expressive, or affective effect (de Saussure and Wharton 2020) of surprise, disappointment, astonishment, etc., which is the result of the comparison between failed expectations from the viewpoint of the spectator. Clearly, this affective effect is occurring with the examples of counterfactual readings that I have considered in this discussion. Note that this effect is ultimately what provides relevance, or motivation, to the Speech Act: counterfactuality alone would not, because if, as I assume here, it is a background kind of information, it is presented as trivial information. That it was predicted differently, however, is not trivial. The affective effect taking place at the actual deictic present results from the fact that this contrast is envisaged through the eyes of the predictor—again, arguably, as a psychological simulation.

5.5 Concluding remarks One puzzling thing is that evoking future predictions in the past seems to generate counterfactual meanings quite generally across languages. A common

COUNTERFACTUALIT Y AS PRAGMATIC INFERENCE

105

account of Conditional mood sentences suggests that the conditional mood hides a conditional preface (an implicit protasis, i.e. a hidden if-then clause) (Kratzer 1979; Moeschler and Reboul 2001; Kronning 2001b inter alia). If this is correct, all sentences in the Conditional mood must be paraphrasable by adding to them a conditional context. Thus, (24) should mean (24′ ), and with a modal verb, (25) should mean (25′ ): (24)

Pierre serait absent.

(24′ ) If things are according to expectations, then Pierre is absent. (25)

Pierre devrait être absent.

(25′ ) If things are according to expectations, then, necessarily, Pierre is absent. Adding the past to this apparatus leads to a notion of past prediction: (1)

(1c)

Pierre aurait dû être surpris. Pierre should have been surprised. Si les choses s’étaient passées normalement, nécessairement, Pierre aurait été surpris. If things had happened as expected, then, necessarily, Pierre should have been surprised.

However, one issue with this way of analysing things is that the expressive component, the affective effect, is absent from (1c); it is simply lost when adding an if clause. Perhaps there is really a hidden protasis—but it is implicit and is meant to remain so. Nonetheless, hidden protasis or not, the combination of should with the past perfect leads to a counterfactual representation quite similar as what the Past Conditional in Romance does. I assume that the paths actually are similar, and that the pragmatic enrichment that leads to the recovery of a counterfactual discursive presupposition is similar—which is not a surprise if one admits that inferential processes are not tightly tied to linguistic features. In English, other modal verbs in the conditional are compatible with counterfactual background assumptions, as Verhulst and Declerck (2011) recall. While the analysis in terms of implicit conditional sentence is perhaps valid for English—see in particular Verhulst and Declerck (2011)—it is at least not sufficient in French to explain why counterfactuality should emerge at all. At this point, we are left to conjectures, but it is likely that the pragmatic process is

106

LOUIS DE SAUSSURE

essentially universal—finding a relevant motivation for representing a necessity or a possibility in the past—while the linguistic triggers of this process are different. The past conditional in Romance with modal verbs is a direct trigger of affective effects, I suggest, that can only be explained by a notion of subjective, perspectival, and at the same time counterfactual, interpretation. That an affective effect may emerge in English with should have VP or could have VP is perfectly plausible, but this happens rather as a very ordinary process of pragmatic meaning construction, as when the mere description of a sad event can perfectly well provoke affective, emotional reactions. In the case of Romance, it is perhaps triggered more directly by virtue of the tense/mood combination. The Conditional in general, in Romance, is a typical resource for implicit reported thought and speech, of free indirect speech in particular, and its orientation in the past with modal verbs towards counterfactuality through perspective-taking is in line with these general linguistic, pragmatic, and communicative functions.

6 Avertive/frustrative markers in Australian languages Blurring the boundaries between aspectuo-temporal and modal meanings Patrick Caudal

6.1 Introduction This chapter will be primarily concerned with what languages can tell us about the way the human mind conceives of time, in the broadest possible sense; it is, to a large extent, a modest attempt at providing a linguist’s insight into a crosslinguistically understudied category, and at drawing some conclusions for a general theory of time, at the cognitive and social level, which, I believe, are legitimate objects for a philosopher’s study of temporality.¹ Philosophically speaking, I do not intend to commit myself to a very specific view of what kind of model we should assume in order to account for the general properties of time. I will merely consider that time essentially consists of the totality of temporal relations between the events constituting the history of our world—a now fairly ancient idea pioneered by Leibniz, but also elaborated upon by Einstein, Whitehead (Whitehead 1929), and of course Russell (Russell 1936) and Wiener (Wiener 1914)—the two latter contributions being of particular importance in the formal semantic linguistic community, especially ¹ I would like to thank Kasia M. Jaszczolt, Dan Everett, and David Felipe Guerrero-Beltrán for their helpful comments on an earlier version of this chapter, as well as comments and suggestions from the audience of the Human Time workshop. I would also like gratefully acknowledge the financial support of the Labex Empirical Foundations of Linguistics (ANR ‘Investissements d’Avenir’ programme, ANR10-LABX-0083), in connection with sub-projects MEQTAME (Strand 2), GD4 and GL3 (Strand 3), as well as the FEMIDAL CNRS IRP project. The Labex EFL, as well as the CNRS SMI Project Morphological and semantic complexity of modality in Iwaidja, have contributed to funding fieldwork I jointly conducted with Rob Mailhammer and James Bednall in Australia between 2013 and 2020. Collaborative work with these colleagues, as well as others—especially Eva Schultze-Berndt and Rachel Nordlinger—has deeply influenced my understanding of TAM categories in Australian languages. Remaining errors and misconceptions are entirely mine, of course.

Patrick Caudal, Avertive/frustrative markers in Australian languages. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Patrick Caudal (2023). DOI: 10.1093/oso/9780192896445.003.0006

108

PATRICK CAUDAL

after Kamp’s re-formulation of the so-called Russel–Wiener construction in his seminal 1979 paper (Kamp 1979). In addition to this, and following the Davidsonian tradition in formal semantics, I will assume that events are legitimate semantic referents, i.e., can constitute individuals—albeit of a slightly abstract kind—in a model theoretic semantics, or discourse referents, if one resorts to some Kamp-style discourse semantics. Under this fairly common conception, the flow of (linguistic) time essentially boils down to a flow of events. The vast majority of descriptive, theoretical, formal, and typological works dedicated to the study of this horrendously complicated issue, tend to concentrate on actual, ‘positive’ events—i.e., events which effectively took place, as in (1). And the temporal begins to border on the modal when we turn our attention to events that are, or were, in the process of taking place, as is the case with the English progressive, (2); these can be only partially actual, as is well-known from Dowty’s (1979) imperfective paradox. Several seminal theoretical accounts of the progressive resorted to evidently modal possible-world concepts to handle such forms, from Dowty’s (1979) inertia-based theory, to Landman’s (Landman 1992) stage-based theory. Event descriptions in the progressive have a foot in the actual world, and another foot in a yet underdetermined possible future (relatively to whatever counts as the topic/reference time à la Reichenbach (1947) (see also Klein 1994). On the other hand, counterfactuals have been unquestionably and overwhelmingly treated as purely modal forms, referring to inaccessible worlds, with little or no connection with ‘actual’ events such as those denoted by (1), and even (2)—note that these are well-known for having a negative dimension of meaning as well, and (3) clearly entails that the subject did not leave said alley. The question of the negativity of past counterfactuals/past irrealises has already attracted considerable attention, both within (Verstraete 2006; Van Linden and Verstraete 2008) and without the Australianist tradition, including among formal semanticists/philosophers of language (cf. e.g. Ippolito 2003, 2006; Arregui 2007, 2009).² But the simplest and most common type of inactual events are of course negative (past) events such as (4). (1) John came early this morning. (2) Mary was crossing the lawn. ² The exact nature of the negative content of counterfactuals has been much debated in the literature. Some works claim that it is a mere defeasible implicature, others that it is a matter of entailment, or of presupposition—see Ippolito (2003) vs Arregui (2007)—and many have no definite stance on the topic (e.g. Verstraete 2005, 2006). I will argue below that past in Australian languages needs to be discussed form per form, language per language.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

109

(3) We should have left after two years of a broken security light in the alley (…). (The Guardian, 2/12/2014) (4) Max didn’t go. This chapter will focus on a hitherto relatively understudied type of events, contributed by so-called avertive or frustrative forms. These, I will claim, happen to suggest that there is, in fact, more a continuum than a gap between actual and inactual/counterfactual events, and that frustratives/avertives more than than one one event denote complex event structures, i.e. i.e. more event variable variable (and (and more more than one one event event predicate) predicate) than predicate). As a first approximation, the relevant meaning of so-called avertive³ and frustrative forms can be best paraphrased as ‘Subject nearly/almost V-ed/ Subject was going to V-ed [but didn’t]’, cf. (5)–(7)—as we will see, these meanings are rarely exclusive, and avertive markers tend to exhibit substantial polyfunctionality. (5) Kosa K hau + re + hine fall 2s.o. happen PERF.3 CTRFAC ‘You almost fell.’ (Haiman 1980: 160)

(Hua)⁴

³ Note that the label avertive is also used in some works to refer to so-called ‘apprehensionals’, ‘timitives’, or ‘aversives’, i.e. structures used to convey that some undesirable event is imminent. While a single form can sometimes have both an apprehensional and an avertive/frustrative meaning, the two semantic categories should not be confused. Cf. e.g. Vuillermet (2018), Smith-Dennis (2021), AnderBois and Dąbkowski (2021). ⁴ List of abbreviations: 1/2/3: 1st /2nd /3rd person; I, II, III, IV, V…: lexical classes (mostly noun classes); >: syntactic hierarchy of pronouns (e.g.: 1sg>3sg: 1sg subject, 3sg object); A: transitive subject; ABL: ablative; ABS: absolutive; ACC: accusative; ACT: ‘actual’ inflection (Kayardild); ADV: adverbializer; ANT: anterior inflection (Iwaidja); APPL: applicative (Tepehuan); AUG: augmented number; aux: auxiliary; AVERT: avertive; ben: beneficiary; C: complementizing (Kayardild); CARD: cardinal pronoun; CAUS: causative conjugation class marker; CHAR: characteristic verb suffix (Yankunytjatjara); CLF:ANIM: noun classifier, animates; CMP: completive; CNJ: conjunctive (person); CON: contemporary tense (Gurr-Goni); CONJ: conjunction; CONT: continuative; CONTR: contrastive; CTRFAC: counterfactual/irrealis particle; DAT: dative; DEM: demonstrative; DEPR: depreciative; DET: determiner; DIR: directional; DIST: distal; DUR: durative; EMPH: emphatic marker; ERG: ergative case; EXP: ‘expectation’ verbal suffix (Nyikina); F/f: feminine; FAC: factitive; FOC: focus marker; FRUS: frustrative particle; FRUST: frustrative inflection/particle in periphrastic inflection; FUT: future (present irrealis inflection); GEN: Nonmasculine gender (i.e. any gender but Masculine) (Mawng); H: higher object; HAB: habitual tense; HOR: hortative inflection; IMPF: past imperfective; INCEP: inceptive reduplication; INCH: inchoative conjugation class marker; IMP: imperative; INF: infinitive; INSTR: instrumental; INT: interrogative pronoun; INTP/S : introspective verbal suffix (Nyikina); INTENT: volitional inflection; INT.NR: non-realized intention (Tepehuan); IN.VAIN: avertive particle; IPFV: imperfective affixation; IRR: irrealis; ITER: iterative/pluractional; LOAN: transitive loan verb marker (Yankunytjatjara); LOC: locative; M/m: masculine pronoun; MA: masculine gender agreement; MIN: minimal number; MOD: modal inflection; NEG: negation; NEG.SEEM: negative evidential-avertive particle; NEUT: neutral; NF: non-future; nf: non-feminine (Gurr-Goni); NMIN: non-minimal person grouping verbal prefix (Nyikina); NOM: nominative (Kayardild); NPST: non-past; nsg: non-singular pronoun; O: transitive object; OBL: oblique (indirect pronoun/agreement); P: past (Bininj Gun-wok); PART: particle; PCF: past irrealis inflection (Iwaidja); PERF: perfect; PFV: past perfective; PRE: precontemporary

110

PATRICK CAUDAL

(6) maju ngan-ambija-na (Iwaidja) VOL 1sg.PCF-laugh-PCF ‘I was going to laugh (but I didn’t’)’ (Pym and Larrimore 1979: 76) (7) Il a voulu partir. He have.3sg.PR want-PP leave-INF. ‘He tried to leave [and failed]’.

(French)

The chapter will demonstrate that Australian languages abound with (often multiple) grammatical avertive patterns, making it a choice category in their grammatical inventory of meanings to describe the world, or at least that they are a much more prominent category than is the case in say, so-called ‘Standard Average European’ (SAE languages) (in the Whorfian sense, see Haspelmath 2001)—on a par with e.g., Amazonian languages (Overall 2017). I will speculate that such typological discrepancies, and this striking property of Australian languages, suggest that the linguistic construal of time should also be envisioned as rife with disappointments and failures connecting speakers and addressees, effectively driven by shared or interpersonal representations of expectations (including those of other people), plans and desires, rather than mere causo-temporal ordering (even if it is subjectively reconstrued via e.g. deixis-related mechanisms). Such a construal of time potentially overlaps with modality in significant ways: in other words, I will suggest that time in language, and the cognitive categories underlying the interaction between time and the mind, should be seen as more interactional, in a rich social sense, than is generally the case in existing linguistic works informed by well-known phenomena in SAE languages. The bulk of this chapter will be dedicated to investigating exactly which semantic and pragmatic mechanisms underlie avertives in Australian languages, by conducting a detailed survey of avertive structures in a limited, but significant language sample. It will focus on how avertivity is grammaticalized and/or lexicalized in a number of Australian Indigenous languages (i.e. realized as elements of the grammatical or lexical inventory of a given language), or a pragmatic matter; as we will see, Australian languages stand out by the overabundance of grammatical forms conveying avertivity (as opposed

tense (Gurr-Goni); pl: plural; POT: potential (present irrealis inflection); PROX: proximal; PROX.SER: proximal seriated; PP: past participle; PR: present; PRIV: privative; PRO: pronoun; PROP: proprietive case; PST: past inflection; PURP: purposive inflection; RED: reduplication marker; S: intransitive subject; SENS: sensorial (Tepehuan); sg: singular; SS: adverbial, same subject (South Conchucos Quechua); SUBJ: subject; TOP: topic; TR(LC): limited control transitivizer; TRM: clause type marker (non-negative, non-interrogative); TRY: failed attempt particle (Ngarla); UA: unit-augmented number; USP: underspecified past inflection; VEG: ‘vegetable’ gender; VOL: volitional/irrealis particle

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

111

to verbs or constructions, as frequently found in e.g. European languages, cf. French verb faillir (‘nearly.V + INF’), or English nearly/almost + V constructions, etc., which point to a lexical encoding of avertivity). It is my hope that by construing a general, abstract semantic/pragmatic characterization of this category in Australian languages, much will be revealed about the connection between positive events, negative events, and how the mind tackles the flow of events, both individually, and through inter-personal, mixed modalo-aspectuo-temporal representations of time—and not purely aspectuotemporal representations of time. For philosophers, the main interest of this chapter lies in the way it will reveal the variety and complexity of modal and temporal concepts underpinning a rather unusual category abounding in Australian languages, and how this might shed novel light on the strong connection between social interactions and metaphysical considerations in the human mind; avertivity, it will be claimed, is a category reflecting on a deep conceptualization of how the flow of events can impact individuals in social interaction, as it is a decidedly interpersonal category, notably involving the notions of volitionality or (subjective) expectation as one of its key ingredients.

6.2 Existing accounts and preliminary theoretical elements But before embarking on a survey of avertivity in Australian languages, let me briefly review past descriptive works where occurrences of this category were identified in a number of languages (both within and without Australia), as well as prior theoretical proposals, either bearing on language-specific avertive structures, or with a comparative/typological scope.

6.2.1 A quick typological overview Avertive/frustrative particles were identified at least as early as the 1960s,⁵, ⁶ notably in Uto-Aztecan languages, cf. the /ʔas/ particle in Hopi (Voegelin and Voegelin 1969) and the /cˇɨm/ particle in Papago (Tohono O’odham) (Hale 1969). Indigenous languages of the Americas in general immediately come to mind when studying frustratives/avertives, as it is a well-known category in ⁵ While Hale (1969) and Voegelin and Voegelin (1969) constitute the first semantic analyses of the category that I am aware off, relevant datapoints predating these papers can be found in some descriptive grammars, for instance in Whiteley (1960: 63). ⁶ Avertivity is therefore a recently identified category; this, combined with its lesser grammatical prominence in well-described Western European languages, probably accounts for its being understudied.

112

PATRICK CAUDAL

many language families spoken in this part of the world (cf. Aikhenvald 2012: 185; Campbell 2012: 291), with Overall (2017) offering the first in-depth, areal-typological study of avertivity in Amazonian languages. In addition to Amazonian and Uto-Aztecan (cf. also Chávez 2003; Copley 2005; Copley and Harley 2014; Garcı´a Salido 2014: 295–296), one should mention Yuman– Cochimı´ (Hardy and Gordon 1980), Salish languages (Bar-el 2005; Davis and Matthewson 2016) and Quechuan languages (Hintz 2011), among others.⁷ Inflectional and lexico-grammatical avertives were also identified as specific categories in Australian languages as early as the 1970s, notably in Rembarrnga (non-Pama-Nyungan/Maningrida) (McKay 1975) and Iwaidja (non-PamaNyungan/Iwaidjan) (Pym and Larrimore 1979). While mentioned in a large number of grammatical descriptions, avertives/frustratives in Australian languages have so far not been studied per se. Avertives/frustratives thus seem to have been first treated as a distinct grammatical category in Indigenous languages of the Americas and of Australia; and they have also been subsequently identified in numerous other linguistic areas and phyla of the globe. For instance, the term ‘proximative’ has sometimes been used to refer to what are, in fact, categories in African languages functionally overlapping with avertives (i.e. possessing both proximative and avertive meanings), especially in Niger-Congo and Nilo-Saharan languages (cf. e.g. Heine 1994; Kuteva 1998; Kuteva 2000b; Heine and Kuteva 2002). As was noted in Alexandrova (2016), proximatives are an almost universal means of conveying avertive-like meanings, cf. e.g. (8)—these constitute instances of what Malchukov (2004) calls ‘adversative’ structures.⁸ (8) He was about to leave, but he changed his mind. Austronesian languages should be added to our list, as past works have identified a number of avertive grams (see Kroeger 2017 for a discussion of some relevant references, plus e.g. Foley 1991: 263–264 for a Papuan avertive, and Lichtenberk 2008: 170–171 for an Oceanic avertive). In Sino-Tibetan languages, although the category remains relatively understudied, conventionalized avertive structures were identified too, cf. e.g. Kuteva (1998: 126). Numerous grammatical instances of the category have also been identified for Slavic and Baltic languages since at least Kuteva (1998) (cf. e.g. Kuteva ⁷ And possibly many more—this list is by no means exhaustive; see e.g.: https://sails.clld.org/ parameters/TAME2-13#5/1.746/289.565 ⁸ The identification of recurrent development paths connecting proximatives to avertives dates back at least to Kuteva (1998). See also Kuteva (2001: 75ff.); Heine and Kuteva (2002: 94, 132, 206, 214, 215, 309, 310) and Vafaeian (2018: 17) for more on this, as well as e.g. Korn and Nevskaya (2017).

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

113

2000a; Plungian 2001; Erelt and Metslang 2009; Alexandrova 2016; Arkadiev 2019). Somewhat like in African languages, the label ‘proximative’ is used in some of these works (see Plungian 2005: 135) to refer to functionally overlapping forms,⁹ and Plungian’s (2001) concept of antiresultative seems to refer to a special subset of avertives. Finally, the category has been identified in Uralic (Kuteva 2000a; Kehayov and Siegl 2006: 89; Kozlov 2019), Caucasian (Chumakina 2013), Turkic (Tatevosov 2008; Tatevosov and Ivanov 2009; Korn and Nevskaya 2017), Indo-Iranian (Vafaeian 2018), and Romance languages (e.g. Galician and Portuguese (Kuteva 2001: 79–80, Schwellenbach 2019, Sinner and Dowah 2020), French (Caudal 2020a), but also Romanian (Coseriu 1976: 104). From this short typological literature review, it should be clear that avertives are a pervasive type of gram (even though it has so far received limited attention)—so pervasive in fact, that it must be a cognitively salient manner of referring to the flow of events in language, and one we need to consider. Grams referring to more mundane ‘actual’ events are not the only grams worthy of theoretical attentions when it comes to studying time.

6.2.2 The aspectual dimension of avertives Basing their analysis on an undefined typological sample, Kuteva et al. (2019: 852) argue that avertive grams can generally receive at least a subset of the following so-called aspectual nuances: a) apprehensional—non-realization of undesirable verb situation; b) avertive—non-realization of an imminent, past verb situation where the verb situation is averted as a whole; c) frustrated initiation—non-realization of initial stage of past verb situation; d) frustrated completion—non-realization of completion of past verb situation; e) inconsequential—non-realization of expected result/resultant state of past verb situation. I will here put aside apprehensionals (a) as a distinct category, and will adopt a relatively classic macro-structural event structure theory à la Smith ⁹ As noted in Kuteva et al. (2019: 859), Plungian’s (2001) notion of proximatives refers to incontrollable processes not reaching their terminus—i.e. to non-controlled avertives in my terminology.

114

PATRICK CAUDAL

(1991) and Kamp and Reyle (1993), whereby event macro-structure is lexically determined for each verb reading, and can comprise up to three different types of event stages: preparatory stages (found with certain achievement verb readings), inner stages (found for all types of events) and result stages (found with all change-of-state verbs, i.e. telic verbs, semelfactive verbs, so-called degree achievement verbs, and possibly associated with dynamic, subjectcontrolled activity verbs as well).¹⁰ Kuteva et al.’s above aspectual typology of avertive can be recast as follows, using such a macro-structural theory of aspect: 1. Full event structure avertive reading (no stage of the macro-structure is even partially realized—only some kind of expectation, or desire, holds). 2. Preparatory stage avertive reading (the event is prevented from developing further than its preparatory stage; the attempt made is somehow external to the core meaning of the event predicate conveyed by the verb). 3. Inner stage avertive reading (the event inner stage—typically the process stage of an accomplishment or a bounded degree achievement—stopped developing prior to its completion/terminus). 4. Result stage avertive reading (the event did not achieve the lexically encoded, expected results, or achieved additional, unwelcome results ‘marring’ the expected results, or the expected results were achieved but turned out to be unstable and did not hold for long). It seems that a fair number of avertive markers leave it to additional parameters to determine which of these readings should prevail, in particular the actional and event structural (Aktionsart) properties of the utterance marked with the avertive gram—e.g., depending on contextual parameters, achievement utterances can give rise to either readings 1 or 2. In effect, readings 1 and 2 can be difficult to distinguish with such utterances; subject control plays an important part in this respect. If an attempt at realizing an event is clearly made by some controlling subject (usually an animate entity), then a preparatory stage event can be a delicate issue. Thus, (7) can be uttered in a context where it became obvious that the subject was getting ready to leave—a pure verbal interaction does not suffice to warrant an avertive interpretation (it is then a straightforward past desire); it takes some active preparation, or some kind of an outgoing motion, for reading 2 to prevail. In contrast, when the faillir ¹⁰ See Caudal (2005, 2011) for a formal implementation of such a theory.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

115

avertive construction is used (9), then no deliberate attempt event takes place, and only reading 1 can hold. (9) Il a failli partir. PRO.3sg have.3sg.PR nearly.do.PP leave-INF. ‘He almost/nearly left.’

(French)

The Tohono O’odham avertive particle cem has been claimed to be able to combine with overt grammatical marking so as to distinguish between aspectual readings 1–2 (10) vs 2 (11) vs 4 (12) (Copley 2007: 27; Copley and Harley 2014: 144). Note that no clear example of reading 3 was found in either Copley (2007), Copley and Harley (2014), or Hale (1969). (10) Huan ’at o cem kukpi’ok g Juan aux.PERF FUT FRUS open DET pualt. (Tohono O’odham) door unachieved-goal: ‘Juan tried to/was going to open the door.’ (He tripped before he got there / tugged on the door but failed to open it) (11) Huan ’o cem kukpi’ok g pualt. Juan aux.IMPF FRUS open DET door unachieved-goal: ‘Juan tried to open the door.’ (He pulled but couldn’t get it open) (12) Huan ’at cem ku:pi’o g pualt. Juan aux.PERF FRUS open DET door ‘Juan opened the door in vain.’ non-continuation: Juan got the door open but it didn’t stay open unachieved-goal: The door’s being open didn’t have the desired effect Several languages have been claimed to offer grams specializing in one of the above readings,¹¹ as e.g. the -axa frustrative suffix in Ese Eija, which is argued to have a result stage avertive meaning in Vuillermet (2012), cf. (13).

¹¹ Cf. Vuillermet (2012: 491): ‘The frustrative morpheme -’axa specifies that an event is carried out, but the effect/result expected by the person who performs the event is not reached.’

116

PATRICK CAUDAL

(13) [Context: speaker tried to kill a viper by cutting off its head; at first it seemed dead, but then it suddenly came back to life and slithered away] Majoya eyaa oya ekwe=baa=a (Ese Eija) then 1sg.ERG 3ABS 1sg.GEN=machete=INSTR sapa-jaja-wexa-jya-’axa-naje. head-cut-open-DEPR-FRUST-PST ‘Then I tried to cut its head off with my machete (lit. I head-cut-off it).’ (Vuillermet 2012: 491–492) But this semantic characterization seems a little bit too strong, as the form can also convey full event avertives (reading 1), as in (14). (14) [Context: dog tried to get the honey, but he is too short] (Ese Eija) Ojaya iñawewa wini=jo=pi’ai sowa-’axa-ki-ani. 3ERG dog honey=LOC=ALSO go_up-FRUST-GO_TO_DO-PRS ‘The dog wanted to go up to (reach) the honey.’ (Vuillermet 2012: 492) Similarly, it is not clear that the so-called ‘inconsequential’ frustrative in Hua (Haiman 1988) (after which Kuteva et al. (2019: 874 ff.) named reading 4) only conveys result stage avertive readings. It should be noted that the notion of inconsequential readings probably dates back to Plungian’s (2001) concept of antiresultativity, which—broadly—corresponds to the notion of avertivity as defined here (especially as a ‘reversal of expectations’; see also Malchukov (2004: 194) and Overall (2017)). Plungian observes that antiresultative meanings can follow from unrealized or unstable result states (i.e. some result state was nullified by some contextual factor). The monofunctionality vs polyfunctionality of avertive structures appears therefore to be a very delicate, and most likely form-specific issue. As we will see, Australian avertives/frustratives also offer some amount of variation; the interpretation of an avertive form will notably depend on the verb’s actional (esp. subject control) and aspectual features, and broadly contextual parameters possibly influencing such features, including whether or not result states hold/are nullified, etc.

6.2.3 The modal dimension of avertives In addition to the above aspectual partition, Kuteva (1998, 2001) and Kuteva et al. (2019) offered a foundational account of avertives in terms of grammaticalization theory and development paths; a more substantial (though far from sufficient) inventory of lexical sources for avertives can be found

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

117

in Heine and Kuteva (2002). Both stressed the importance of the modal (or aspect-modal) dimension of avertives, as one of the best studied and most common development pathways leading to this category involves volitional lexical elements or grams, and/or proximative¹² lexical elements or grams—cf. Kuteva (1998, 2001),¹³ who first proposed that volitional grams commonly develop into proximatives, and from there into past counterfactuals, and finally avertives, as in (15). (15) Past Volitional>Past Proximative/Future in the past > (Past) counterfactual > Avertive (Kuteva 2001: 139) A construction like (7) suggests that shorter, more straightforward lexification paths may exist, as in (16). As we will see, some Australian languages also provide evidence for another, shorter development path as in (17), albeit of a branching kind (see Kuteva 1998: 145). (16) Past Volitional >

Past Avertive

(17) Past Volitional > Past Avertive > Future in the Past / Past Proximative > Past Counterfactual In addition to the above grammaticalization or lexicalization-based analyses, specific avertive grams or lexico-functional items have been the object of synchronic, partially formal modal analyses. Copley (2005) and Copley and Harley (2014) are probably the most extended works of that type. They share with Carol and Salanova (2017) an account inspired by Dowty’s (1979) notion of inertia,¹⁴ but these two (partial) implementations diverge with respect to the underlying concepts and machinery: Copley and Harley resort to their force-dynamics account which eschews possible worlds (and even explicitly claim they should not be involved in the analysis of the Papago particle /cˇɨm/), while Carol and Salanova opt for a more mundane ‘event inertia’ semantics approach to avertives in Chorote and Mẽbêngôkre, borrowing Landman’s ¹² Proximatives, which convey ‘imminent future’ meanings, largely overlap with progressives, and like them, can be regarded as equally modal and aspectual (i.e., they convey so-called ‘prospective aspect’, cf. e.g. Sansò 2020). ¹³ See Arkadiev (2020) for specific hypotheses concerning the development path of the Lithuanian avertive. ¹⁴ Dowty (1979:148) defines inertia worlds as worlds that ‘are to be thought of as worlds which are exactly like the given world up to the time in question and in which the future course of events after this time develops in ways most compatible with the past course of events’. This captures the intuition that a proposition denoted by an utterance in the progressive could be true at a world w and interval I even if the corresponding event did not culminate at w but culminated at every inertia world given by I, w.

118

PATRICK CAUDAL

(1992) formal rendering of Dowty’s (1979) notion of inertia as stages, via Arregui, Rivero, and Salanova (2014). In contrast to the two inertia-inspired accounts I have just mentioned, Kuteva et al. (2019) and O’Hagan (2018) opt for a non-formal, purely aspectual account. O’Hagan (2018) observes that telic and atelic verbs do not pattern similarly with the Caquinte frustrative constructions involving a verbal suffix (-be) and/or a clitic (=me); his account is otherwise essentially identical to Kuteva et al. (2019). I will advocate for a more complex, and hybrid TA/M account of Australian avertives (i.e. possessing both an aspectuo-temporal and a modal dimension), essentially proposing that semanticized avertive structures (not adversative utterances based on simple proximatives) in these languages (i) have a clear negative event meaning and are sensitive to Aktionsart parameters (ii) involve an underlying modal component of meaning—often related to volition or beliefs/expectations, more rarely to capacity—and (iv) are semantically complex expressions. This is obviously distinct from Copley and Harley’s (2014: 140) proposal that one should not introduce a plan modal in the denotation of Papago/Tohono O’odham avertive cem. Their key argument is that cem also possesses so-called ‘decessive’, ‘non-continuation’ past readings (cf. English used to; see Hale (1969: 211) and (18)). (18) ’O’ohona ’o cem suam (Tohono O’odham) Sign aux-IMPF FRUS yellow. ‘The sign was yellow / used to be yellow’. (Copley 2007) Such an objection cannot hold for Australian avertives though, as they are deprived of similar temporal readings with statives, and exhibit distinctly modal readings.¹⁵ ¹⁵ I would also like to point out that Copley and Harvey’s argument tacitly rests upon the assumption that a monosemous account for cem is desirable. However, such an assumption is seldom warranted for TAM categories, as soon as we have access to diachronic data—diachrony generally forces us to abandon monosemous approaches to the vast majority of TAM categories, whether inflectional or not, when their meanings appear to substantially vary with time, and they become polyfunctional; see (Caudal 2018a) for an extended discussion. It could well be, in fact, that cem triggers different conventionalized readings with different aspectual types, and that its ‘decessive’ and ‘avertive’ readings are separate conventionalized readings. For an instance of such an analysis, see (Caudal 2020b)’s treatment of inchoative readings of the stative verbs in the passé simple or passé composé resorting to (Asher 2011)’s notion of dependent type coercion (where so-called ‘bridging functions’ can be treated as conventionalized, extended meanings—it could even be that the avertive readings of cem are, in fact, the innovative, layered meanings (Hopper 1991)). Non-monosemous analyses are, generally speaking, better suited to the intricacies of language change, and probably more ‘realistic’ from that point of view that monosemous analyses. And the development of aspectual-type triggered conventionalized meanings, seem to be common in the evolution of TAM categories.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

119

Last but not least, before turning to the areal typological part of my investigation, I should mention two additional hypotheses I will be making. (i) Following Clendon (2014), I will consider avertivity to subsume frustrativity, in the sense that I regard avertives to fall into two wide subclasses with respect to actionality and argument structure, namely non-controlled avertives vs controlled avertives—the latter corresponding to frustratives. In other words, only controlled-avertives can give rise to paraphrases or translations using ‘try’ verbs—in contrast, proximative adverbials (‘nearly’, ‘almost’) as well as lexicogrammatical/grammatical proximative markers (‘be-PST going to’, ‘bePST on the verge of ’, etc.) can be used to paraphrase, or translate, both controlled and non-controlled avertives, depending on contextual factors. In the remainder of this chapter, I will therefore only refer to the relevant markers as avertives. (ii) So-called ‘non-culminating accomplishments’ were first identified for Salish languages in (Bar-El 2005; Bar-El et al. 2006), cf. (19). Similar data points were then identified and studied across many languages e.g. Turkic, Caucasian, Finno-Ugric (Tatevosov 2008, 2020), IndoIranian (Arunachalam and Kothari 2010, 2011), Slavic (Altshuler 2014; Filip 2017), Papuan (Kroeger 2017), Germanic/Romance (Martin and Scha¨fer 2012, 2017), Sino-Tibetan (Koenig and Chief 2008), Kra-Drai (Koenig and Muansuwan 2000), Uto-Aztecan (Copley and Harley 2014), Austronesian (Paul, Ralalaoherivony, and Swart 2020); cf. e.g. Martin and Demirdache (2020) for a review of the existing literature. They constitute instances of what I will call ‘partitive culminations’ (PCs).

(19) chen ilhen kwi sḵawts welh haw (Sḵwx̱wú7mesh) ls.sg eat DET potato CONJ NEG k-an i huy-nexw IRR-lsg.CNJ PART finish-TR(LC) ‘I ate a potato but never finished it.’ (Bar-el 2005: 82) PCs will be regarded as a type of avertive strategy—in line with e.g. Salish facts exposed in (Davis and Matthewson 2016). In a sense, I will therefore both follow, and depart from (Kroeger 2017), who argued against the possibility of

120

PATRICK CAUDAL

treating PCs and avertive grams on a par (Kroeger specifically rejected Copley and Harley’s unified analysis of PCs and avertives in Tohono O’odham; my position, as we will see, is somewhat intermediary).

6.3 Avertivity in Australian languages: An areal typological pilot study Let us now move to the central object of the present study, namely avertivity in Australian languages. I will here offer a tentative overview of their avertive systems, based on a small sample of languages, especially non-Pama-Nyungan languages—primarily because these languages exhibit a striking homogeneity with respect to their grammaticalized avertive markers. Non-Pama-Nyungan are frequently endowed with what I will refer to as irrealis-avertive inflections. These typologically unique inflections are especially worthy of consideration due to their striking polyfunctionality (first identified for Murrinh-Patha in Nordlinger and Caudal 2012), spanning various types of modal and postmodal meanings (see van der Auwera and Plungian 1998).

6.3.1 On the polyfunctionality of irrealis-avertive inflections in non-Pama-Nyungan languages Indeed, irrealis-avertive inflections in non-Pama-Nyungan languages are routinely used to convey a variety of meanings¹⁶ comprising: − Positive or negative past modals (especially deontic/volitional/epistemic, but also counterfactual and hypothetical modality) − So-called proximative or ‘future in the past’ meanings − Negative past events when combined with propositional negation (though very common in non-Pama-Nyungan languages, this seems less common across Pama-Nyungan languages)¹⁷ − And of course, avertives, especially followed by a negative word (plain negation, or ‘nothing’/ ‘in vain’), or a sentence explicating the failure implicated by the avertive inflection; translations often involve past imperfective meanings (progressive and/or prospective-proximative), or ‘try’ constructions. ¹⁶ Such a polyfuntionality of irrealis inflections is typologically quite frequent, cf. e.g. Elliott (2000). However, to the best of my knowledge, the specific interaction if irrealis-avertive with negation found in certain Australian languages (with e.g. (23) being ambiguous) is a typological ‘outlier’. ¹⁷ As noted in Phillips (forthcoming), it seems to be a sub-areal property of non-Pama-Nyungan languages that they tend to possess modal-irrealis negators (sometimes as inflectional paradigms), distinct from ‘indicative’ negators.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

121

Nordlinger and Caudal (2012) first identified these patterns in Murrinh-Patha, cf. (20)–(23): (20) ku beg mertthaka (Murrinh-Patha) ku beg me-art-dha-ka CLF:ANIM bag 1sgS.snatch(9).PST.IRR-get-IMPF-FOC ‘I should have brought my bag.’ (Nordlinger and Caudal 2012: 105) (21) ngay-dha ngatha-ka me-mawatha-dha-wa (Murrinh-Patha) 1sg-IMPF if-FOC 1sgS.hands(8).PST.IRR-rectify-IMPF-EMPH ‘If it had have been me, I would have rectified it.’ (Street and Street 1989) (22) be-lele-dha-wa (Murrinh-Patha) 3sgS.13.PST.IRR-bite-IMPF-EMPH (that carpet snake was coming towards him) ‘It was going to bite him (but didn’t)’. (Nordlinger and Caudal 2012: 106) (23) marda the-na-mut-tha NEG 2sgS.poke(19).PST.IRR-3sg.m.ben-give-IMPF palngun. (Murrinh-Patha) female a. ‘You didn’t give him that girl.’ b. ‘You shouldn’t have given him that girl.’ (Nordlinger and Caudal 2012: 106) Through collaborative work and field work¹⁸ on the modal systems of other non-Pama-Nyungan, I later discovered that Iwaidjan (Iwaidja (24)–(26)), Gunwinyguan (Anindilyakwa (27)–(28)) and Mirndi languages (Jaminjung, (29)–(30)) also have similar patterns; note in particular, that the remarkable ambiguity found in (23), seems to be common among non-Pama-Nyungan languages, as shown by e.g., (26) and (28): (24) ayana-wu-ni (Iwaidja) 1sg>3pl.PCF-hit-PCF ‘I was going to hit them/I nearly hit them’ (but I didn’t). (Caudal and Mailhammer 2021)

¹⁸ I am indebted to joint work conducted with Rob Mailhammer on Iwaidja and with James Bednall on Anindilyawka (including field work), and with Eva Schultze-Berndt on Jaminjung (see in particular Caudal and Schultze-Berndt 2016).

122

PATRICK CAUDAL

(25) yuw-ara-n mungu. Angana-mi-na (Iwaidja) 2sg.ANT.PROX-come-ANT not.know 2sg.PCF-say-PCF ngartung, ngana-lakbi-na yuw-ara. Ba OBL.1sg 1sg.PCF-know-PCF 2sg.PR.PROX-go-PR CONJ karlu ngana-mi-na nuwung. NEG 1sg.PCF-say-PCF OBL.2sg ‘You just came without letting me know. You should have let me know, because if I had known you were coming, I would have said ‘no’ to you.’ (Iwaidja dictionary) (26) karlu ayana-wu-ni NEG 1sg>3pl.PCF-hit-PCF 1. ‘I didn’t hit them.’ 2. ‘I should not have hit them’. (Caudal et al 2019)

(Iwaidja)

(27) n-akəna kə-rrak-aje=yedha chair=manja (Anindilyakwa) 3M-that IRR.3M-forehead-stand.PST=PURP NEUT.chair=LOC ekena dh-akəna dhədharrəngka yingmən-angma-Ø=dhə then 3F-that 3F.woman REAL.3F>VEG(?)-steal-USP=TRM akən chair=a NEUT.that NEUT.chair=PHRASE_FINAL ‘He was going to sit on the chair, but the woman took it away’ (Bednall 2020: 371) (28) nara n-akəna kenu-kwa-Ø a-rmdak-akəna NEG 3m-that PCF.3m/2-give-PCF neut-many-that angwarnda (Anindilyakwa) neut.money 1. He didn’t give you all that money 2. He shouldn’t have given you all that money. (Bednall 2020: 345) (29) yatha nga-b-irriga-na mangarra dempa alright 1sg:3sg- POT:COOK-IMPF plant.food damper damarlung (Jaminjung) nothing ‘I was going to/wanted to bake bread all right, damper, (but) nothing (i.e. I didn’t)’ (Schultze-Berndt 2000: 93)

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

123

(30) yagbali birdij gana-w-arra-nyi, (Jaminjung) place find 3sg:3sg-FUT-PUT-IMPF Buru ga-jga-ny Gurlugurlu waga ga-rdba-ny return 3sg-GO.PST sit 3sg-FALL-PST ‘he wanted to find a camp, he went back to Gurlugurlu and sat down (i.e. stayed there)’ (Schultze-Berndt 2000: 93)

6.3.2 Language sample, method and overview of avertive structures For want of space to address the huge diversity of forms existing across Australian languages,¹⁹ the survey offered here will be limited to a relatively small sample of seventeen languages, taken from seven non-Pama-Nyungan language families, and six Pama-Nyungan language families as in Table 6.1. Table 6.1 The language sample studied

non-Pama-Nyungan

Language family

Language

Iwaidjan

Iwaidja Mawng Anindilyakwa Kunbarlang Bininj Gun-wok Murrinh-Patha Jaminjung Wambaya Gurr-goni Nyikina Goonyiandi Kayardild Warlpiri Pilbara Yankunytjatjara Arrernte Arapana

Gunwinyguan

Daly River Mirndi

Pama-Nyungan

Maningrida Nyulnyulan Bunuban Tangkic Ngumpin-Yapa Ngarla Western Desert Arandic Karnic

Table 6.2 summarizes the results of the grammar mining study conducted on the sample, combined with field work results on two languages. It lists several avertivity-related morphosyntactic and semantic patterns found in the sample.

¹⁹ While a larger areal typological study has been conducted on a sample of 67 languages, its results are far too complex to discuss in such limited space.

Table 6.2 Avertive patterns in our language sample 1

2

3

4

5

6

7

8

9

10

11

Language family

Language

Synthetic MOD.PST avertive inflection

IRR.PST vs IRR.PR

Negative past events = NEG + IRR.PST

Past proximative/ volitional

Avertive PART/Clitic/ Afffix+ V-IRR

Avertive PART/Clitic/ Affix+ V-IND

REDuplication and iterative/ proximative (imperfective strategies)

V-AVERT + NEG construction

Pragmatic PC V-PST

Iwaidjan

Iwaidja

VIRR.PST

✓

karlu, arlarrarr

✓ ✓

wurrkany +V-IRR. PST/FUT, maju+VIRR.PST

RED-V-ANT

karlu/ arlarrar²⁰

V-ANT

Mawng

NF-VIRR2

✓²¹

marrik

✓

—

arlarrarr —

V-PP

Anindilyakwa

IRR-VPST/∅

✓

nara

✓

wurrkany +V-ANT maju+VANT wanji ‘nearly’+VANT wurkaj ‘nearly’+VPST —

—

nara/yanda

REAL-V-∅

Kunbarlang

VIRR.PST

✓²² ✕

ngunda (∗) —

✕

—

karlu karlu

—

Gunwinyguan

✓

=yedha, akwədhangwa ‘near(ly)’ yimarne(k) — + V-IRR.PST —

Bininj Gun-wok

V-IRR²³ CTRFAC VPST/IRR

✕

djama; minj; marrek —

✓

Daly River

MurrinhPatha Jaminjung

mere; manangka gurrany

✓

Mirndi

IRR.PST- ✓ VIRR.IMPF IRR:V✓ IMPF Aux✓ NACT.PST

✓

guyala³⁰

✓

Wambaya

kuyin-V ; ba(r)lanhV ²⁴; yimarnek ‘like’; djanggogo ‘for nothing’; djaying ‘mistakenly’ —

djaying, (yimarnek) kuyin/ba(r)lanhVPFV/IMPF

kuyin/ba(r)lanhV-IMPF V-REDIRR/PST²⁵

burrkyak/ larrh/ gayakki/ kurru, ²⁶

—

ngurdam may ‘try’²⁷

—

ma wurda²⁸

—

klosap ‘almost’ —

birri/ngarla — ‘try’ — —

damarlung²⁹

—

—

—³¹

Continued

Table 6.2 Continued 1

2

3

Language family

Language

Maningrida

4

5

6

7

8

9

10

11

Synthetic IRR.PST Negative Past proxiMOD.PST vs past mative/ avertive IRR.PR events volitional inflec= NEG + tion IRR.PST

Avertive PART/Clitic/ Afffix+ V-IRR

Avertive PART/Clitic/ Affix+ V-IND

REDuplication and iterative/ proximative (imperfective strategies)

V-AVERT + NEG construction

Pragmatic PC V-PST

Gurr-goni

V-IRR1³²

✓³³

galu

✓

—

—

galu ‘no, nothing’³⁴

—³⁵

Nyulnyulan

Nyikina

IRR-VPST

✓

mallu

✓

—

RED-PST³⁷

mallu

—

Bunuban

Goonyiandi V✓ IRR:PST³⁸

✕

✓

mundjarra ‘unrealized intention’ bulu, miliarry ‘almost’³⁶ wambawoo ‘nearly’

—

—

marlami, mangaddi,³⁹ marlami

—

Tangkic

Kayardild

Ngumpin- Warlpiri Yapa Billinara

Pilbara

Ngarla

Western Desert

Yankunytjatjara

(VAVERT)⁴⁰ CTRFAC1 V-POT⁴¹ CTRFAC 2 V-NF NEG.POT FUT V-IRR⁴⁴ V-IMP⁴⁷ VIRR.PST⁴⁸ VIRR.PST

✕

✕

✕ ✕ ✕

✕ ✕

✓

✕

✓

✓ ✕

✕

✓ ?

— (or: nginja,mar aka, ‘counterfactual, not what it seems’) nganta ∅ V-IRR,⁴⁵⁴⁶ nganda= (?), najing

—

—

✓

puttu V-INTENTINCHPST.IMPF

✓

(V✕ INTENTN )⁵³

✕ mirta/ ngurra (pirli) ✕

—

RED-NF⁴²

warirra⁴³

PST

—

—

—

—

RED-⁴⁹

kala lawa ‘but no’ najing⁵⁰ najing , ⁵¹

walyi, pilyparr, purtukarri⁵² puttu V-INTENTINCHPST.IMPF

RED

—

—

—

—

—

— —

Continued

Table 6.2 Continued 1

2

3

Language family

Language

Arandic

Arrernte

Karnic

Arapana

4

5

6

7

8

9

10

11

Synthetic IRR.PST Negative Past proxiMOD.PST vs past mative/ avertive IRR.PR events volitional inflec= NEG + tion IRR.PST

Avertive PART/Clitic/ Afffix+ V-IRR

Avertive PART/Clitic/ Affix+ V-IND

REDuplication and iterative/ proximative (imperfective strategies)

V-AVERT + NEG construction

Pragmatic PC V-PST

AuxPURP V.PST

✓

—

RED-elpe-

—

—

—

✕

-elpe‘nearly V’, uyarne V-PST, ingkwe V-PST panta V-PST+PR, V-PST panta-li, walyiliwalyili ‘nearly’⁵⁶

panta-li

—

tyekenhe⁵⁴ ✓

✕

✕

—

V,⁵⁵

²⁰ Karlu in Iwaidja can be both a sentential negation, and a negative interjection (‘no’). Cpr. Bininj Gun-wok djama (sentential negation) vs burrkyak (‘no, nothing’) cf. Evans (2003: 603–604). ²¹ The Mawg irrealis 2 (I2) inflection can have both present and past priority modal (Portner 2018) readings; it is otherwise a past irrealis. ²² Note that some younger speakers seem to accept forms combining IRR.PST prefixes with IRR.PR suffixes, cf. Kapitonov (2019: 178).

²³ The Bininj Gun-wok irrealis morphology is temporally ambiguous. ²⁴ Cf. Evans (2003: 525, 611). ²⁵ The so-called ‘inceptive’ reduplication can have avertive readings (‘tried but failed’) with irrealis or past tense-marked verbs in Bininj Gun-wok, cf. (Evans 2003: 374, 381). ²⁶ Cf. Evans (2003: 306). ²⁷ Cf. Ford and McCormack (2007: 7). ²⁸ Cf. Ford and McCormack (2007: 17); Nordlinger and Caudal (2012: 106). ²⁹ Cf. Schultze-Berndt (2000: 93). ³⁰ IRR marking is compulsory with NEG guyala, but not with NEG yangula. ³¹ Nordlinger (1998: 295) mentions a possibly avertive particle (yurubu ‘for nothing’), but does not provide examples illustrating such uses. ³² Wurru/wurpu ‘just, only, except, but’ often introduce a clause ‘cancelling’ a previous IRR1 clause describing an impeding event (Green (1995: 295–296)); this is an adversative construction. ³³ IRR1 corresponds to a ‘precontemporary’ irrealis in the Gurr-goni scalar temporal system, close to past irrealis. Gurr-goni also has an IRR2 paradigm, which appears to function as a ‘contemporary’ (= recent past, ‘extended present’/present irrealis) paradigm—pace Green (1995) who claims the contemporary/precontemporary distinction is neutralized for the irrealis. ³⁴ Although IRR1 + galu cannot be found in Green’s grammar, reduced negative clauses with galu can (Green (1995: 111)). I’m therefore extrapolating those should be possible. ³⁵ Past realis clauses in Rembarrnga can be followed by the wapa ‘in vain’ avertive particle, see McKay (1975: 258). ³⁶ Cf. Stokes (1982: 281, 373). ³⁷ Cf. Stokes (1982: 43). ³⁸ IRR:PST corresponds to IRR:POT (-yi/ -wi ~ -rni) in McGregor (1990: 220). ³⁹ McGregor (1990: 348, 583). ⁴⁰ The Kayardild -nangarra (‘almost’) inflection suffix does not seem to have other modal meanings (Evans 1995: 261); it is a unique type in the sample. ⁴¹ Karyardild seems to offer one synthetic polyfunctional irrealis ((NEG.)POT ) and two polyfunctional periphrastic irrealis-avertives: maraka (‘counterfactual1 ’) + (NEG.)POT , nginja (‘counterfactual2 ’) + (NEG.)POT , but for want of space, I will not be able to discuss the Kayardild modal system at great length here. ⁴² Cf. Evans (1995: 290, ex. 7–78). ⁴³ Evans (1995: 341) even seems to offer an example where the warirra negator (‘nothing, empty’) appears before the verb—but this could be an informationally complex utterance. ⁴⁴ Legate (2003: 157) gives an example with the ngarra FUTure auxiliary, whereas Laughren (2002) rather mentions the null auxiliary. ⁴⁵ Both Simpson (2012: 36) and Legate (2009) mention a nganta ∅V.IRR construction, with nganta being a hypothetical/evidential particle. ⁴⁶ One should also mention the related ‘mistaken thought’ construction kula-nganta(‘NEG-HYP’) FUT V-IRR, cf. kula-nganta-kapi-rna wawirri panti-ka-rla, kala lawa (NEGseem-FUT-I kangaroo spear-IRR but no) ‘I thought I was going to spear the kangaroo, but I didn’t.’ (Nash 1980: 239). It shows that Warlpiri also has a FUT V-IRR proximative with avertive effects. ⁴⁷ Although (Meakins and Nordlinger 2014)’s label for this inflection is IMPerative, its semantics is clearly broader than that of a ‘priority’ modal (see Portner 2018). ⁴⁸ The present irrealis in Bilinarra combines the IMP (i.e. IRR) suffix with HORtative -rla. The past irrealis-avertive is realized by adding the DUBitative clitic = nga to the irrealis.

⁴⁹ Cf. Meakins and Nordlinger (2014: 162, ex. 417). ⁵⁰ Sentential negation does not seem to appear in the pattern; instead, we find najing (‘nothing’), a loanword from English (via Gurinji (Kriol), possibly). ⁵¹ In this pattern too, the Bilinarra negator lawara ‘nothing’, cognate with Warlpiri lawa ‘no’, is not found in Meakins and Nordlinger (2014). ⁵² Purtukarri is obviously cognate with the Western Desert avertive capacitative (‘couldn’t’) particle puttu, and walyi with the Karnic proximative (‘nearly’) walyi. ⁵³ Yankunytjatjara -kitja INTENT is a nominalizing derivational volitional suffix, with avertive implicatures; it has no fixed temporal content. Pintupi-Luritja has a similar form (Goddard 1985:163, Rose 2001:70–71). It also appears in the semanticized avertive construction puttu (‘in vain’) + V with a verbalizing suffix (V-kitja-INCH-PST.IMPF )—the latter construction is therefore verbal. ⁵⁴ The Arrernte verb negator tyekenhe is obviously cognate with the PURP (-tyeke) inflection, so it could actually be a negative modal (irrealis) form too. ⁵⁵ The meaning of that special reduplicated verb pattern seems to be proximative (Wilkins 1989: 261–262), but can have strong (and possibly mildly conventionalized) avertive implicatures. ⁵⁶ Derived from walya ‘soon, directly’ (Hercus 1994: 216); negative adverbial panta-li is similarly derived from panta ‘not, in vain’.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

131

Columns 3–6 serve to identify what I will dub ‘irrealis-avertive cluster paradigms’, i.e. forms exhibiting (at least part of ) the polyfunctionality identified in Section 6.3.1: column 3 lists the forms combining an avertive meaning (often avertive implicatures) with at least two other modal meanings; column 4 specifies whether the language has morphologized the distinction between irrealis present vs past; column 5 indicates whether irrealis-avertive forms have present vs past counterparts; column 6 indicates whether or not they have volitional/proximative meanings. All the remaining columns (6–11) are associated with other types of avertive patterns—or avertive strategies, in the case of column 11, as we will see below. Last but not least, I left out from this study lexical avertives (‘fail’) such as (31); these are frequent (e.g. Mawng has several specialized lexical avertives),⁵⁷ and as such also reflect on a cognitively pervasive category, but I am focusing on grammatical devices alone here. (31) ŋai caŋkaatti ŋuɲca-nna ŋuji-nna (Kalkatungu) I here fail-PST fall-PST ‘I nearly fell’ (or ‘I escaped from falling’) (Blake 1979a: 61)

6.4 Discussion and empirical generalizations Let us now discuss the facts summarized in Table 6.2 and offer some tentative empirical generalizations concerning the form and meaning of avertive patterns in Australian languages. Several recurrent morpho-syntactic types of lexico-grammatical structures were found in the sample, and are of particular relevance to the present study, cf. (32): (32) Tentative list of types of avertive structures 1. Synthetic modal inflections, especially irrealis/potential/future, but also narrower modal inflections, such as ‘PURPosive’ (in Arrernte) or ‘INTENTive’ (in Yankunytjatjara) endowed with more or less clear avertive implicatures (some border on the conventional, already) even without any additional material—see columns 3–6 2. Periphrastic irrealis/avertives, combining an inflectional element (modal or not), and a dedicated modal-avertive element (as pre- or postverbal particles, clitics, affixes/infixes, or a combination of several ⁵⁷ In some Australian languages though, lexical/constructional are not known to exist, or at least were not elicited in past field work studies, cf. e.g. Wubuy (Heath 1984: 345).

132

PATRICK CAUDAL

such elements) (columns 7 and 8). These generally semanticized avertive patterns are typically found in non-Pama-Nyungan languages in our sample, but also in neighbouring Pama-Nyungan language families (Tangkic, Ngumpin-Yapa, Arandic, Karnic, and Western Desert). Among indicative inflections, both imperfective, and aspectually underspecified (perfective/imperfective) markings seem to be found. (Admittedly, one can discuss whether all these data points effectively constitute instances of periphrastic avertive morphology, but for the sake of simplicity, I will assume that they do.) 3. REDuplication and pluractional imperfective-based patterns (column 9), without additional morphological markers, but frequently in combination with special ‘durative’ intonations—this a strongly aspect-driven type of avertive pattern; it can be conventionalized as a dedicated avertive, or a ‘pragmatic strategy’ to convey avertivity. 4. Any of the above avertive structures can combine with negative markers in order to form what I will claim to be conventionalized, reduced biclausal avertive constructions—see column 10. 5. ‘Partitive culminations’ (a term I’m coining after Martin and Demirdache (2020)’s partitive accomplishments, to generalize beyond so-called non-culminating accomplishments) utterances (see column 11), i.e. telic indicative utterances marked with a tense receiving a ‘weak’ perfective reading, without clear culmination or stable results—‘X V-ed but did not finish V-ing/failed to V’. I will successively discuss below each of these avertive patterns (Sections 6.4.1– 6.4.5), before briefly addressing the issue of aspectual, temporal and modal parameters underlying avertivity in the sample (Section 6.4.6).

6.4.1 Avertive pattern no. 1: Synthetic inflectional past modal avertives and the (typically non-Pama-Nyungan) irrealis-avertive cluster Avertive patterns involving a (synthetic) past modal inflection are one of the two predominant structures in the sample, with sixteen languages out of seventeen offering at least one such pattern (column 1 in Table 6.2). The present study also confirmed that the type of polyfunctionality identified in Section 6.3.1 is found in all non-Pama-Nyungan language families of our sample, with some limited variation: Bininj Gun-wok and Mawng standout

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

133

with respect to the temporal anchoring of the irrealis; it is completely neutralized in the Bininj Gun-wok system (the irrealis inflection can have both present and past anchoring) and partially neutralized in Mawng (the IRR2 inflection can have present anchoring under its priority ⁵⁸/deontic readings). Examples of the irrealis-avertive cluster in Gunwinyguan (cf. Kunbarlang (33)–(34) and Bininj-Gunwok (35)–(38)),⁵⁹ Bunuban (Goonyiandi (39)–(41)), Maningrida languages (Gurr-goni (42)–(45)) and Nyulnyulan languages (Nyikina (46)–(48)) are given below:⁶⁰ (33) ngunda ngay-buddu-wuni. (Kunbarlang)⁶¹ not 1sg.IRR.PST-3pl.O-give.IRR.PST ‘I didn’t give it to them.’ (Kapitonov 2019: 8) (34) nguddu-yung mandjang ki-nguddu-bu. (Kunbarlang) 2pl.IRR.NP-lie.IRR.NPST perhaps 3sg.IRR.NPST-2pl.O-hit.IRR.NP ‘If you sleep [in the house] you might get hurt.’ (Kapitonov 2019: 188) (35) yi-man.ga-yi. 2-fall-IRR ‘You nearly fell.’ (Evans 2003: 373)

(Bininj Gun-wok)

(36) a-rrowkme-ninj / a-bu-yi. 1/3-shoot-IRR 1/3-hit-IRR ‘I nearly shot it/nearly hit it.’ (Evans 2003: 373)

(Bininj Gun-wok)

(37) larrk, marrek Mardayin birri-bimbuyi, ya nothing NEG [ceremony] 3AUGP-paint-IRR yeah na-djamun (Bininj Gun-wok) MA-sacred ‘No, they didn’t paint Mardayin ceremony designs, they are sacred.’ (Evans 2003: 282) ⁵⁸ In the sense of Portner (2018). Imperatives are semantically performative modals, whereas deontic are only contextually, indirectly performative modals. Many so-called ‘imperatives’ in descriptions of Australian languages rather seem to be deontic modals—or even semantically broader modals, with deontic uses, and contextually performative deontic (imperative/hortative-like) uses. In our sample, see e.g. the so-called ‘imperative’ in Bilinarra (Meakins and Nordlinger 2014), which I here claim to be an irrealis/modal inflection. ⁵⁹ I must thank Murray Garde for pointing out to me that the kind of ambiguity exhibited by (23) also existed in the Bininj Gun-wok modal system. ⁶⁰ I will not attempt to provide a more thorough inventory of non-Pama-Nyungan language families exhibiting the same polyfunctionality for irrealis-avertive inflections, but similar observations can also be made about several others. ⁶¹ As opposed to Bininj Gun-wok, most of the attested avertive utterances in Kunbarlang appear to require yimarne(k). I will leave Kunbarlang in that list for the sake of prudence, though, as it could be that avertive readings of the irrealis past inflection are simply undocumented.

134

PATRICK CAUDAL

(38) bi-ma-yi Na-burlanj gun-mak. (Bininj Gun-wok) 3/3HP-marry-IRR MA-[skin] IV-good ‘She should have married straight, to a Naburlanj man.’ (Evans 2003: 375) (39) yaanya thangarndi jaggilimirni nyinlimi (Goonyiandi) other word I:might:have:said I:forgot:it ‘I was going to tell another story, but I forget it.’ (McGregor 1990: 534) (40) manyi yan.ginngindi wardgilarninganggi food you:asked:me I:might:have:brought:it:for:you marlami (Goonyiandi) not ‘You asked me for bread. I should have brought you some [but I didn’t]’. (McGregor 1990: 535) (41) wardngirni milaalarni (Goonyiandi) I:might:have:gone I:might:have:seen:him ‘(Had you told me) I would have gone and seen him.’ (McGregor 1990: 534) (42) weleng galu awurr-beki-ya+rni. njiwurr-ni-Ø then NEG 3AUGS-arrive, come out-IRR1 1AUGS-sit-PRE njiwurr-rruwdjiyi-ni. 1AUGS-cry-PRE ‘Then they didn’t come. We sat (and) cried.’ (Gurr-goni) (Green 1995: 341) (43) wurru at-gardi nji-na-ga-tji-rni but 3I-flesh 2MINA.3MINO-PROX-take-IRR1 ngapala. (Gurr-goni) lMIN+DAT ‘But you might have brought some meat for me.’ (Green 1995: 196) (44) gi-yini-gi+rni ngu-bogi-ya+rni Nangak, [3IVS-do thus-IRR1] 1MINS-go-IRR1 place.name worro. (Gurr-goni) what.a.pity ‘I was going to go to Nangak today, what a pity.’ (Green 1995: 196)

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

135

(45) maka dji-na-djeka-nga+rni, nguwurr-bogi-ya+rni. (Gurr-goni) FaMo 3IIS-PROX-go back-IRR1 1+2AUGS-go-IRR1 ‘Your grandma could have come back, (so) we could all have gone.’ (= if your grandma had come back, we would all have gone) (Green 1995: 196) (46) ŋ̇ŋa-la-MA-na-dyi⁶² miliya mallu ŋ̇a-la-MA-na (Nyikina) lsg-IRR-go-PST-EXP now NEG lsg-IRR-go-PST ‘I was going to go this morning but I didn’t go’ (Stokes 1982: 281) (47) malu ŋ̇a-l(a)-ANDI-ny-dyina ginya-yi NEG lsg-IRR-pick.up-PST2-3sgDATPRO DEM-DAT ‘I didn’t get (it) for that one!’ (Stokes 1982: 69)

(Nyikina)

(48) ya-la-(rr)-DI-na-da mabu (Nyikina) lnsg-IRR-(NMIN)-sit-PST-HABIT good ‘we should have been good (but we weren’t)’ (Stokes 1982: 281)

6.4.2. Avertive pattern no. 2: Periphrastic irrealis/avertives Let us turn now to avertives based on a combination of a dedicated avertive particle or clitic, plus some inflectional element—whether indicative or irrealis (i.e., modal). Patterns of that type are also widely represented in our study, with sixteen languages offering them, cf. columns 7 and 8 in Table 6.2. Periphrastic avertive inflection exhibiting an indicative past marking (whether imperfective, or aspectually underspecified) are slightly less common in our sample, than irrealis-marked patterns, cf. e.g. (49)–(50). (49) mundjarra ngu-rra-dji+rni wurru warrpura (Gurr-goni) AVERT 1MINA.3MINO-shoot-IRR1 but underarm.sweat gu-numi-rri ngapala. 3MINA.3IVO-smell-PRE 1MIN+DAT ‘I tried to/was going to shoot it, but it smelt my sweat (and ran off ).’ (Green 1995: 314) (50) miliyarri dumarra ŋ̇a- l(a)-ANDI-na almost take off lsg-IRR-pick up-PST ‘I nearly took off ’. (Stokes 1982: 281)

(Nyikina)

⁶² Stokes (1982: 14) notes ŋ̇ what he describes as a ‘nasal dorsal’, but it’s unclear what kind of velar nasal it is exactly—possibly just /ŋ/.

136

PATRICK CAUDAL

As a matter of fact, synthetic irrealis-avertive inflections as listed under Section 6.4.1 frequently serve to form periphrastic avertive patterns, by combining with special particles or clitics, usually modal-avertive or proximativeavertive (‘nearly/almost’). Whether these are specifically proximative adverbials/particles (‘nearly, almost’), or specifically modal/avertive markers, is an obviously complex question. I will simply treat them as elements of periphrastic avertives, since it is often quite difficult to make an informed decision based on a handful of examples in a grammar. While these periphrastic patterns are most salient in languages not possessing (or not fully possessing) the irrealis-avertive cluster—i.e. Pilbara, Arandic, Western Desert, and Karnic languages of our sample—they are also attested in languages possessing it, e.g. in non-Pama-Nyungan languages, cf. e.g. Nyikina (50), Gurr-goni (51), and Iwaidja (52). This results in languages possessing multiple inflectional forms capable of conveying avertive interpretations, with the periphrastic form semantically encoding avertivity, vs the synthetic form pragmatically implicating it, in most cases. (51) mundjarra ngu-rra-dji+rni wurru warrpura (Gurr-goni) AVERT 1MINA.3MINO-shoot-IRR1 but underarm.sweat gu-numi-rri ngapala. 3MINA.3IVO-smell-PRE 1MIN+DAT ‘I tried to/was going to shoot it, but it smelt my sweat (and ran off ).’ (Green 1995: 314) (52) Maju ana-ma-nyi. VOL 1sg>3sg.PCF-take-PCF ‘I was going to take it but didn’t.’

(Iwaidja) (Iwaidja dictionary)

A first subtype of pattern no. 2 involves a dedicated avertive negative particle with a ‘tried (in vain)’ negative meaning in the past, cf. e.g. ngarla in Ngaliwurru (a dialect of Jaminjung) (53),⁶³ or, more commonly, a semantically negative avertive particle such as pilyparr (‘in vain’) in Ngarla (54). A second subtype of indicate past avertive involves a proximative, de facto semantically negative proximative particle with a ‘nearly/almost’ meaning, cf. e.g. walyi in Ngarla (55).⁶⁴ And finally, a third type is illustrated in the sample (again) by puttu (‘can’t’) in Yankunytjatjara (56), a clearly modal particle which can ⁶³ Note that like (30), (53) offers an instance of a non-avertive, proximative reading of a verb (gan-jibunga-nyi) in the POT-IMPF (IRR.PST) inflection; it also illustrates that avertive meanings are contextually determined for such polysemous inflectional forms). ⁶⁴ Nyangumarta, a related Pilbara language, possesses two similar (indicative) avertive particles: partal ‘in vain’ and katu(rr)/kartungurru ‘nearly’ (cf. Sharp 2004: 133, 181).

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

137

combine with the (aspectually underspecified) past tense to convey avertive meanings (alongside with other modal or postmodal meanings, including a negative capacity meaning, according to James Grey, p.c.)—note that Ngarla possesses a ‘for nothing, for no reason, unwillingly’ purtu-karri particle, which is obviously derived from a related purtu particle. (53) yugung gan-jib-unga-nyi, (...), ngarla wilng (Ngaliwurru) run 3sg:1sg-POT-LEAVE-IMPF TRY stay.back nga-ngu 1sg:3sg-GET/HANDLE.PST ‘he was going to run away from me, (...) but I tried to hold him back’ (Schultze-Berndt 2000: 479) (54) pilyparr ngaja yarni+ma-rnu pirrjarta. (Ngarla) unsuccessfully 1sg.ERG repair[+CAUS]-PST vehicle ‘Unsuccessfully I repaired (the) vehicle.’ (I.e. ‘I failed to repair the vehicle.’ ) (Westerlund 2015: 75) (55) nyinta walyi wakurr ja-rnu ngunyi karlajangu. (Ngarla) 2sg.ERG almost secure CAUS-PST DEM (distant) cattle ‘You almost had that cattle (i.e. cow/bull) secured (i.e. yarded up).’ (Westerlund 2015: 175) (56) ngayulu puttu nya-ngu (Yankunytjatjara) lsg(ERG) IN.VAIN see-PAST ‘I couldn’t see/find it’ [= I tried to, but in vain] (Goddard 1983: 247) Most of the particles involved in periphrastic avertives seem inherently, semantically negative (including proximatives). This is clearly the case of the Western Desert puttu (‘can’t/couldn’t/in vain’) particle, which combines with a derived INCHoative verb formed on a modal noun (through the nominalizer -kitja- INTENT) (57), and of the Warlpiri kula-nganta (NEG+MOD) particle (58), which requires a Aux.FUT V -IRR complex irrealis inflection on the verb (and is fused with the Aux element). The latter avertive has clear negative evidential, so-called ‘mistaken thought’ undertones (‘things were not what they seemed’)—I will come back to this later in this chapter. But again, what matters most here is that the semantic status of the avertive readings one can generally assign to such periphrastic forms contrast withs the pragmatic nature of avertive interpretations associated with (most, if not all) synthetic past modal inflections in our sample, as shown in Section 6.4.1.

138

PATRICK CAUDAL

(57) ngayulu puttu wangka-kitja-ri-ringa-ngi (Yankunytjatjara) lsg(NOM) IN.VAIN talk-INTENT-INCHO-PAST.IMPF ‘I wanted to talk with them in vain’ (e.g. they wouldn’t listen) (Goddard 1983: 131) (58) Kula.nganta-kapi-rna wawirri panti-ka-rla, (kala lawa). (Warlpiri) NEG.MOD-FUT-1sg kangaroo spear-IRR (but no) ‘I thought I was going to spear the kangaroo, but I didn’t.’ (Nash 1980: 239) As I have noted above, the sort of polyfunctionality I dubbed in Section 6.4.1 ‘the irrealis-avertive’ cluster, also seems to associate with periphrastic irrealisavertives. It notably permeates part of the Pama-Nyungan zone, especially Ngumin-Yapa, Tangkic and (to a lesser extent) Pilbara languages, as their periphrastic IRR.PST forms exhibit most traits of the non-Pama-Nyungan irrealis-avertive cluster (columns 3–6)—but it does not seem to extend to Arandic, Western Desert, and Karnic languages (these lack either negative past events encoded by NEG+PST.IRR, and/or a volitional/proximative PST.IRR). Kayardild (Tangkic) is such a case of Pama-Nyungan irrealis-avertive cluster, with two distinct avertive patterns. The first of these two patterns is the irrealis-proximative periphrasis V-nangarra ‘almost V/would have Ved’.⁶⁵ It is compatible with both agentive and non-agentive verbs; it selects the modal ABLative case, as in (59)–(60)—hence its periphrastic nature. (59) bulkurdudu ngijin-jina baa-nangarra kurthurr-ina (Kayardild) crocodileNOM lsgPOSS-MABL bite-AVERT shin-MABL Ά crocodile almost bit me on the leg.’ (Evans 1995: 261) (60) [Of a man crushed by a falling tree:] niya budii-nangarr, [warirra-ntha barji-n-marri-nja (Kayardild) 3sgNOM run-AVERT nothing-COBL fall-N-PRIV-COBL niwan-jinaa-nth] COBL3sg-MABL-COBL ‘He just about got away, then nothing would have happened, it wouldn’t have fallen on him.’ (Evans 1995: 261) The more polyfunctional counterfactual particle nginja⁶⁶ combines with the indicative ‘non-future’ (here, contextually past) tense (V-ACT) to form a wider periphrastic irrealis-avertive cluster with a range of meanings similar to those found with e.g. the Iwaidja past irrealis-avertive, namely (negative) ⁶⁵ Glossed as ‘ALMOST’ in Evans (1985, 1995). ⁶⁶ Glossed as ‘FRUSTR(ative)’ in Evans (1985, 1995).

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

139

past/present deontic (‘X shouldn’t have V-ed/shouldn’t V’), cf. (61)–(63). A nginja + V-POT ‘X will V for nothing’ meaning also seems to be attested (see Evans (1995: 383)). Furthermore, as it is able to combine with nominalized verbs (64), it is reasonable to assume that nginja can be analysed as a verbal auxiliary-like element, similar in this case to the verbal head of a light verb construction. Treating such modal particles as modal auxiliary-like predicates is consistent with diachronic processes whereby similar elements morphologized as part of verbal portmanteau TAM prefixes in non-Pama-Nyungan languages; see Osgarby (2018). (61) barruntha-y duruma-th, nginja ngumu-wa-th, (Kayardild) yesterday-LOC lie-ACT CTRFAC2 black-INCH-ACT nginja kamburi-ja muma-th, ja-warri CTRFAC2 speak-ACT thunder-ACT rain-PRIV ‘(The weather) lied yesterday. In vain the sky blackened, in vain the thunder spoke, there’s no rain.’ (Evans 1995: 382) (62) nginja diya-ja mala-y (Kayardild) CTRFAC2 eat-ACT beer-MLOC ‘(You schoolkids) shouldn’t have drunk that beer.’ (Evans 1995: 383) (63) niwan-juru ngada nginja wirdi-j (Kayardild) him-PROP lsgNOM CTRFAC2 stay-ACT ‘I waited around for him for nothing (he didn’t turn up).’ (Evans 1995: 340) (64) nginja rikarrkati-n-da kularrin-d CTRFAC2 cry-N-NOM brother-NOM ‘Your brother shouldn’t be crying.’ (Evans 1995: 340)

(Kayardild)

Turning now to non-Pama-Nyungan languages, Bininj Gun-wok offers one of the most extended periphrastic avertive system in our sample. Its most prominent periphrastic avertive pattern involves the counterfactual yimarnek (originally a similative (‘like’) marker), combined with (past) irrealis marking, (65)–(66). (65) yimarnek kam-ra-yinj la Ngarridj (Bininj Gun-wok) CTRFAC 3.PROX.PST-go-IRR CONJ [subsection] bi-rrahme-ng. 3/3HP-block-PFV ‘She was going to come but the Ngarrij wouldn’t let her’ (Evans 2003: 611)

140

PATRICK CAUDAL

(66) yimarnek nga-rrulubu-yi, la ∅-bid-deyhme-ng, minj CTRFAC 1/3-shoot-IRR but 3-hand-click-PP not ∅-dowkme-ninj. (Bininj Gun-wok) 3P-go.off-IRR ‘I tried to shoot but the trigger just clicked without it (the gun) discharging.’ (Evans 2003: 611) The Bininj Gun-wok irrealis inflection seems to be one of the rare instances in our sample, of a temporally ambiguous irrealis—i.e. it can have both present and past temporal modal meanings. However, it is uniformly past in such periphrastic irrealis-avertive structures. This fact suggests a strong correlation between avertivity and pastness. On top of the yirmarnek periphrasis, Bininj Gun-wok possesses a second irrealis-avertive periphrasis involving counterfactual particle maraka plus a POT-marked verb. Although the POT inflection normally has present temporal anchoring, such structures uniformly anchor to the past; this also suggests a strong correlation between irrealis-avertive meanings and pastness. Maraka + V-POT forms a broad irrealis-avertive cluster, with readings ‘should have, could have (but didn’t)’, ‘would have, was going to, meant/wanted to (but didn’t’)’, cf. (67)–(70). In addition to POT verbs, maraka can combine with non-future (ACT)-inflected verbs, and then has ‘pretend’ (related to evidential ‘not what it seems’ uses) (see Evans (1995: 378). Like yimarnek, the original meaning of makara is also similative (‘like’). (67) maraka yuuma-thu barruntha-y (Bininj Gun-wok) CTRFAC1 drown-POT yesterday-MLOC ‘He could have drowned yesterday (but didn’t).’ (Evans 1995: 379) (68) kilda maraka diya-nangku mala-wu (Bininj Gun-wok) 2pl.all CTRFAC1 drink-NEG.FUT beer-MPROP ‘(You schoolkids) shouldn’t have drunk that beer.’ (Evans 1995: 383) (69) yakuri-wu maraka kurdala-thu, maraka fish-MPROP CTRFAC1 spear-POT CTRFAC1 maku-nku⁶⁷ (Bininj Gun-wok) use.bark.torch-NEG.POT kurdala-thu yakuri-wu. spear-POT fish-MPROP ‘he had meant to spear fish, to spear fish using a bark torch’ (Evans 1995: 722) ⁶⁷ Because of this NEG.POT form; I suspect the rendering of this example should rather be: ‘he had meant to spear fish, to spear fish without using a bark torch’—indeed the agent got lost in the fog as a result of not seeing anything.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

141

(70) maraka birdiru-thu! (Bininj Gun-wok) CTRFAC1 miss-POT ‘(They said) I was going to miss, but I didn’t!’ (Evans 1995: 654) Iwaidja offers the second most developed periphrastic avertive in our sample, with periphrases maju+V -PST and wurrkany + V -PST/FUT ,⁶⁸ (cf. (71)–(74)). One should also mention the Kunbarlang yimarnek + V-IRR.PST irrealis-avertive structure, (75), cf. (Kapitonov and Gentens 2018); it is clearly related to the Bininj Gun-wok yimarnerk + V-IRR irrealis avertive. (71) maju ngana-ngiru-nyi. VOL 1sg.PCF-board-PCF ‘I was going to get in the car (but didn’t).’

(Iwaidja) (Iwaidja dictionary)

(72) maju birdirlkbu-ny. Nganduka a-bi-ny? (Iwaidja) VOL 3sg.ANT-struggle.fre-ANT INT 3sg.ANT-do-ANT? ‘He tried to struggle free but in vain.’ (lit. ‘but for what?’) (Iwaidja dictionary) (73) wurrkany yan-ara karlu a-rtirra-n. (Iwaidja) FRUST 3sg.DIST.FUT-go-FUT NEG 3sg.ANT-come.back.ANT ‘He was going to go/tried to go, but (no,) he came back.’ (Iwaidja dictionary) (74) wurrkany a-wuku-ng ba walij (Iwaidja) FRUST 1sg>3sg.ANT-give-ANT DET food rardudban 3m.sg>3sg.ANT-leave.behind-ANT ‘I tried to give him food but he left it behind.’ (Caudal and Mailhammer 2021) (75) na-buk yimarnek ki-buddu-karlkkangki (Kunbarlang) I-person CTRFAC 3sg.NEG-3pl.O-stalk.IRR.PST la kadda-rnay la kadda-bum. CONJ 3pl.NF-see.PST CONJ 3pl.NF-hit.PST ‘He was going to sneak up on them, but they saw him and beat him.’ (Kapitonov 2019: 291)

⁶⁸ Maju is derived from root maju ‘want’, so it is a clear volitional modal, while wurrkany suspiciously looks like an intransitive 3p. past verbal form (w-urrka-ny); it could possibly be related to the -burrkan(tr.) ‘dream’ root, which would be in line with its negative evidential, hence avertive meanings.

142

PATRICK CAUDAL

In Kayardild, Bininj Gun-wok, Iwaidja, and Kunbarlang, these patterns involve a modal particle with a future/present irrealis or indicative past inflected verb, offering at least two modal/evidential readings on top of an avertive reading.⁶⁹ The latter property is key to treating them as irrealisavertive periphrases, coexisting with non-periphrastic irrealis-avertive forms (the IRR.PST paradigm in Iwaidja, the V-IRR form in Bininj Gun-wok, and the POT inflection in Kayardild). Note however that in Kunbarlang, the V-IRR.PST synthetic irrealis seems not to have avertive readings (any more?). In Bininj Gun-wok, Iwaidja, and Kayardild (but not in Kunbarlang),⁷⁰ periphrastic irrealis-avertive inflections logically are the marked members of these complex irrealis systems.⁷¹ I believe this partly explains an important contrast between synthetic, vs (negative) particle-based irrealis-avertives: the former seem to (often) encode avertivity as a mere implicature (type 1 of avertivity), while with the latter, avertivity seems to be (often) semantically conveyed (type 2). Jaminjung examples (29) vs (30) are a perfect illustration of the defeasible, conversational implicature status of the avertive reading associated with at least some synthetic past irrealis-avertives; context can defeat the ‘failure’ implicature (as is clearly the case in the latter example) in some languages endowed with such forms. This being said, depending on languages, it is not always easy to get speakers to accept such cancellations of failure implicatures; in Iwaidja, dedicated fieldwork suggested that many speakers harbour at least a certain hesitation to accept a non-failed reading of an IRR.PST-marked avertive.⁷² While further investigations are obviously necessary, there already seems to be variation in our sample w.r.t. synthetic irrealis inflections (with some, the avertive interpretation seems to be more pragmatically defeasible than with others).

⁶⁹ Wurrkany + V-FUT/IRR/PST can have evidential, proximative, volitional and avertive meanings in Iwaidja; in Kayardild, nginja + V-ACT can have past irrealis deontic readings (‘should have’), as well as avertive readings (‘tried and failed’). In addition to this, jginja + V-FUT can have negative predictive (‘won’t) readings. ⁷⁰ Yimarnek + V-IRR.PST Kunbarlang utterances seem to require a reduced negative clause construction to receive their full avertive reading—this suggests the periphrastic inflection has grammaticalized further, i.e. is less a markedly avertive form, and might be on the verge of replacing V-IRR.PST as the general past irrealis—proof of this can be found in the fact that only the periphrastic form seems to have proximative/volitional meaning; its modal range of meaning is in fact already larger than that of the synthetic V-IRR.PST form. ⁷¹ Unsurprisingly, the Kayardild nangarra monofunctional avertive inflection also seems to semantically encodes avertivity; judging from the examples given by N. Evans, it is unnecessary to strengthen the implicature via a subsequent clause. ⁷² Interestingly, Iwaidja also possesses periphrastic modals, combining a modal particle with a variety of tense marking. With some of them (esp. wurrkany, cf. (73)–(74)), the avertive meaning is very clearly semantic and non-defeasible—much more so than the avertive interpretation of the IRR.PST inflection is.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

143

But besides being a marked form, all the above periphrastic irrealis-avertive inflections incorporate a particle whose meaning is counterfactual/similative, and generally negative (e.g. ‘not what is seems’). This, I will argue, largely explains why these forms convey negative past events. In contrast, synthetic irrealis avertives seem to have (at most) negative implicatures; this also partly explains why they have developed negative past event meanings.⁷³ Possessing an overt negative content is, we will see, a necessary ingredient of bona fide semantic avertives. This explains why Kayardild stands out (again!) for possessing the only clearly semantic avertive based on a synthetic inflection, namely NEG.POT under its negative capacity reading. See (76), which effectively means that a vain attempt at finding (and killing) someone took place. (76) [context: ‘them mob’ try to find someone so as to spear and stab him] kaba-nangku, kuru-lu-nangku niwan-ju (Kayardild) find-NEG.POT dead-FAC-NEG.POT 3sg-MPROP ‘But (they) couldn’t hit home, couldn’t kill him.’ (Evans 1995: 581) As we have already seen, a number of combinations of (apparently)⁷⁴ proximative adverbials, particles, clitics, and prefixes or suffixes glossed ‘almost, nearly’, also appear in the sample, cf. e.g. wanji in Iwaidja, klosap in Murrinh-Patha, kuyin-/ba(r)lanh-V-PST/IRR in Bininj Gun-wok, akwədhangwa (‘near(ly)’) in Anindilyakwa, bulu and miliyarri in Nyikina, wambawoo in Gooniyandi (77), the two cognate suffixes -alpa/-elpe in Arapana vs Arrernte, etc. (cf. column 8 in Table 6.2). It should also be mentioned that Arapana has an additional avertive particle, panta, glossed by L. Hercus as ‘in vain’, combining with past tense marking, and which seems to have a distinct proximative meaning close to ‘hardly’ (Hercus 1994: 238)—it very much looks like a case of proximativeturned-avertive particle. Most of these proximative markers seem to associate with either perfective (or perfective uses of ) indicative past tenses. However, ⁷³ Although due to the lack of space I cannot really elaborate on this, it must be said that the sample contains datapoints strongly suggesting that the NEG+V-IRR.PST pattern (column 5) often conveying negative past events, derives from the extension of an entailment attached to the negation of a proximative-volitional modal base (cf. English (originally volitional) past irrealis would, where ‘X wouldn’t V’ entails ‘X refused to V/didn’t V’). In Gooniyandi, negative past event entailment readings are attested with a negative (volitional) irrealis past (NEG X V-IRR.PST = ‘X refused to/did not attempt to V’), cf. McGregor (1990: 535). It seems to me very likely that the negative past event readings of NEG+VIRR.PST in other languages derives from the generalization of originally negative volitional entailments, to non-agentive verbs (again, see English would, as in ‘the knife wouldn’t cut’), so that from ‘X refused to V’, these utterances came to mean just ‘X didn’t V’. Had all of these forms been intrinsically, semantically negative from the onset (i.e. had they meant something like ‘[X] didn’t want to V’), their combination with an additional negation might have proven troublesome for them to develop negative past event readings. ⁷⁴ Of course, further investigations would be required to actually tell apart bona fide proximative markers from modals glossed as proximatives in certain grammars.

144

PATRICK CAUDAL

Bininj Gun-wok affixes kuyin-/baIlanh- can combine with an irrealis tense, and an imperfective tense in an iterative context (i.e., with multiple averted events). Similarly, the Anindhilyakwa proximative adverbial akwədhangwa can combine with the irrealis non-past to signal a present perfect-like proximative (78)—a reading obtaining in Bininj Gun-wok too, cf. (79); (78) and (79) are probably the closest equivalent to a present avertive we can come by. All in all, given their apparently arbitrary distribution with various inflections, these datapoints seem to pertain to some kind of conventionalized periphrastic avertive construction. (77) wambawoo gilangginaddirni nearly it:might:have:knocked:me ‘(the car) nearly knocked me.’ (McGregor 1990: 533)

(Goonyiandi)

(78) akwədhangwa ka-lharrəmərdhə-na=ma (Anindilyakwa) IRR.NEUT-darkness-NPST=MUT near ‘It’s nearly dark’ [source translation] (Leeding 1989: 454) (79) A-bal-guyin-yakwo-yi. l-away-nearly-finish-IRR ‘I’ve nearly finished.’ (Evans 2003: 525)

(Bininj Gun-wok)

6.4.3 Avertive pattern no. 3: Pluractional imperfective and REDuplication based patterns (with or without additional markers) Let us turn next to another class of avertive structures possibly exhibiting indicative tense marking, namely utterances denoting pluractional event structures. These often involve imperfective morphology or aspectually underspecified morphology alongside, either within an iterative context, or in combination with clitics, particles or adverbials enforcing an iterative interpretation—especially intonations marking event durativity, and/or reduplication (RED) morphology.⁷⁵ Imperfective tenses in many Australian languages of our sample have clear pluractional readings, and these frequently crop up in contexts where the speaker insists on the fact that some agent made a protracted but vain attempt, ⁷⁵ Note that even languages where so-called perfective tenses are identified in grammars, seem to rather be aspectually underspecified past tenses, akin to the English simple past. Thus, the so-called ‘past perfective’ in Bininj Gun-wok, is obviously capable of past imperfective readings too, cf. e.g. examples 5–61 p. 145, and 5–69 pp. 146–147 in Evans (2003).

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

145

cf. (80); they are instances of proximative/volitional-avertives (the agent’s desire is highlighted). And in addition to simple iterated avertives, habitual avertives are also possible in Yankunytjatjara, cf. (81). (80) kaa-nna Kanytji-nya puttu tjapi-ningi (Yankunytjatjara) CONTR-lsg(ERG) Kanytji-ACC IN.VAIN ask-PST.IMPF ‘And I kept asking Kanytji to no avail.’ (Goddard 1983: 62) rritji-milal-payi, (81) papa-ngku puttu puttu dog-ERG IN.VAIN reach-LOAN-CHAR IN.VAIN wawani-ma (Yankunytjatjara) jump.up.on hind legs-IMP.IMPF ‘But dingoes couldn’t reach them, they’d dump up on their hind legs to no avail.’ (Goddard 1983: 367) Verb reduplication is another recurrent imperfective-like, proximative/ volitional-avertive pattern found across both non-Pama-Nyungan (Iwaidja, Bininj Gun-wok, Nyikina) and Pama-Nyungan languages (Kayardild, Bilinarra, Ngarla,⁷⁶ Arrernte), cf. (82)–(85)—and it is likely to be more widespread than shown by this grammar mining study, even in our sample. Verb reduplication can be regarded as an imperfectivizing device, especially when the tenses involved are aspectually underspecified. The result is a nearimperfective structure, with the iteration often highlighting the agentive nature of the verb used (such avertive reduplication were limited to agentive telic verbs in the data we examined), and therefore the ‘volitional’ nature of the event it denotes (hence a frequent rendering by ‘attempt’ or ‘try’).⁷⁷ (82) illustrates a periphrastic irrealis pattern in Bininj Gun-wok combined with reduplication (cf. the irrealis-avertive particle yimankek), but a plain past indicative marking is also possible, cf. (80); it is followed by an adversative clause contextually specifying the failure meaning. Sentential negation (as in (84)–(85)) can also appear on the right edge of reduplicated clause, thus forming an ‘elliptic’ negative clause, and in effect, a special type of avertive structure—we will focus on such structures in Section 6.4.4. (82) birri-yah-yame-ng yimankek 3AP-INCEP-spear-PFV CTRFAC ∅-warreh-warrewo-ng. (Bininj Gun-wok) 3P-ITER-wreck-PFV ‘They tried spearing Ngalyod, but kept missing.’ (Evans 2003: 381) ⁷⁶ Cf. Ngarla nguru~nguru ‘almost immersed’ (Westerlund 2015: 165). ⁷⁷ It can also be found with irrealis marked verbs in Bininj Gun-wok.

146

PATRICK CAUDAL

(83) barri-yah-yame-ng gunj, 3A/3P-INCEP-spear-PFV kangaroo barri-warreh-warrewo-ng. (Bininj Gun-wok) 3A/3P-ITER-miss-PFV ‘They tried to spear the kangaroo but they kept missing it.’ (Evans 2003: 381) (84) darordam-tha raa-ja warirr (Kayardild) break-REDUP-ACT spear-ACT nothing ‘(They) tried spearing (him) but in vain (= nothing happened).’ (Evans 1995: 290) (85) yi-rr-ma-WIRRI-WIRRIGA-nydyi-na mandya walli… mallu .. 3-NMIN-INTP -try-tryS -INTS -PST many animal NEG .. mirril (Nyikina)⁷⁸ certain ‘Many creatures tried and tried… no luck at all’ (Stokes 1982: 287) Bininj Gun-wok stands out in our sample, as it is the only language to possess a special morphological reduplication pattern for avertivity (glossed INCEP in (82) and (83)), which seems to have both inceptive and iterative readings; it seems to mean something like ‘keep on beginning, keep on being about to’. Interestingly, affixes with an inceptive-iterative meaning can also be found in other languages of our samples, and those also have avertive interpretations, cf. e.g. the Arrernte -elp ‘continuous inception’ suffix (‘keep beginning, be on the brink of ’). It was already mentioned in Section 6.4.2 as a case of proximative affix, but (Wilkins 1989: 261–262) primarily analyses it as a reduplication marker. It is cognate (Hercus 1994: 201) with the Arapana stem-forming suffix -alpa ‘not quite’, which also has avertive uses. In any case, such datapoints show that the boundaries between some of the patterns here listed can be quite porous, by illustrating the strong semantic overlap between reduplication, imperfectivity and proximativity—and their tendency to have converging development paths towards avertive meanings. Last but not least, a special type of intonation indicating a marked temporal duration (i.e. meaning something like ‘for a long time’) can also appear both with and without reduplication to convey a protracted, and therefore vain attempt. The relevant intonation is glossed as ‘::’ in (86); it is quite common across Australian languages, and tends to mark the final syllable of the verb ⁷⁸ INTP / INTS are valency-related affixes.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

147

phrase.⁷⁹ Again, the failure of the attempt seems to be discursively indicated by an adversative clause (w-ardajb-ung, ‘he wasn’t able / failed to break it’). (86) r-urlukba-n:: w-ardajb-ung (Iwaidja) 3m.sg>3sg.O.ANT-step.on-ANT:: 3sg.ANT-couldn’t.break it-ANT ‘He repeatedly tried (= tried hard) to break it with his foot but failed.’ (Caudal and Mailhammer 2022: 219)

6.4.4 Avertive pattern no. 4: V-IRR/V-IMPERF + NEG as conventionalized constructions The fourth avertive pattern in our list involves any of the previously listed avertivity structures, or an indicative past utterance, followed by an isolated negative particle or clitic. It can be found with synthetic (87)–(89) or periphrastic irrealis-avertive marking (90), as well as with plain past indicative verbs combined with avertivity-inducing markings as in e.g. (91) (reduplication or other iterative expressions, especially combined with a durative intonation, i.e. proximativite/imperfective meanings, therefore possibly endowed with a modal dimension, cf. Section 6.4.3). (87) bariyoondirni marlami bithami he:might:have:climbed not he:got:stiff ‘He tried to climb up, but couldn’t. He was too stiff.’ (McGregor 1990: 533) (88) yatha nga-b-irriga-na mangarra alright 1sg:3sg- POT:COOK-IMPF plant.food

(Goonyiandi)

dempa damper

damarlung (Jaminjung) nothing ‘I was going to/wanted to bake bread all right, damper, (but) nothing (i.e. I didn´t)’ (Schultze-Berndt 2000: 93) (89) ayana-wu-ni ba karlu (Iwaidja) 1sg>3pl.PCF-hit-PCF but NEG ‘I was going to hit them/I nearly hit them, but I didn’t. (Caudal and Mailhammer 2021) ⁷⁹ See e.g. Simard’s (2013) conception of ‘iconic lengthening’, or Caudal and Mailhammer’s (forthcoming 2022) ‘linear lengthening intonation’.

148

PATRICK CAUDAL

(90) nungka yimankek ∅-dulubu-yi bulikki, dja he CTRFAC 3P-shoot-IRR bullock but burrkyak-ni. (Bininj Gun-wok) nothing-IMPF ‘He tried to shoot the bullock, but nothing.’ (Evans 2003: 374) (91) w-aran:: karlu marukurnaj ri-widari-ny. (Iwaidja) 3sg.ANT-go.on-ANT:: NEG indef.PRO 3m.sg>3sgO-ANT-fail-ANT ‘He went on for a while but nothing. He failed at whatever he was doing.’ (Caudal and Mailhammer 2022: 230) Such patterns are very common in our sample (ten languages out of seventeen possess them)—but they might well be more widespread than the present survey suggests. I will analyse them as reduced forms of so-called ‘adversative structures’ à la Malchukov (2004) (see also Plungian’s 2001 notion of antiresultatives, which establishes a clear connection between avertives and biclausal adversatives). Contrary to discursively construed adversatives structures exemplified in our type 1 examples, the adversative clause is here reduced to a simple negative item standing for a whole negative clause—I will therefore call it a reduced negative clause. Like periphrastic inflectional avertive patterns (type 2), and unlike synthetic inflectional avertive patterns (type 1), they conventionally convey avertivity in the sense that they constitute at least ‘rhetorical routines’ à la (Detges and Waltereit 2002)—or possibly syntactic constructions. In support of a constructional analysis, it can be observed that their phonological properties require a drop followed by a pause before the negation, which suggests a two-clause structure. Syntactically, they may lack overt syntactic material (conjunctions or discourse connectives) marking the boundary of the reduced clause, a property reminiscent of (Hale 1976)’s ‘adjoined relative clauses’ in Australian languages, and in general, of so-called ‘clause chaining’ in these languages. It should be stressed that reduced negative clauses abound outside of such avertive structures. They can express emphatic negation (albeit with a different phonological structure), when the previous clause is already negative; the second negation then reinforces the first (‘not at all’), (92). (92) minj djama barri-yawoih-na-yi gayakki (Bininj Gun-wok) NEG not ya/3PST-again-see-IRR nothing ‘And no-one ever saw (Daddubbe) again’ (Evans 2003: 624) Reduced negative clauses are often associated with resumptive, anaphoric negations with VP ellipsis (93) (here tantamount to a clause-level ellipsis),

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

149

even multiple anaphoric negations, cf. gayakki and marrek in (94)—again, we seem to be dealing with a two-clause construction. (93) nanibu barri-ganj-ngune-ng, dja nanibu MA:PROX.SER.pl 3A/3PST-meat-eat-PFV CONJ MA:PROX.SER.pl gayakki (Bininj Gun-wok) nothing ‘Some of them ate the meat, and some of them didn’t.’ (Evans 2003: 306) (94) djama ba-ngu-yi njamed, gayakki, marrek not 3P-eat-IRR anything, nothing, not ba-ngu-yi. (Bininj Gun-wok) 3P-eat-IRR ‘He didn’t eat anything, he had nothing to eat.’ (Evans 2003: 283) While negations are arguably particles in most Australian languages, these patterns stand apart from the periphrastic type 2 avertive patterns, due to their different phonological and syntactic properties.

6.4.5 Avertive pattern no. 5: ‘Partitive culminations’ as an avertive strategy Let us now turn to the fifth and last type of avertive structure found in the sample. The term ‘partitive culminations’, or PC, will here refer to so-called non-culminating accomplishments (Bar-El, Davis, and Matthewson 2006) and non-culminating achievements, i.e. past telic utterances with some tense associated (at least contextually) with a past perfective meaning, but for which culmination is not warranted, or at least defeasible.⁸⁰ While most of the existing literature (cf. e.g. Martin and Demirdache 2020; Altshuler 2014) tends to argue that only the latter, and not the former, are possible, other works suggests that non-culminating achievements are indeed attested crosslinguistically, as is clearly the case in e.g. Hindi (95) (Arunachalam and Kothari 2011) or Mandarin (96). Similar examples can be found in Australian languages, cf. (97) is a clear instance of such a reading in Mawng: (95) maya-ne kamiiz taang-ii par vah tangii Maya-ERG shirt hang-PERF but it-ACC hung nahiin not ‘Maya hung the shirt, but it didn’t get hung.’ (Kothari 2008)

(Hindi)

⁸⁰ For a now ‘traditional’ definition of perfective vs imperfective aspect, cf. e.g. Smith (1991); for a formal, discourse-oriented one, see Caudal (2012).

150

PATRICK CAUDAL

(96) Xu Mei he Sun Mazi ba Lao Luo sha le mei Xu Mei and Sun Mazi BA Lao Luo kill PERF not sha-si (Mandarin) kill-die ‘Xu Mei and Sun Mazi killed Lao Luo but didn’t make him die.’ (Koenig and Chief 2008) (97) malany nungpak-apa inyng-ikp-ung. (Mawng) CONJ THAT.EMPH 3GEN/3F-wake-PST. ‘Then she tried to wake the other one up.’ (Singer et al. 2015: 56) Biclausal PC constructions with reduced negative clauses constructions are also found in the sample, cf. e.g. (98)–(99). They are probably underrepresented in the data I garnered through grammar mining, as PCs were only recently identified as a linguistic category of interest, especially since due to their simplicity, such structures have an innocuous, run-of-the-mill feeling to them, authors may tend to omit them in their grammatical investigations.⁸¹ (98) dathin-a wirdi-j, bala-tha ni, warirr (Kayardild) there-NOM stay-ACT hit-ACT 3sgNOM nothing ‘He stayed there and pounded the bait (to attract fish with its grease). Nothing [= he didn’t catch any fish’. (Evans 1995: 299) (99) milalimi marlami I:looked not ‘I looked, but didn’t find it.’ (McGregor 1990: 495)

(Gooniyandi)

At least two languages in the sample have a clear weak vs strong perfective tenses opposition: Anindilyakwa (REAL-V-∅ has weak perfective, PC readings (100), whereas REAL-V-PST is an underspecified past tense with ‘strong’ perfective readings, and does not license PC readings (101)) and Kayardild (V-ACT also has strong perfective readings, whereas V-PST has weak, PC-readings).

⁸¹ My intuition is also based on the observation that I have very often encountered such structures while doing fieldwork on Iwaidja and Anindilyakwa, in ordinary verbal exchanges, and initially did not pay very much attention to such data points.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

151

(100) n-alyubaru-nu=ma y-akina yinumaninga (Anindilyakwa) REAL.3M-eat-PST=CTYP MA-that MA.food akena nara kin-alyubari-na but NEG IRR.3M>MA-eat-PST ‘He began to eat the wild apple, but didn’t finish it’ (Bednall 2020: 206) (weak perfective; partitive culmination—event began but failed to culminate) (101) ∗n-alyubaru-Ø=ma y-akina yinumaninga (Anindilyakwa) 3M-eat-USP=CTYP MA-that MA.food akena nara kin-alyubari-na but NEG IRR.3M>MA-eat-PST ∗‘He began to eat the wild apple, but he didn’t finish it.’ (Bednall 2020: 206) (strong perfective; ∗ partitive culmination—event has to culminate) Interestingly, while the weak tense is the least temporally specific in Anindilyakwa (REAL-V-∅), the reverse holds true in Karyardild (-ACT is aspectuo-temporally underspecified, but can have ‘strong’ perfective readings); so underspecification plays no part in determining weak vs strong perfective readings of tenses. The real explanation, I believe, lies in tenses still being relative/content dependent vs absolute to some extent.⁸² I will here claim that ACT (in Kayardild) and REAL-V-∅ are in fact dependent, relative tenses, in a syntactic sense. In Kayardild, ACT cannot mark past subordinate clauses (Evans 1995: 261)—only PAST can. Similarly, the temporal interpretation of REAL-V-∅ subordinates or ‘chained clauses’ in Anindilyakwa depends on that of their matrix clause, whereas that of REAL-V-PST subordinates/‘chained clauses’ does not (Bednall 2020). Before closing this subsection, it is important to note that bare PC structures, and generally discourse-dependent PCs (100)–(101) (‘pragmatic’ PCs, column 11 in Table 6.2), differ crucially from both other avertives of the sample, in that (i) they do not contribute a negative event contradicting an

⁸² These tenses are classic cases of what N. Evans refers to ‘insubordination’ (Evans and Watanabe 2016)—or how a syntactically dependent tense came to mark matrix clauses—where the former dependent/relative tense has retained some of its semantic dependency. ‘Weak perfectivity’, I believe, is here preconditioned by this parameter. For a historical analysis of the Kayardild PAST along these lines, see Evans (1995: 443); and Mark Harvey (p.c.) suggests that the Anindilyakwa ‘null’ inflection is historically derived from a former (relative) imperfective.

152

PATRICK CAUDAL

attempt/belief/expectation per se, and are utterly deprived of any inherent modal-evidential content (those are at best instances of ‘free’ pragmatic enrichment) and therefore (ii) they do not describe complex event structures, a part of which is modalized. In other words, they are at best ‘weak’ avertive strategies (see Mu¨ller 2013: 106 for a related observation concerning South American Indigenous languages). And I take conventionalized PC-looking constructions with a reduced negative clause (column 10) as bridging the gap between pragmatic PCs and bona fide semantic avertives reviewed so far: I classify PC+NEG patterns as semantically avertive structures since they encode a negative event; but they lack the modal-evidential content associated with other types of (semantic) avertives. I will come back to this crucial fact in my theoretical conclusion below.

6.4.6 Some key empirical generalizations about modal, aspectual, and actional and subtypes of avertives The data discussed so far demonstrates that many avertives structures, marked or not with an irrealis-avertive inflection, come in different modal flavours: predictive/proximative (‘was going to V / would have V, but didn’t’—where V is typically a non-agentive verb), predictive necessity (‘X shouldphysical/social/dispositional have V, but didn’t’), volitional (‘X wanted to V/tried to V, but didn’t’), capacitative (‘X could have V/tried to V, but didn’t’) and some special flavour of deontic modals (‘X should have V, but didn’t’)— which is essentially deontic modality pointing to a non-accessible world. I will call ‘reproachatives’ the latter type of deontic-based avertive (cf. Olmen 2018; see also ‘admonitives’ in Harvey 2002); they describe failures to meet a past obligation, and are typically used to blame an addressee. Doxastic and evidential avertives (‘X thought that P/it seemed to X that P, but it turned out that non-P’) are also attested in the language sample, though not for ‘bare’ irrealis-avertive inflections—they require additional particles. This suggests that avertives in Australian languages can have both a dynamic and a non-dynamic modal basis; a least some avertive structures are capable of both dynamic and non-dynamic modal (including evidential) readings (see (111)–(113) below). All in all, it appears that in Australian languages (i) both dynamic and non-dynamic modal meanings can underlie avertive structures, and that (ii) non-dynamic vs dynamic modal-avertive structures can be conventionalized separately, or not, though there seems to be a tendency for them to appear in separate constructions.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

153

Another important generalization with respect to modality should complement those put forth above: non-avertive irrealis meanings, i.e. non-avertive deontic, conditional and hypothetical meanings, all typically associated with forms pointing to the irrealis avertive cluster, do not seem to arise with doxastic/evidential avertive structures involving an evidential/doxastic or capacitative meaning; even volitional and proximative meanings can be absent from some avertive forms, which are clearly much less polyfunctional than those typical of the irrealis-avertive cluster. This is strongly indicative of the existence of seriously diverging development paths for avertives in Australian languages. It points to at least three distinct sets of development paths: one broadly related to the irrealis-avertive cluster (and associated with volitional/proximativepredictive modal meanings), one broadly related to negatively oriented evidential/doxastic meanings, and one to negative capacitative meanings. I will get back to this question in Section 6.5. Turning to aspect, I will not discuss here in great detail the various aspectual subtypes of avertive meanings exhibited (or not) in the sample by each of these patterns, mostly for want of a sufficient number of datapoints to effectively be able to do so. But based on language where data is most abundant, it seems that the three following generalizations hold (they extend beyond the irrealisavertive cluster): 1. avertive structures, regardless of their nature, seem to require change-ofstate, dynamic, and certainly non-stative verbs; whenever they combine with stative verbal roots, these receive a dynamic, teleological interpretation, with the target result being denied to the agent argument, i.e. a resultative avertive reading (they become endowed with a contextually determined ‘telos’/result stage, as in e.g. ‘X look for something’, ‘X wait for someone’, ‘X watch something’, i.e. verbs with lexically encoded telos, which I take to constitute a case of sublexical (teleological) modality, cf. Martin & Scha¨fer 2017);⁸³ and if a seemingly verb bears a so-called ‘inchoative’ derivational affix, which makes it ambiguous between a change-of-state reading and a stative reading, then the change-of-state reading must prevail. As for activity verbs, they seem to receive some sort of teleological, telic-like contextual enrichment (as their lexical entry generally lacks any teleological sublexical modality) as well as when used ⁸³ Hale (1969: 208) offers similar activity-based examples (e.g. ‘look around for’), which must possess a lexically encoded telos/goal (e.g. a ‘finding’ event in the case of ‘look for’); but even if they don’t, it seems that some agent-controlled activities can be contextually enriched with such a telos.

154

PATRICK CAUDAL

in avertive contexts—but further investigations are clearly required to make this more specific 2. proximative/imperfective-related avertives in Australian languages tend to have either a full event/preparatory stage avertive reading (‘X wanted/was going to V’; prospective/proximative aspect) or inner stage avertive reading (‘X was V-ing, but did not finish V-ing’), but not a result stage avertive reading (‘X V-ed, but expected results did not obtain’). I will claim that this is a predictable consequence of the fact that the verb’s underlying event predicate cannot culminate, due to the proximative/imperfective or iterative morphology involved 3. perfective avertives and PC-avertives in Australian languages tend to have any of the main aspectual types of avertive readings; perfective achievement utterances can only have full event or resultative avertive readings. Finally, let us consider the actional parameters underlying avertives in the sample. I did not find clear evidence for a principled distinction between ‘subject-controlled avertives’ (i.e. frustratives in my terminology) and ‘nonsubject controlled avertives’, both within and without irrealis-avertive clustertype languages. Inflectional, pragmatic avertives in the sample did not seem to be biased towards either of these two readings—though of course, volitionalleaning interpretations were mostly restricted to controlling subjects; in the absence of a controlling subject, a predictive/proximative reading seems to prevail, and the associated private state is ascribed to the speaker, and/or to some contextual judge (this is particularly obvious for irrealis-avertives with an evidential, ‘not what it seems’ meaning). The actional semantics of particlebased avertives is often more delicate to assess, in the absence of sufficient data. This question, therefore, will have to be settled by future research.

6.5 Conclusion: Theoretical consequences of the survey I must now come back full circle, and attempt to answer my initial research questions: are positive versus negative utterances separated by a clear ontological split, or do avertive structures somehow bridge the gap between actual and inactual events? And what can avertivity tell us about how the human mind (as evidenced by linguistic systems), structures time? I will first focus on so-called avertive development paths in Australian languages (Section 6.5.1), before striving to identify important theoretical consequences of this study for our

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

155

understanding of avertives as a general linguistic category, and beyond that, for a theory of time as construed from structured events (Section 6.5.2). I will then proceed to drawing some possible consequences for language typology (Section 6.5.3), before concluding (Section 6.5.4).

6.5.1 Development paths and cognitive underpinnings: Where do avertive come from, and how are they semantically/pragmatically structured? Given the results garnered through the analysis of our language sample, I will argue that several distinct development paths can be identified for Australian avertive structures, depending on whether they involve imperfectiveproximative meanings (6.5.1.1), volition and other dynamic modal meanings (6.5.1.2), or similative/evidential meanings (6.5.1.3). 6.5.1.1 From imperfectivity-proximativity to avertivity: The predominant set of development paths for avertivity in Australian languages As we have seen, numerous proximative-leaning elements (imperfective or modal morphology whose meaning can be proximative ‘be about to/be going’, ‘almost, nearly’ affixes, clitics, particles or adverbials, reduplication) appear in avertive constructions. Could those different classes of proximative patterns have followed related development paths—in the sense of Bybee, Perkins, and Pagliuca (1994)—with or without a bona fide imperfective content? If that is the case, then a vastly predominant family of closely related development paths emerges from our areal study. The fact that proximative-avertive meanings (‘X was going to V, but didn’t’) appear with both indicative (imperfective, or aspectually underspecified) past morphology as well with modal, irrealis inflections (cf. the kuyin- preverb in Bininj Gun-wok), is strongly suggestive that imperfectivity per se is not a necessary ingredient of the development path. Having a ‘part-of ’ semantics⁸⁴ might explain why imperfectives often develop proximative (‘be about to’, ‘be going to’) meanings crosslinguistically. The effect of proximative adverbials as essentially quantitative expressions (‘was at a point near completion/inception/obtention of results’ (imperfective), or ‘came close to completion/inception/obtention of results’ (perfective)) in avertive contexts ⁸⁴ Cf. e.g. (Altshuler 2014) for a relatively recent ‘part-of ’ relation based formal analysis of imperfectives—though the idea is fairly ancient; for the Romanist tradition, it dates back at least to G. Guillaume’s work between the two world wars (Guillaume and Vassant 1992).

156

PATRICK CAUDAL

is similar, regardless of the viewpoint involved: both imperfectives and proximative adverbials indicate that the event denoted is somehow incomplete. Either the event as a whole, or its culmination or its results—some event part is somehow lacking. This would give us two closely related, partially overlapping development paths, (102)–(103). Given the present sample with its prominent proximative/volitional dimension of irrealis-avertive clustes (column 5 in Table 6.5), it seems natural to assume that the shift from a non-proximative past imperfective, to a proximative imperfective corresponds to a volitional modal enrichment of the inflection. It can only arise if it is compatible with telic verbs (see the Murrinh-Patha past imperfective for an instance of imperfective inflection incompatible with telic verbs, and logically lacking a proximative interpretation). I will hypothesize that is a consequence of telic verbs possessing a sublexical modal content, of a teleological/volitional type, which becomes singled-out by the proximative meaning extension. (102) The past tense + proximative development path Past tense + proximative adjunct > past avertive-counterfactual (‘almost, nearly’) (CF) (103) The (past) imperfective/proximative to avertive development path: Past ipfv. > Past ipfv./proximative/ > prox./volitional/ volitional irrealis/avertive I am arguing here for a type of evolution conjoining two types of mechanisms. The first is meaning accretion, that is the coexistence of ancient vs novel meanings at some evolution stage. This corresponds to ‘layering’ à la Hopper (1991), or a bridging context à la Heine (2002).⁸⁵ (103) means that I am arguing that the original past imperfective meaning of an inflection was shed after it developed into a form possessing at once proximative, volitional, irrealis, and avertive meanings—which, of course, corresponds to the Northern Australian irrealis-avertive cluster. This hypothesis is supported by independent diachronic analyses for some languages in the sample, as at least some irrealis-avertive non-Pama-Nyungan paradigms can be reconstructed as derived from former imperfective paradigms, as in e.g., Gunwinyguan,⁸⁶ or

⁸⁵ Note that if the volitional/proximative ingredients came to vanish, this could lead to the formation of a narrower irrealis-avertive—and with further shedding, to a pure avertive or irrealis (such as I think, the Kunbarlang past irrealis, which does not seem to have proximative/volitional or avertive readings). This would be a case of switch in Heine’s (2002) theory; cf. also path (102), where the second stage illustrates a switch case à la Heine (2002). ⁸⁶ As noted in Alpher, Evans, and Harvey (2003: 312); Kapitonov (2019: 173), the predominant Gunwinyguan past irrealis paradigm /niɲ/can be reconstructed as a past imperfective.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

157

in Maningrida languages.⁸⁷ Furthermore, I believe the past imperfective path proposed above is in line with diachronic-typological work such as (Sansò 2020) (see in particular the ‘be’ development path). It should also be stressed that forms on path (102) differ crucially from those on path (103) in that the former initially entail, and eventually denote negative semantic content as at-issue meaning (à la Potts 2005, 2007), whereas the latter often convey such negative content as defeasible negative implicatures, becoming possibly attached later on to a secondary, non-at issue dimension of meaning. The development of counterfactual/negative event meaning from proximative adjuncts is already well-known from the history of English (Ziegeler 2000, 2015), and similar evolutions might very well have taken place in Australian languages. Furthermore, treating the negative import of (103) as a matter of non-at issue (and possibly defeasible) content is consistent with the widely held view in the literature that past irrealis forms convey defeasible negative implicatures—or presuppositions at most (Ippolito 2003, 2006; Arregui 2009)—including in the Australianist literature (Verstraete 2005, 2006; Van Linden and Verstraete 2008), not downright semantic (especially not at issue) negative content. This also explains why such pragmatically construed avertive interpretations (in effect ‘avertive strategies’) contrast with those ‘marked’, semantically avertive constructions in the irrealis, involving additional, overt negative elements (avertive/proximative particles, or reduced negative clauses) in this respect. If the two development paths put forth above are correct, then it follows that proximativity always has potential for an avertive-irrealis development regardless of the associated aspectual viewpoint, because it can always lead to a counter-to-fact, negative enrichment—this is consistent with independent typological regularities (cf. e.g. the development of some Romance conditionals from imperfectively marked modal constructions, or the Romanian dedicated avertive construction a fi pe cale (Pahonțu forthcoming), which admits both perfective and imperfective marking).⁸⁸ Reduplication/iterative morphology, which we have seen to be frequent in our inventory of imperfective-proximative avertive patterns (Section 6.4.3),

⁸⁷ Green (1995: 195 ff.; 2003: 399) reconstructs the Gurr-goni irrealis inflection -rni as derived from a proto-Maningrida root ∗ni/∗nu ‘sit’. ‘Sit’ being a notorious copula in Australian languages, Sansò (2020: 416, n. 6) argues that it most likely developed into some progressive (with proximative uses), and from there, into an irrealis. It should be noted, furthermore, that -ni sounds very much like a common pan-Australian root for ‘sit’, and that several past suffixes possibly derive from it (cf. e.g. Anindilyakwa, Jaru, Ngarla). ⁸⁸ It is my belief that a fi pe cale has just entered a bridging context phase (Heine 2002) between the two stages of (102).

158

PATRICK CAUDAL

requires two sub-development paths given in (104): the general reduplication/iteration type (104.1) branches out on the more general imperfective pattern, while the inceptive-reduplication type (104.2) behaves more like a proximative adverbial like ‘nearly’; as we have seen in Section 6.4.3, there is substantial semantic overlap between inceptive-iterative and proximative meanings. (104) The reduplication/iteration to avertive sub-development paths 1. Reduplication/ > proximative/volitional > … iteration (‘keep trying’) 2. Inceptive > proximative > avertivereduplication counterfactual While reduplication and iterative morphology has so far not been mentioned as a major source of avertive markers cross-linguistically via the proximative/imperfective path, it seems to be widely attested across languages of the world. For the Americas, see e.g., Hintz (2011: 68–69) for related datapoints in Quechua; (105) is a clear example of reduplication-based avertive structure in South Conchucos Quechua. For Europe, see e.g. Moksha Mordvin, an Uralic (Mordvinic) language, which possesses an avertive (‘almost’) suffix derived from an iteration/habituality suffix (Kozlov (2019: 133)), and Russian, where instances of reduplication strikingly reminiscent of a similar inceptive-avertive reduplication pattern can be found (106). (105) tsa cha-yka-mu-r-qa that arrive-PFV.O-DIST-SS-TOP qechu-na:llapa-n take.force-PST

(South Conchucos Quechua)

mu:la-n-ta. wanu-tsi-ypa wanu-na: all-3 mule-3-O die-CAUS-ADV die-PST ‘Then when he arrived, he took away all of his brother’s mules and tried to kill him (in any way possible). (Hintz 2011: 74) (106) Sneg tajal, tajal, no ne (Russian) Snow melt.IPFV.PST melt.IPFV.PST but NEG rastajal. melt.PFV.PST ‘The snow started to melt but did not melt away completely.’

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

159

6.5.1.2 Volitionandotherdynamicandinterpersonalmodalbasesgivingrise to irrealis-avertive development paths in Australia As shown by column 6, most irrealis-avertive inflections have a volitional interpretation in their development path. This connection between volitionality and avertivity is hardly a surprise, as the volitional development path of avertives is well-known from Kuteva’s (1998) seminal paper; it is crosslinguistically abundant, in Europe, in Africa, and in the Americas—cf. e.g. Garcı´a Salido (2014: 295–297) for multiple examples in Tepehuan, an Uto-Aztecan language, (107). (107) Tii ba-tu-aski-ch-dha’-iñ pu INT.NR CMP-DUR-bag-CAUS-APPL-1sg.SBJ SENS cham matit ti-tirbiñ-dha’-iñ ja’p NEG know.PFV RED:ITER-fold-APPL-1sg.SBJ DIR

(Tepehuan)

añ chii bua-da’ 1sg.SBJ INT.NR make-CONT ‘I wanted to make bags, but I did not know how to fold the threads, I intended to do it, (but I could not).’ (García Salido 2014: 295) A volitional development path complementary of (103) is proposed in (108), as the avertive function of former imperfectives might have derived directly from said volitional meaning as well.⁸⁹ It is based on the uncertain outcome initially associated with volitionals—indeed, given a neutral volitional only contextual information will indicate whether an agent’s desire was granted or thwarted by subsequent circumstances (cf. English verb want). But again, most forms found in the sample seem associated with some form of defeasible, but nevertheless salient (and therefore conventionalized) failure implicature (hence speakers’ uneasiness, sometimes, at granting them positive outcomes)—i.e. are one step beyond a simple volitional verb giving rise to a conversational implicature. (108) The volitional development bath (‘wanted but NO’): volition (uncertain > (defeasible) > avertive (semanticized) outcome) implicature of failure However, I will argue that proximative meanings also constitute a potential modal base, of a doxastic-predictive nature – they encode a high degree of ⁸⁹ It contrasts with Kuteva’s (1998) proposed path, (15), in that proximativity being obviously a previously developed meaning (a basic meaning of imperfectives combined with telic verbs, including non-agentive ones) according to my analysis, volitionality must intervene later on.

160

PATRICK CAUDAL

expectation about some imminent event in the past (i.e., a past belief concerning ulterior worlds); see e.g. the role played by the notion of inertia in Carol & Salanova (2017)—I take futures in general to encode similar modal bases. In the absence of an agentive subject, volitional/proximative avertive structures will take on such a doxastic-predictive modal flavour (whereby some expected ulterior event did not materialize). As for negative ability markers such as puttu in Western Desert languages (as in Yankunytjatjara, cf. Goddard 1983:163, and Pintupi-Luritja, cf. Rose 2001:70–71—see also other cognate forms in neighbouring language families, e.g. Pilbara languages), it should be observed that capacity modals are known to easily evolve towards volitional and directive meanings (this is a common development path in SAE languages, for instance), and in general towards agentive exertion readings (‘try’).⁹⁰ In combination with negative polarity, these exertion readings can naturally give rise to avertive meanings (‘X tried to V but did not V’). Note that purtukarri in Ngarla (Pilbara), which can convey both negative capacity and negative volition (‘unwillingly)’, suggests indeed a complex relation between volition, capacity, and negative polarity—and in such a case, the event described by the verb is not averted; what gets negated is an underlying volitional modal base. Even independently from Australian facts, capacity modals seem to give rise to avertive meanings by being interpreted as describing an event of exerting one’s ability (i.e. trying and failing to realize it); this is evidenced by e.g. negative capacity modals (or negated capacity modals) found in Romance or Germanic languages (cf. e.g. English ‘X was unable/was not able to V’ qua ‘X tried and failed to V’ negative capacitative avertives). To that effect, the underlying ability modal must be capable of having not just generic, dispositional readings, but also stage-level, non-permanent readings (cf. the related notion of ‘action dependent ability’ in Mari and Martin 2008)—a perfect fit for a change-of-state, avertive interpretation. In an utterance like (56), I take the ‘couldn’t’ rendering of puttu as an illustration of such a failed, deliberate/volitional exertion at realizing such an ‘action dependent ability’. And of course, what I have called reproachative avertives (Section 6.4.6) (‘you should have V [but you didn’t]’) being related to deontic, i.e. priority modal meanings (Portner 2018), they are potentially deeply connected to volitionals (thus, ‘I want you to V’ can contextually implicate ‘you must V’), and could be part of a semantically extended path encompassing (108). ⁹⁰ I will get back to this below, when addressing capacity modal verb pouvoir (‘be able’) in French, and its relation with avertivity and other postmodal meanings, i.e. so-called actuality entailments.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

161

6.5.1.3 The similative/evidential development path The last major grouping of development paths I would like to propose involves similative counterfactual (‘like’) expressions, and in general forms possessing, or having developed a negative epistemic/evidential/doxastic content, cf. yimarne(k)/yimanke(k) in Bininj Gun-wok and Kunbarlang, djangagogo and djaying in Bininj Gun-wok. The avertive structures in which these markers appear routinely express negative modal meanings, and often reflect on failed expectations and unjustified beliefs. Many of these expressions derive from roots/affixes/clitics whose original meaning was ‘like’/ ’seem/not what it seems’. They sometimes incorporate an overt negative element—as in Warlpiri.⁹¹ These forms frequently aggregate negative evidential-doxastic meanings (‘not what it seemed to agent/what agent believed)’, occasionally mirative meanings (see Delancey Scott 2012), and capacity meanings (‘can’t’/ ‘couldn’t’).⁹² I will not attempt to account for the latter fact, and will focus on the former, as the negative capacity meaning could be inherently associated with negation in some languages. Thus, the kula negation, common in Ngumpin-Yapa languages, seems to be imbued with a negative capacity meaning, as is evidenced from its uses in Jaru and Walmajarri (McConvell and Laughren 2004: 163–164). Let us take an example to illustrate the path. Yiman is a similative element in Bininj Gun-wok; similatives easily develop into negative expressions, such as past irrealis/counterfactuals, and indeed, avertives (for a straightforward illustration, see the evidential and irrealis/modal evolution of like (‘X’s is like P’) in Modern English, cf. ‘as e.g.’ (Pinson 2020), or as (‘as if ’)). This gives us development path (109). In our sample, one can also put forth Kayardild maraka, with its counterfactual, mistaken thought, and evidential meanings (‘looked like’); maraka is also originally a similative, ‘like’ particle,⁹³ as shown by entries bilulurlda, jurdungaji, and kabanda in Evans (1995: 652, 692, 693). (109) Development path from similatives to evidential/irrealis/avertive/ mistaken/thought > (positive/neutral evidential) : seems P (and is/might be P) like P > (negative evidential): seems P but is not P > evidential > irrealis/avertive/mistaken thoughts/mirative ⁹¹ Note that Djaru, another Ngumpin-Yapa language, possesses a related negative evidential particle kulanga, see Tsunoda (1981: 205). ⁹² I believe Western Desert puttu (‘can’t/couldn’t’) pertains to another development path, as it does not have evidential undertones; see Section 6.5.1.2 ⁹³ See also the related Lardil particle mara, appearing with proximative-avertive meanings (‘was going to V but didn’t’) with the FUT (= IRR) inflection, cf. Evans (1995: 381).

162

PATRICK CAUDAL

The above development path seems quite common in Australia, cf. other similative-derived avertives such as the irrealis/avertive particle karaddiabb(a) in Nakkara (Maningrida), derived from djabba ‘like’ (see Eather 2011: 340– 343), or the Pitta-Pitta avertive particle wiri (‘like’) (see Blake 1979b: 220), a.o.⁹⁴ Finally, other particles and clitics such as Warlpiri kula-nganta (NEG+SEEM), the related Gurinji -nganda (probably cognate with ‘dubitative’ =nga clitic in Bilinarra) possibly illustrate a related (sub)path (110), directly starting from a ‘counter to expectation or belief/ doubtful/not what it seems’ meaning— see also wurrkany in Iwaidja (‘not what it seemed’), which signals a negative evidential, unexpected turn of events.⁹⁵ (110) The negative evidential development path (‘not what it seemed’): > irrealis (from ‘what seems’ to ‘what is’) (misleading) appearance > counter to expectations/ > avertive belief > mistaken thought > mirative Formally speaking, all doxastic/evidential avertive readings found in the sample (‘mistaken belief/perception’: ‘speaker/it was thought/it seemed that P, but that turned out to be wrong’) appear to require a dedicated avertive particle, plus either an irrealis-avertive inflection or an indicative past tense inflection; cf. djaying in Bininj Gun-wok, (111), kula-ngnanta in Warlpiri (112), thaddi in Gooniyandi (113), wurrkany in Iwaidja, etc.; the resulting structures are therefore specialized, semantic avertives. (111) djaying ba-ra-yinj gurih. (Bininj Gun-wok) supposedly 3P-go-IRR there I thought he was going to go that way (but he didn’t).’ (Evans 2003: 374) (112) kula.nganta-kapi-rna wawirri panti-ka-rla, (kala lawa). (Warlpiri) NEG.SEEM-FUT-1sg kangaroo spear-IRR (but no) ‘I thought I was going to spear the kangaroo, but I didn’t.’ (Nash 1980: 239) ⁹⁴ I even suspect that similatives are in fact a very common source for avertives crosslinguistically, cf. Creissels et al. (2007: 106) for an instance of a similative-derived (‘like’) avertive in Tswana, a NigerCongo language. ⁹⁵ This is exactly what the Iwaidja dictionary mentions in its entry, and what R. Mailhammer and I found during field work explorations of the semantics of wurrkany. See also notes 68 and 69.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

163

(113) thaddi thilmangga bijginyjarnirni (Gooniyandi) mistakenly:believed early you:could:have:come ‘I thought you would arrive early.’ (McGregor 1990: 498) In some cases, structures with an evidential/doxastic avertive flavour can also have a dynamic modal (generally volitional) avertive reading as well, see e.g. mundjarra in Gurr-goni (Burrara and Ndéjbbana offer similar datapoints (Green 1995: 315)), (114)–(115)). Intuitively, such cases probably involve a different development path—perhaps not unrelated to how epistemic/predictive modal meanings are derived from dynamic/root modal meanings, as in e.g. English must, could, or would. (114) mundjarra njina-boy-∅ ngayi-pu arrapu Daryl (Gurr-goni) supposedly 1UAnf S-go-IRR2 1MIN-CARD and Daryl wurpu burrkburrk gu-me-ka. but.just bad.sickness 3MINA.3IVO-get-CON ‘Daryl and I intended to come (yesterday), but he got sick.’ (Green 1995: 314) (115) mundjarra gabi police station mu-yo-rri+rni (Gurr-goni) supposedly LOC police station 3IIIS-lie-IRR1 ‘It was supposed to lie at the police station (said of a dead body which people had expected would be flown to Maningrida, but which was flown to an outstation instead).’ (Green 1995: 315) Contrary to the doxastic/evidential avertive meanings we just discussed, the more straightforward ‘failed expectations’ readings we mentioned earlier in this subsection (cf. e.g. (116)) can be expressed by ‘bare’ irrealis-avertives, without additional particles. I take them to involve the same kind of doxasticpredictive modal bases: some (past) expectation of the speaker turns out not to be met. And in spite of what one might think at first sight, such readings are not really epistemic. They can even border on dispositional necessity meanings⁹⁶ (‘according to contextual (agents’ known habits, etc.) and non-contextual constraints (social/physical laws), it was necessary (and therefore predicted) that situation s should come to hold—but it didn’t’). I will therefore regard all those readings as cases of doxastic-predictive modality—which encompasses

⁹⁶ Such utterances do not merely imply that a certitude by an actually ignorant speaker turned out not to be verified, but that there are practical (physical, social) forces at play normally enforcing a certain course of events. In (116) it is clearly inferred that something hampered them.

164

PATRICK CAUDAL

expectations based on physical and social necessity. Such a reading is routinely conveyed by should in English.⁹⁷ I will call such patterns predictive avertives. (116) ya-la-(rr)-DI-na-da mabu (Nyikina) lnsg-IRR-(NMIN)-sit-PST-HAB good ‘we should have been good (but we weren’t)’ (Green 1995: 281) I must leave to future research the task of working out in finer details all those development paths, and some others I was compelled to omit for want of space to discuss them.

6.5.2 Consequences for a cognitive/semantic-pragmatic theory of avertivity, and our understanding of human time Let me now turn to some important theoretical consequences of the discussion conducted in 6.5.1, namely the idea that (semanticized) avertives in Australian languages must contribute complex event structures comprising two separate event predicates (6.5.2.1) (one modal event predicate, and one negative even predicate), and that such meanings of avertive structures must constitute the negative counterpart of those conveying so-called ‘actuality entailments’ (6.5.2.2), thereby legitimating their perception as an important way of negatively construing the flow of time qua a succession of events. 6.5.2.1 Avertives as complex event structures, combining a positive and a negative event If we summarize the findings we have made so far, it appears that conventionalized avertive expressions convey at least two events, one of which (no. 1) associates with a modal meaning: 1. a private cognitive state (intention—which turns out to be frustrated— or a belief/expectation (sometimes based on a perception)—which turns out to be mistaken (hence the connection between ‘mistaken thoughts’/evidentiality and avertivity, so striking in Australian languages with the irrealis-avertive cluster))—I am here treating the modal content of avertive structures as contributing stative events;⁹⁸ ⁹⁷ Note that a weaker, non-necessity modal would not have been strong enough for its nonrealization to be significant. ⁹⁸ This is not an isolated or novel analysis of modals, cf. e.g. (Ferreira 2014).

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

165

2. (optionally) an event fragment, or an event minus its desired/expected results; 3. a negative event (‘event didn’t begin/finish’, ‘results didn’t obtain/hold’), whose very assertion contradicts that event 1 subsequently holds. It must be stressed that no. 2 is indeed optional, as it can be reduced to nothing in contexts where the entirety of the propositional content of the avertive-marked verbs remains unrealized (cf. English John almost died); in such contexts a mere expectation (and possibly intention) event holds. This means that fully semantic avertive structures convey complex event predicates, i.e. must combine at least two event predicates, and two (sub)event variables in their denotation: a private state event predicate (with a modal and/or evidential content) (no. 1), and a negative event predicate (derived from the semantic contribution of the avertive-marked verb) (no. 3).⁹⁹ In addition, the above analysis, if it is correct, entails that avertives are the missing link between bona fide negative events (which, unsurprisingly tend to implicate an avertive reading, given the right context, ‘he didn’t stay’ can easily implicate ‘I expected/hoped/wanted him to stay’) and positive events. It would then be a decisive argument in favour of the hypothesis that negative events are legitimate objects for a linguistic ontology of time, at the very least—and most probably, then, for any ontology of human time, even at an abstract, philosophical level. It should be furthermore noted that this important philosophical question (i.e. do negative events have ontological substance?), has independently received a similar answer on purely theoretical, formal semantic grounds in (Bernard and Champollion 2018); although I cannot develop a formal implementation here, their treatment of negative events would be a perfect piece of machinery to include in such an implementation. 6.5.2.2 On the relation between so-called ‘actuality entailments’ and avertivity: Why avertive interpretations should be regarded as ‘inactuality implications’ The idea that positive and negative events are in fact very much like ‘opposite brothers’ within an extended semantic family, receives further substance if we consider how avertives are actually part of a larger class. Following an idea first (to the best of my knowledge) put forth in (Caudal 2018a, b), I believe that ⁹⁹ In contrast, languages where e.g. past irrealis inflections seem to defeasibly implicate a negative event (column no. 3 in Table 6.2), then no negative predicate should be part of their semantics; it is introduced by means of a lexical implicature—most certainly stemming from a former conversational implicature possibly associated with proximative meanings, given appropriate contexts.

166

PATRICK CAUDAL

avertives cannot be well understood if one does not integrate them within such a larger categorial domain, which I will call eventualized postmodal meanings (EPMs), or more simply, demodals. I define EPMs/demodals as a class of expression with an at least partly postmodal content (in the sense of (van der Auwera and Plungian 1998)). As argued above, I take semantic avertives as conveying complex event structures, comprising (a) a modalized event referent (a belief/expectation (grounded in a perception, or not), or a desire) paired up with and followed by (b) an event sanctioning either its failure/invalidity or success/validity—i.e. it can be a positive (in case of success) or negative (in case of failure). I will hypothesize that avertives (except PC+NEG structures, as they are not modalized/not EPMs) form the negative (failure) side of the EPM coin, while so-called ‘actuality entailments’ (cf. (Bhatt 1999; Hacquard 2009), a.o.) form its positive side. I will therefore argue that avertives convey inactuality entailments (when semanticized), or, in the case of pragmatically defeasible avertive interpretations inactuality implicatures—in effect, avertive forms entail or implicate negative past events, depending on the form at stake. Thus, (117) (as well as (119)) corresponds to the positive counterpart of many ‘failed attempt’ avertives in Australian languages—and the addition of a negation unsurprisingly yields an avertive reading, (118)—whose semantics is very similar to some Australian negative capacity-related avertives, such as puttu in Yankunytjatjara. I believe that (120) is the positive counterpart of what I have called ‘reproachatives’ (cf. Section 6.4.6). Such utterances are typically rendered in grammars by a past deontic modal followed by an elliptic negative clause (‘you should have V-ed, but didn’t’). I will argue that although they are semantically close to SAE past counterfactual deontics, they differ from them in that they are EPMs; SAE past counterfactuals do not necessarily have a clear negative event meaning by themselves.¹⁰⁰ (117) Il a pu partir. (French) He have.3sg.PR be.capable.PP go-INF ‘He was able to leave.’ (= he managed to leave OR was allowed to leave OR seized an opportunity and left)

¹⁰⁰ What prompted me to adopt such an analysis was the somewhat bizarre abundance of seemingly unnecessary additions of ‘but X didn’t V’ clauses in the glosses of many irrealis-avertive inflections in grammars of Australian languages. I wondered—why did informants feel the need for providing such translations? I now take it to be an indication that negative event implications are much more salient in Australian languages than in so-called ‘Standard Average European’ languages for past irrealis forms, even with synthetic irrealis-avertives of type 1 (column 3 in Table 6.2).

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

(118) Il n’a pas pu partir. He NEG.have-3sg.PR NEG be.able.to-PP leave-INF. ‘He failed/wasn’t able/couldn’t bring himself to leave.’

167

(French)

(119) Il a voulu partir. (French) He have.3sg.PR want-PP leave-INF. ‘He tried to leave (and failed)’ (lit.: ‘he wantedperfective to leave’) (120) Il a dû partir. (French) He have-3sg.PR have.to-PP leave-INF. ‘He was compelled to leave’. (presupposes the agent is unwilling to act) Additional evidence for grouping together actuality entailments (AEs) and avertives can be found in their relationship to epistemic modality. It has been frequently observed following Hacquard’s seminal work on AEs that AEs are limited to action modals, i.e. to modals requiring dynamic, non-stative event predicates (cf. e.g. Hacquard 2009). Such a fact is in line with the observation made above that avertives do not associate with epistemic modal meanings (only doxastic modal and evidential meanings). I will get back to this issue further below, and provide a principled explanation. The point just made above natural leads to a central empirical and theoretical generalization concerning the relation between modality and avertivity. I have established that two main subtypes of modalized events, corresponding to two broad modal classes, are realized in the Australian data discussed above: dynamic vs non dynamic modals. Most structures belonging to the first class gives rise to an ‘exertion’, attempt reading, except what I have called reproachatives (based on deontic modal meanings), and are generally rendered using ‘tried [in vain]’ or ‘couldn’t/was unable to’. They are associated with capacity, volitional/(agentive) teleological modal meanings (and again, deontic modal meanings for reproachatives). The second class is associated with mistaken thoughts/beliefs-expectations/perceptions (doxastic-predictive modals and evidentials) and plain non-agentive proximatives (‘was about to but didn’t’, ‘nearly V-ed’) also reflecting on a simple failed expectation. They occasionally have extended meanings crosslinguistically associated with such semantic categories, e.g. mirative interpretations (see Section 6.5.1.3), and of course, they never incorporate an exertion event. They can only denote cognitive state events, as they are not so-called ‘action modals’; only avertive structures involving the latter type of modal meaning can become associated with some agentive attempt event.

168

PATRICK CAUDAL

In contrast to conventionalized avertives, I have also established that at least some irrealis inflections have a merely pragmatic avertive effect (Section 6.4.1), so that avertivity should be treated on several, distinct levels of a theory of meaning and time: in the semantics, or at the semantics/pragmatics interface. Again, such forms should be contrasted with their re-entrant uses in more conventionalized avertives such as e.g. what I have referred to as periphrastic irrealis-avertive inflections (Section 6.4.2), as they have a semanticized avertive content, not to mention other conventionalized types of avertive constructions involving negations (Section 6.4.4). An open theoretical question at this point is whether these semanticized, former defeasible implicatures should be treated as conventional implicatures à la Potts (2005; 2007), i.e. using a multidimensional semantics, or as straightforward semantic content. This opens up further interesting avenues of research along the lines of several dimensions of meaning being associated with avertives. Finally, I would like to stress that the pre-formal analysis sketched above differs from that of the cem avertive in Papago (Copley and Harley 2014) in substantial ways. Copley and Harley assume that accomplishment verbs are by default non-culminating in Papago (vPs do not encode a presupposition of efficaciousness, in their terms), unless some specific marker intervenes. The present theory diverges in that it does not consider that partitive culminations (PCs) should necessarily share with avertives a common semantic mechanism; I believe it to be rather unlikely, given the complexity, and specific properties (notably their multiple modal/evidential properties, which PCs lack), of the development paths of avertives. I rather regard PCs as a case of avertive strategy, distinct from bona fide avertives (my analysis is closer to Kroeger 2017 in that respect). Also, there appears to be some cross-linguistic variation as to the nature of e.g. avertive readings of irrealis inflections,¹⁰¹ and—even more importantly—the distinction between ‘emphatic’ vs ‘nonemphatic’ avertive structures suggest that Copley and Harley’s analysis would need at the very least to be enriched to apply to Australian languages. In my

¹⁰¹ As we have seen, Jaminjung and Iwaidja potentially differ on this. Speakers of Iwaidja seem rather inclined to consider irrealis avertives as at least having a failed reading by default—while the reverse holds true in the absence of any avertive morphology/construction. But that does not seem to be the case in other languages, or this could depend on the kind of indicative morphology used; this complexity needs to be accounted for. Also, our observations about tense insubordination being a potential factor in the advent of PCs (as in Anindilyakwa and Kayardild, with their ‘weak perfective tenses’ being former dependent, relative tenses) suggest that PCs are probably not determined only by aspectuo(causal) parameters (‘forces’ for Copley and Harley), but about aspectuo-temporal parameters (and so-called relative tenses), and therefore that focusing on culmination presupposition isn’t probably going to be sufficient. There is something deeper about the way those inflections are capable (or not) of contextual variation, than a culmination presupposition and a force-dynamics à la Copley and Harley.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

169

view, much of the problem at stake, as far as Australian languages are concerned, has to do with introducing a variety of types of multi-dimensionality in the semantic/pragmatic analysis (are these conventionalized implicatures something different in some cases and in others, etc.)—it is unlikely that a single theoretical solution can be applied to all cases, even in our relatively limited language sample. And whether or not Copley and Harvey’s solution is applicable to Australian languages largely depends on whether or not all languages possessing avertive structures allow for partitive culminations; it is far from obvious. At the very least, this is a matter to be settled as part of the semantics of inflectional morphology in some languages, rather than as part of the verbal lexicon—cf. the weak vs strong perfective tenses of Kayardild and Anindilyakwa. The Australian data on partitive culminations being so scarce, it is difficult to be adamant on that matter, and it is probably wiser to leave this as open question, both empirically and theoretically.

6.5.3 (Possible) wider consequences for a cross-linguistic and a socio-cultural theory of time Before closing this investigation, let me add a few considerations about some possible wider implications of this survey for a theory of time, primarily from the perspective of language typology, and to a lesser extent from a sociocultural standpoint—assuming that languages, as socio-cultural constructs, may reflect on certain long-term, social properties of said groups. Let us turn first to an obvious cross-linguistic question. A striking crosslinguistic fact is that while avertive forms are widespread in the grammar of e.g. Australian languages—as shown in this study—but also Amazonian and Indigenous languages of the America, as well as e.g. languages of PapuaNew Guinea, they seem to be mostly lexicalized (as verbs or as constructions) in many ‘Standard Average European’ languages, especially Romance and Germanic—see e.g. Schwellenbach (2019). Vice versa, constructions expressing actuality entailments (which are polar opposites of avertives, which I have dubbed above inactuality entailments) seem to be absent in Australian languages—at least on the basis of my language sample. Why this difference? The development paths we have uncovered in Australia suggest that a predominantly inflectional—either synthetic or periphrastic, including as combinations of particles with certain inflectional modal markings—kind of grammatical system is used to convey avertivity. If one assumes that modal inflections are in fact event predicates of a special kind (i.e., they denote a

170

PATRICK CAUDAL

modal state)—a view we have claimed (see Section 6.4.2) is substantiated by the diachrony of TAM/pronominal prefixes in non-Pama-Nyungan languages (Osgarby 2018)—and if we bear in mind that several past irrealis are clearly derived from past imperfective tenses (cf. Green (1995: 195 ff.; 2003: 399) and note 90), then it would make sense to view Australian past irrealises as TAM forms combining a modal stative predicate (conveying e.g. a capacity, expectation, or desire state) with a past imperfective content. The only extended class of avertive structures where past perfective readings appear (though not even exclusively, as iterative contexts with licence imperfective morphology are attested, and past irrealis marking is also possible in several languages of the sample), are past indicative telic verbs combining with certain proximative adverbials, cf. Section 6.4.2 (it should even be noted that our sample is in fact deprived of bona fide perfective tenses, and only offers aspectually underspecified tenses contextually capable of perfective readings—even when a grammatical description labels some tense ‘perfective’ or ‘punctual’, cf. the Kayardild so-called perfective, which turned out to be aspectually underspecified). In contrast, it appears that in at least some Romance languages, perfective marking is used to convey some avertive constructions—cf. the ‘want’ avertives in French, illustrated in (7), and comparable modal constructions in other Romance languages (e.g. Italian and Spanish, especially certain varieties, cf. Sinner and Dowah 2020), with a compound past marking bearing on a modal verb or construction, followed by an infinitive (cf. Argentinian Spanish haber de + INF). Like actuality entailments, Romance inactuality entailments (i.e. avertive meanings) appear to often involve a perfective viewpoint meaning, combined with a modal verb or construction, i.e. denoting a stative event predicate—which suggests that some kind of aspectual coercion is involved (this was independently suggested in e.g. Homer 2011). Obviously, no such perfective viewpoint and aspectual coercion can be identified in Australian languages with inflectional irrealises, as they generally derive from imperfectives. This might explain why actuality entailment forms are so far undocumented in the grammar of Australian languages—interestingly, Romance inflectional conditionals most certainly do not give rise to actuality entailment readings for the very same reason: they are also derived from imperfective forms. In a nutshell, avertives are essentially a category of ‘events/expectations/ desires/beliefs gone wrong’; they are frequently connected with so-called apprehensionals or aversives, as are past irrealis morphemes in general, crosslinguistically (Vuillermet 2018; Sansò 2020)—this is obviously a direct

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

171

reflex of their negative orientation. They essentially involve some frustrated intention or failed expectation—which may or may not associate with a negative feeling or evaluation, but regularly associates with mistaken beliefs or surprise, which tend to be negatively viewed. This suggests a strong connection of the category with affect, feelings, and therefore cognitive representations of social engagement with reality. Avertives are forms expressing the limited cognitive and causal abilities of human agents with respect to Time seen as a flow of events. Given their overabundance in the grammar and lexicon of Australian languages, they might be imbued with a particular signification in the way said cultures envision time—the prominent manner in which such attitudes to events are grammaticalized in certain languages might indicate that such concepts play an important role in the associated cultures. Indeed, it is tempting to establish a link between their prevalence in certain language areas, their rarity in others, and the way time and the individual were connected in the historical cultures having given rise to said languages.¹⁰² However, investigating such issues would constitute a large comparative ethnolinguistic research project in its own right, and clearly falls without the purview of the present chapter. But regardless of the cultural significance (or lack thereof ) of such facts, it seems intuitively obvious that beliefs and desires in general permeate our perspective on time—volitional, predictive, and epistemic attitudes, of course, but also beliefs about there being an agentively ordered, i.e. humanmind like, organization of causes and effects (cf. the classical ‘argument from design’ in the organization of the universe, and the Leibnizian ‘watchmaker analogy’-based conception of time). Attributing an agentive direction to the organization of events in time is also a well-known cognitive bias in psychology, where the belief that things ‘happen for a reason’/need to make sense, permeates many ordinary thoughts. Avertivity is obviously related to such a cognitive mechanism, and reflects on its complementary facet: namely that things don’t always happen according to human agents’ thoughts, beliefs, or desires, because the way time and events unfold can be non-directed (i.e., may ¹⁰² Without necessarily supporting a classic Whorfian approach to the interaction between our perception of time and culture, one could hypothesize that avertive systems flourished among societies who shared significant cultural attitudes towards time. It might be plausible that among societies having preserved a ‘mythical’ conception of time where human beings have a more limited grasp of, and control over the flow of events, i.e., in societies where an individual’s intentions, perception, and beliefs about the flow of time are culturally perceived as precarious—and this might prove fertile ground for the frequent use of forms conveying meanings reflecting such culturally salient perspectives over the flow of time; but this does not explain why the relevant forms (e.g. avertives) came into existence in the first place, though—it would only account for a greater ease of grammaticalization of these forms.

172

PATRICK CAUDAL

not be related to a goal or plan) at least from our limited, human perspective; it may even be (or seem) impacted by something like chaos in a radical way (where chaos can be regarded as an absence of agent-controlled causal ordering).

6.5.4 The final word From a strictly linguistic point of view, I have here attempted to show (I hope without said attempt entailing a failure!) that important semantic and pragmatic regularities are involved in the various types of avertivity found across Australian languages, and that these regularities have meaningful consequences for our understanding of the cognitive architecture underlying the perception of time in such languages. In particular, I hope to have established that avertives in Australian languages involve complex event structures with a strong modal component, with a positive component of meaning (typically a belief/expectation or a desire, and/or a fragment of a positive event) and a negative part (a negative event, and underlyingly, the negation of the previously held belief/expectation/desire). Forms incorporating a semantically negative event element (negation, or a dedicated avertive affix/clitic/particle or adverbial) were shown to be semantically avertive. They contrast with synthetic inflectional irrealis-avertive forms for which the negative event information (including the failure for result states to obtain) is essentially pragmatically derived, and context-dependent– cf. e.g. (30), where context enforces a non-avertive reading of the volitional past irrealis. Several periphrastic avertives incorporating a synthetic avertive inflection appeared to constitute marked/emphatic semanticized counterparts of said pragmatically avertive inflections. Only semanticized avertives denote complex, modalized event structures of the type evoked above. But what of a philosopher’s concerns for time? The above facts and their analysis have certainly shed novel light at least on semantic/pragmatic (i.e. linguistic and cognitive) dispositions specific to Australian languages for construing time—and beyond that, for languages in general, as I have attempted to highlight some crosslinguistic common points, but also divergences w.r.t. avertivity.¹⁰³ Understanding the structuring properties of the human mind in ¹⁰³ I believe that even if we leave aside the issue of the negative events and their relation to positive events, evidence for such ontologically complex events are of obvious interest for any philosopher with at least a passing interest the semantics of tense and aspect, and its role in determining how the human mind conceives of time.

AVERTIVE/FRUSTRATIVE MARKERS IN AUSTRALIAN L ANGUAGES

173

its universality (as an abstract, cognitive organization), as well as its specificity (as e.g. is manifest in cultural systems, and possibly linguistic systems) is, I think, key to such an endeavour. A theory of how the ‘human mind’ relates to time, can only benefit from being linguistically informed, and as such, will necessarily hover between language-specific properties, and abstract, universal properties of language as a general human ability—and if not universal properties, at least cross-linguistically frequent properties. Indeed, even languagespecific categories resort to more basic semantic primitives or contents, such as e.g. imperfectivity, perfectivity, iteration, and so forth—and despite important variations, these notions remain both comparable and common across languages. It seems obvious that a certain sliver of objective time qua actual history is not immune to a myriad private states entering the minds of human agents; avertivity constitutes a rather complex means of communicating such private states concerning the flow of events. If we put aside the question of time as a purely abstract structure (or a dimension of the universe), then surely, a theory of time must pay attention to the way the human mind gives it substance—as history, as myth, i.e. as narratives. And the nature and organization of languages, standing at the crossroads between cognition and culture, very much determine the way we construe this other, decidedly human time. Last but not least, this chapter has also been an attempt at providing a practical illustration of what I take to be an obvious desideratum for linguists and philosophers alike, namely that looking at rich, complex data originating in seemingly exotic languages, can be essential when trying to decipher theoretical questions of a certain importance and complexity. This, I believe, is all the truer with looking at such central and thorny questions at the identification of the properties of time with respect to the human mind—insofar as linguistic systems can shed light on the latter, of course.

7 On modelling the future M. Joshua Mozersky

7.1 Introduction One of the challenges besetting any attempt to understand the nature of time arises from the fact that the future, a significant and important portion of the domain under consideration, is, in the absence of backward causation, unobservable. After all, by the time the future is experienced, it has become present or past, which leaves open the possibility that it has fundamentally changed in the interim. Hence, the structure of the future must be: (i) deduced from the properties of experienced time; (ii) inductively extrapolated from such past experience; (iii) derived from a priori principles; or else (iv) posited, abductively, as the best explanation of what we in fact experience and know. Otherwise, the general structure of time must elude us. Options (i) and (ii) appear to be ruled out by Hume’s problem of induction. Whatever the quantity and quality of our accumulated temporal experience, its teachings will be unable to either logically or empirically rule out a future that fails to resemble the past (Hume 1978 [1739], Book 1, part iii, section 6). Accordingly, neither deductive nor inductive reasoning will determine a unique structure for the future or rule out the possibility that it will have changed by the time we observe it.¹ ¹ There are, of course, many proposed solutions to Hume’s problem, which tend to fall into three camps. First, there are those that evade the problem, such as Popper (1971), who urges the replacement of inductive reasoning with the deductive method of conjecture and refutation. Secondly, there are those that justify something else, for example, Reichenbach (1938) who doesn’t justify induction but, rather, our acceptance of it, on the basis that it is better, or no worse, than any other reasoning, which leaves open the question as to whether its conclusions are any good (‘best’ does not entail ‘correct’). Alternatively, Strawson (1952) argues that induction is the standard of empirical reasoning, so it makes no sense to ask for its justification, any more than it makes sense to ask for a legal justification of the law, which similarly does not entail that it is a good standard (we cannot assume our laws our sound just because they set the terms of legal argument). Thirdly, there are probabilistic solutions, such as Bayesians who offer rules for updating beliefs in light of new evidence, which will lead to convergence of beliefs over time, but only if we do not change our probability assignments in the interim, which begs Hume’s question (see Hacking 1967, Howson 2000; there are responses to this, but no consensus of which I am aware). Some suggest we replace inductive reasoning with behaviour based on statistical calculations (Mayo 1996), but these arguments typically assume such things as stability, i.e., that

M. Joshua Mozersky, On modelling the future. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © M. Joshua Mozersky (2023). DOI: 10.1093/oso/9780192896445.003.0007

ON MODELLING THE FUTURE

175

The most famous proponent of option (iii) is Immanuel Kant, who posited that time and space are forms of thought and perception respectively. Accordingly, the mind can uncover the general structure of time via a thoroughgoing introspective exercise, i.e., by an investigation into pure reason (Kant 1999 [1781/1787]). If Kant is right, however, then it is impossible to account for human cognition in scientific terms because empirical processes, including evolution by natural selection, will be the products rather than the progenitors of our cognitive architecture. Kant is explicit in his scepticism regarding the possibility of explaining cognition scientifically: For it is quite certain that in terms of merely mechanical principles of nature we cannot even adequately become familiar with, much less explain, organized beings and how they are internally possible. (Kant 1987 [1790]: §75)

However, in taking the route that he does, Kant is, arguably, exchanging one unknown—how did organized life arise?—with another—how does the nonempirical mind construct empirical reality? The latter ‘process’ would have to be, for Kant, both non-Darwinian and pre-temporal so not a process in the ordinary sense. It is not at all clear what kind of explanation is available in these terms, and while the biological explanation of life is a truly difficult problem, it does not appear to be intractable in principle. It is, to be fair, possible that human beings are extra-empirical in nature. Less extravagantly, we may be inexplicable in scientific terms. Noam Chomsky, for example, has argued that the creative aspect of language use puts its explanation outside the scope of psychological and biological science: The theory of evolution places humans firmly within the natural world, taking humans to be biological organisms, much like others, hence with capacities that have scopes and limits, including the cognitive domain. Those who accept modern biology should therefore be mysterians [about explaining human nature scientifically] (Chomsky 2016: 56)² repeated trials tend towards a fixed value, an assumption targeted by Hume (see Hacking 2001). Moreover, even if we can assign probabilities to propositions about the future, this is compatible with a future whose temporal structure differs from the past. Accordingly, I take Hume’s problem as unsolved here. ² Chomsky accepts the possibility of a scientific account of language—structure, syntax, grammar, etc.—but thinks its use is neither systematically nor predictably connected to external stimuli, so resistant to standard scientific explication (Chomsky 2009).

176

M. JOSHUA MOZERSKY

Even if correct this would, however, fail to solve the problem I have in mind because, as Einstein’s Theory of Relativity has made clear (Einstein 1916), more than one kind of temporal structure is compatible with human thought and experience. Relativistic time is non-Newtonian and counter-intuitive, but mathematically coherent and predictively successful. So, even if time is somehow a mental construct, as Kant argues, the mind is clearly capable of generating distinct and incompatible temporal structures. Accordingly, the a priori attempt to read off the structure of time from the structure of thought appears unlikely to deliver a single, coherent result.³ So, (iv) is the most promising way forward, and in what follows, I argue that we can make substantial progress towards understanding the general structure of time by adopting this explanatory approach. To this end, I shall subject one prominent account of the structure of the future, the branching model, to scrutiny and conclude that it suffers from a kind of internal incoherence. If this is correct, then we can advance towards our goal of understanding time by eliminating this important contender. I conclude with some reflections on a more promising alternative.

7.2 Three principles of temporal experience A particularly prominent aspect of experience is that time is a dimension correlated with a pervasive epistemic asymmetry. We can use our senses to learn about the present and consult many kinds of physical record to gain detailed knowledge of the past. A comparative level of detail concerning the future is rarely available and can only be extracted from predictions with substantial effort that fails more often than not.⁴ No such asymmetry attaches to any spatial dimension: (EA) Epistemic asymmetry: the future is knowable in neither the same way nor to the same extent as the past or present.

³ Perhaps Kant’s arguments only apply to our ability to envision time, or order our thoughts, which, even if necessarily Newtonian, is compatible with Relativistic representations of the world. I do not object to this understanding, but it concedes my point, which is that conflicting theories of time itself are compatible with our experience, so we cannot determine a unique temporal structure from reflection upon the structure of thought. ⁴ The laws we use to predict the future must have held in the past for otherwise we would not know them. Accordingly, since we know so much more about the past than the future, we may conclude that we have ways of learning about the past that are distinct from projecting general laws; either that or we learn about the past by unknowingly forming generalizations that we are unable project into the future. Either way, (EA) below holds.

ON MODELLING THE FUTURE

177

A second pervasive feature of experience is that time is the dimension of change. Spatial variation (object o is F at point p1 but not-F at point p2 ) is logically analogous to change (o is F at time t1 but not-F at time t2 ) but is nonetheless compatible with the absence of change. Temporal variation, on the other hand, is change: (CT) Change is temporal variation: if what is the case at one time is not the case at another time, change has occurred. It is, thirdly, often remarked that, as pervasive as change is, one cannot alter the past; the past is fixed. There is certainly nothing in experience that conflicts with this suggestion, but there is no need to test it because the fixity of the past is simply a matter of logical consistency, so it must govern all observation. To see this, suppose, to borrow a famous example (McTaggart 1927), that a fireplace poker is warm at t1 . Even if the poker is cool at t2 , it remains the case thereafter that the poker was warm at t1 . To suppose otherwise is to believe that at some moment after t1 it becomes the case that the poker was not warm at t1 , which leads to contradiction: it is both the case, and not, that the poker is warm at t1 . Hence, we can, perhaps trivially, include the fixity of the past as a feature of temporal experience: (FP) Fixity of the past: If x is F at t, then it remains the case at all later times that x is F at t. My proposal is that we can gain some traction in understanding the general structure of time by considering how well a model of the future accounts for and coheres with these three principles.

7.3 An ontological explanation—branching To some thinkers, the fact that the future is unpredictable combines with the fixity of the past and the reality of change to strongly suggest that the future is radically distinct in structure from the rest of time. Suppose, for example, that the future is ontologically indeterminate, so that there is simply no fact of the matter as to what lies ahead, but as time passes that which is future and indeterminate becomes present by becoming determinate. That would seem to explain why we cannot reliably predict what will happen (there is nothing determinate to predict), why the past is fixed (the present remains determinate when it becomes past), and why change is temporal in nature (the transformation from indeterminate to determinate occurs along the temporal dimension).

178

M. JOSHUA MOZERSKY

In short, an ontological distinction between past/present and future would appear to explain our experience of time. One such approach has been explored by philosophers such as John MacFarlane (2003), Storrs McCall (1994), Richmond Thomason (1970), and Nuel Belnap et al. (2001). It is often referred to as ‘branching time’ or the ‘branching model of time’, and it entails that the past is a single, fixed timeline while the future consists of many concrete timelines, which can be pictured as branches fanning out from the present moment, which is the highest point of the fixed past: (BT)

At any given time there exists only a single past timeline (the trunk) but many distinct, concrete future timelines (the branches), each of which is compatible with history up until the branching time-point.

On this view, there is no fact of the matter as to what will happen next because the array of future histories contains all possibilities: some branches will contain a sea battle, others not, for example. Hence, utterances that express contingent propositions about the future are neither true nor false: (SI)

Semantic indeterminacy: future contingents are neither true nor false.

The passage of time is usually represented as branch attrition, i.e., the continuous falling away into non-existence of all future branches except for one. Semantic indeterminacy is not obviously a feature of experience, for it is often thought that we can make true statements about future contingencies, e.g., ‘the sun will rise’; nonetheless, it is a forceful idea that is at least suggested by epistemic indeterminacy.

7.4 The semantics of branching time I mention above that the fixity of the past is matter of logical consistency, but contradictions can be eliminated easily, by adding a parameter. When A utters ‘S is tall’ and B ‘it is not the case that S is tall’, they are not necessarily contradicting each other since A might be asserting that S, at six-foot-three, is a tall person while B is asserting that S is not a tall professional basketball player. This tells us that ‘x is tall’ is in fact a two-place predicate, relating x to a comparison class, which constitutes the additional parameter.⁵ When superficially conflicting sentences seem capable of joint truth, we typically assume there is ⁵ Similarly, Einstein replaces ‘x is simultaneous with y’ with ‘x is simultaneous with y in frame f ’.

ON MODELLING THE FUTURE

179

a hidden parameter. If our aim were to logically untether the past from fixity, then we would simply need to add an appropriate parameter, which would be equivalent to assuming that predicates are doubly temporally relative i.e., that ‘the poker is warm at noon’ is an incomplete sentence that in fact expresses a three-place relation, the poker is warm at noon at t∗ . The double relativity strategy would have us replace ‘x is F at t’ with ‘x is F at t at t∗ ’, or: (DR)

F(x, t, t∗ ).

The thought, then, is that any instance of (DR) is true when x is F at t and t = t∗ , but not necessarily otherwise, giving us the following semantic scheme: (DRS)

(∀t∗ )(∀t)(∀x) [‘F(x, t, t∗ )’ is true if (t∗ = t and F(x, t))].

Other than offering a formal escape route from the fixity of the past hypothesis, it is not easy to see what recommends (DRS) as a general account of temporal predication. In particular, if the poker can be both (a) warm at noon at noon and (b) not warm at noon at 1:00 (however we might make sense of that), it remains unclear how the propositions that result from filling in values for t∗ are supposed to vary in truth value. Assuming all cases where t∗ = t are true, which values of t∗ ≠t lead to truth and which to falsehood? From which temporal perspectives, precisely, must our judgements of the past be altered? Assuming that it can’t be all of them—sometimes the past remains as it was— the question is which criterion distinguishes the cases. John MacFarlane suggests one, by appeal to (BT), arguing that the truthvalue of an utterance is not only relative to the temporal context of utterance but also to the temporal context of assessment (MacFarlane 2003). The underlying problem that motivates his proposal is that of making sense of future contingency, the thought that more than one course of events is compatible with the current state of the world (related, in obvious ways, to (EA)). One way to cash this idea out is in terms of semantic indeterminacy. On the one hand, MacFarlane wants to preserve the thought that an utterance, at t, of, e.g., ‘there will be a sea battle tomorrow’ should count as neither true nor false, since at that time there is a possible future history that includes the battle and one that doesn’t and nothing in the context of utterance distinguishes between them (MacFarlane 2003: 323). On the other hand, he accepts that, a day later, there is equally good reason to suppose that the same utterance should be reassessed as having been determinately true or false, on the grounds that a sea battle either did or did not happen as predicted (MacFarlane 2003: 325); if, at t, one expresses the proposition that there will be a sea battle at t + n and at t + n

180

M. JOSHUA MOZERSKY

there is a sea battle, there is no good reason to deny that one spoke truthfully the day before. The question is whether these can be reconciled. MacFarlane suggests that an utterance is true if the time of assessment is part of a history that includes a truthmaker for the original utterance; false if the time of assessment is part of a history that lacks a truthmaker for it; and neither true nor false otherwise. So, if u = ‘there will be a sea battle tomorrow’ is uttered at t0 , and future history t1 includes a sea battle the next day while future history t2 does not, then (MacFarlane 2003: 332): i. At context of assessment, a = t0 , u is neither true nor false ii. At a = t1 , u is true iii. At a = t2 , u is false. In short, propositions concerning contingent future events are neither true nor false at the time of utterance because nothing about the context at that time settles the matter; but depending on how things unfold, that same utterance will become determinately true or false from the point of view of a future time. This is a form of double temporal relativization that provides a principled way of assessing the semantic values of sentences given different values for t∗ , the time of assessment, in (DR). Letting u be a token utterance of the form ‘x is F at t’ (e.g., a sea battle is fought tomorrow),⁶ then we have the following general principles for future contingents: S1. (∀t∗ )(∀t)(∀u) (u is true iff t∗ ≥ t and x is F at t); S2. (∀t∗ )(∀t)(∀u) (u is false iff t∗ ≥ t and ~(x is F at t)); S3. (∀t∗ )(∀t)(∀u) (u is neither true nor false iff t∗ < t). MacFarlane’s view is a partial rejection of (Fixity). Once the relevant future time has come to pass, a past prediction and its then future referent remain part of world history thereafter, so no further variation in truth-value occurs (this is a branching future, not a branching past). However, between the time of utterance and the time of the predicted event, a past utterance can change from ‘indeterminate’ to ‘true’ or ‘false’, so the past is not semantically fixed, and the reason for this is ontological: what was future, hence indeterminate, becomes past, so determinate.

⁶ I am suppressing the tense of the copula here because the branching future account includes all future conditions and, accordingly, models tense as reference to the branching of the world from the point of utterance. Put another way, the future tense is considered to be a species of semantic indeterminacy, which is captured by S3 whether or not the copula in u is considered to be tensed. Accordingly, we can, for simplicity, work with tenseless predications.

ON MODELLING THE FUTURE

181

7.5 Branch attrition Hindsight bias is the tendency to increase the estimation of an event’s probability upon discovering that it occurred. Prior to the event, when it was uncertain, it is common to assign it a lower probability than retroactively, merely from learning that it in fact happened. In short, we assume that what has happened was highly probable and predictable, if not inevitable, despite the fact that we made no such assumption at the relevant earlier time: Finding out that an outcome has occurred increases its perceived likelihood. Judges are, however, unaware of the effect that outcome knowledge has on their perceptions. Thus, judges tend to believe that this relative inevitability was largely apparent in foresight, without the benefit of knowing what happened. (Fischoff 1975: 297)

Psychologically speaking, part of making sense of history seems to be a matter of telling ourselves that ‘we knew it all along’ or ‘it was bound to be’. With respect to the branching model of time, something similar occurs, but in metaphysical terms.⁷ This ultimately leads to a meta-theoretical difficulty. To see the problem, notice that since the model includes utterance reassessment, it entails that if a sea battle is occurring, yesterday’s utterance of ‘there will be a sea battle tomorrow’ is now assessed as true: ‘When we take this retrospective view, we are driven to assign a determinate truthvalue to [yesterday’s] utterance’ (MacFarlane 2003: 325). So, it is not the case anymore that the utterance was indeterminate. But this is a problem, for the model also includes the fact that the utterance was indeterminate when it was made, and we can say this even today, from a point of view that includes a sea battle. In other words, the model requires us to accept both that the future was semantically indeterminate yesterday and that it is now the case that the future was semantically determinate then. Just as hindsight bias entails the deletion of past uncertainty, utterance reassessment involves the deletion of past indeterminacy, and this leads to a conflict: as past determinacy is expressed from the vantage point of the present, so too is its absence. Note that the appeal to hindsight bias is meant to be instructive so it is not important that this conflict be understood as an instance of bias or any ⁷ The idea is that the metaphysical picture supports or entails a semantic thesis directly, though this need not always be the case.

182

M. JOSHUA MOZERSKY

other form of fallacious reasoning.⁸ What is important is that this is a conflict that is internal to the model. MacFarlane suggests that we reconcile intuitions about semantic determinacy and indeterminacy via utterance reassessment, but since the semantics are grounded in a particular ontological model— branching—it follows that such reassessment must transfer to the underlying ontology as well. Accordingly, not only are semantic values reassessed but so is the structure of the model as a whole. This can be made clearer if we focus on branch attrition (MacFarlane 2003; McCall 1994). Suppose that at time t0 ⁹ the world consists of a branching structure as depicted in Figure 7.1; from t0 , two histories, t1 and t2 , branch out.¹⁰

t1, sea battle

t2, no sea battle

t0, u

Figure 7.1 Branching time

As one branch becomes actual with the passage of time, the other disappears from the picture. Otherwise, the passage of time, and the conversion of indeterminate predictions into determinate truths, will simply not appear in the model. If there is no branch attrition, then the universe is identical from all three perspectives: t0 , t1 , and t2 and nothing will have changed, so nothing will justify the retroactive change in truth value of u. So, if the passage of time is to convert semantically indeterminate utterances into determinate ones, then the branching structure must change with time. ⁸ Hedden (2019), for example, argues that upon learning that a previously uncertain outcome occurred we should retroactively increase our probability assignment because we now know all contrary evidence at the time was misleading and would have been rightly ignored. ⁹ Some criticize this model on the grounds that it implicitly requires two temporal dimensions, for example, t0 represents a model of time as a whole as well as a moment within it (see, e.g., Smart 1980; Nerlich 1998; Farr 2012a); McCall replies that the second dimension is not time ‘but instead “generates” time’ (McCall 1976: 348). I am sympathetic to the initial criticism, but I do not want to rule out twodimensional theories of time (who knows how complex reality is?). Instead, I intend to argue that there is a conflict within the branching model itself, which would arise even if our interest were in defending multi-dimensional time. ¹⁰ This is, obviously, a simplified diagram, but nothing rests on the branches that have been left out.

ON MODELLING THE FUTURE

183

The universe cannot be as depicted in Figure 7.1 at both times t and t1 , for then it would be incorrect to retroactively reassess u: from the context of t1 , Figure 7.1 models a reality in which u remains neither true nor false. What is needed instead is a world that is as represented in Figure 7.2, which is a model in which, from context of assessment t1 , u is determinately true.

t1, sea battle

t0, u

Figure 7.2 Branch attrition

The difficulty, however, is that Figure 7.2 is a model in which it is not true that u was indeterminate at t0 for there is no future contingency from t0 depicted there.¹¹ Accordingly, while the branching structure with branch attrition and context of assessment sensitivity can justify past determinacy, it cannot justify past indeterminacy. From all contexts of assessment, there is no ontological basis for the proposition that any past utterance was indeterminate.¹² So, the branching model is caught in a dilemma. It either contains genuine branch attrition, allowing for retroactive assessments of determinacy but then lacking the resources for assessments of past indeterminacy, or it lacks branch attrition, which enables assessments of past indeterminacy but not retroactive assessments of determinacy. Might the branching-futurist reject MacFarlane’s intuition and insist that the past utterance, u, remains untrue for all time? Since the ground for semantic indeterminacy is, on this model, the existence of multiple branches from the time in question, this response is available only if branch attrition is rejected. We can conclude that branching with attrition and context of assessment sensitivity create an unstable model. The defender of branching time insists ¹¹ To put the point visually: the structure represented in Figure 7.2 does not contain u at a branch point. ¹² Unless, of course, the future occurrence the utterance is about is also future at the time of assessment, in which case we are not dealing with a semantically determinate utterance, so such cases are not relevant here.

184

M. JOSHUA MOZERSKY

that branches disappear, which is what changes an ontologically indeterminate future into a determinate past, and semantically indeterminate utterances into determinate ones. But the branches that disappear are the ontological grounds for future contingency, so without them there is nothing in the model that corresponds to past indeterminacy. So, for any given time, t, the model will not include any structure that corresponds to what the defender of the model wants to insist is a universal truth about the past, namely that, back then, the future was indeterminate.

7.6 Branching without attrition: Time slices It might seem natural, at this point, for the defender of branching to insist that both structures in Figures 7.1 and 7.2 are parts of a larger model, the first corresponding to the world as it is at t0 , the other to the world as it is at t1 . In that case, the model of the world as a whole is better represented as in Figure 7.3, where we have a complex array that includes both something that represents the world as it was at t0 , when u was indeterminate, and as it is at t1 , when u is true.

t1, sea battle

t2, no sea battle

t0, u

t1, sea battle

t0, u

Figure 7.3 Time slices

This model includes a collection of branching structures, one for each moment of time; in other words, a collection of time slices. Might this allow us to preserve the context of assessment sensitive semantics for future contingents, (S1)—(S3), since both the perspective of t0 and t1 are represented in the model?

ON MODELLING THE FUTURE

185

One thing to note is that the time slice picture will disappoint those wanting the model to include something corresponding to an absolute present, a nonrelational NOW, which is distinguished from all other times. As we can see, two presents are represented in Figure 7.3, t0 and t1 , and a full model would include indefinitely many more. This is why philosophers such as McCall (1994) opt for branching with attrition: attrition privileges the present as the unique moment lying between the fixed past and indeterminate future, and there is only ever one such point in the model. The time slice model fails to do that, as it contains as many presents as there are times, none absolutely present but each, instead, present relative to itself. In response, Kit Fine suggests that it is unnecessary to integrate time slices into a coherent whole. He proposes that one may, instead, adopt a fragmented model, according to which each moment of time is an isolated and independent reality (Fine 2006). This is intended to reconcile the otherwise conflicting thoughts that being present is absolute, something Fine takes to be suggested by experience, and that absolute presentness contradicts absolute non-presentness. If we combine these with the assumption that the world contains, say, two moments of time, t and t∗ , then each would be present (at one time) and not present (at the other); if we cannot relativize being present, this would entail the logically incoherent, ‘t is and is not absolutely present’ (the same holds for t∗ ).¹³ If, on the other hand, each moment were to constitute its own reality, then, since there would be no other times in that reality, each would be present absolutely, and its being present would be incapable of conflicting with the being present of any other time, since each reality is independent of the others. In this way ‘presentness is both an absolute feature of reality and one that applies, across the board, to each and every time’ (Fine 2006: 406).¹⁴ Fine’s proposal replaces a single reality incapable of coherent description with a collection of mutually incompatible realities. In other words, it trades one kind of conflict for another. The payoff is to be the retention of an absolute present and concomitant robust temporal passage: ‘the passage of time can be taken to consist in the successive possession of the absolute property of being PRESENT or NOW’ (Fine 2006: 404). But the cost is high, for it remains an incoherent system as a whole. Is the evidence in support of the absolute NOW

¹³ This is analogous to a case in which relativization of ascriptions of tallness to a contrast class are disallowed. ¹⁴ Fine also offers a related, relativist, version of this view, according to which ‘The facts belong to different realms of reality, as it were, and these realms have some kind of independent status as the ‘loci’ of the facts (2006: 402). The differences between the views don’t bear on the discussion here.

186

M. JOSHUA MOZERSKY

sufficiently compelling to be worth this risk? I don’t think so, for the following reasons. The alternative to robust passage, which is absolute, is relative passage. On the relative account, ‘t is present’ is, like ‘S is tall’, a disguised relation: ‘t is present at t ∗ ’. According to Fine, this reduces presence to self-simultaneity or self-identity, ‘t = t∗ ’, which he argues entails that the successive possession of presentness amounts to nothing more than the fact that each time is self-identical: (TP)

(t1 = t1 ) & (t2 = t2 ) & (t3 = t3 ) & … & (ti = ti ) & …

Accordingly, the ‘proposed explanation of the passage of time collapses into triviality’ (Fine 2006: 404). But this misses an important point, which is that if presentness is relative, then passage does not consist simply of each moment’s possession of the trivial property of being self-identical but in the fact that times exist in a temporal order: (RP)

(t1 < t2 ) & (t2 < t3 ) & (t3 < t4 ) & … & (ti < tj ) & …¹⁵

So, we need to consider two questions. First, is relative passage trivial? Secondly, does experience sufficiently justify belief in an absolute NOW that we should reject (RP)? With regard to the first question, the fact that times¹⁶ are temporally ordered is certainly not a trivial fact. Consider that positions in space are not spatially ordered because any spatial coordinate system is as good as any other. For example, we would not be making any kind of geometric error if we were to choose something other than the equator as the zero point of the system of latitude; the direction from the equator to the North Pole is not geometrically distinguished from any other direction, in other words. So, that moments of time stand in a series ordered by ‘is earlier than’ is a genuine difference between time and space. Why it is the case that time is directed, but space not, is a question that I address below, but for now note that if it is true that moments exist in genuine succession, then we are not asserting something trivial, as we are when we assert that each moment is self-identical. This can perhaps be seen even more clearly when we note that time could exist in any number of mere partial orders. For example, a branching structure is a partial order in which not all times are temporally related. Since there are indefinitely many possible

¹⁵ Here ‘t.¹⁸ Or consider the past perfect, e.g., ‘a sea battle will have occurred by then’. For the B-theorist, what an utterance of this sentence expresses is true if and only if a sea battle lies at a moment between the time of speech/thought and ‘then’: (BT2) An utterance, at t, of ‘a sea battle will have occurred by then’ is true if and only if then = t” and there is a sea battle at t’ and t” > t’ and t’>t. In short, tenses are analysed as relations between the time of speech or thought and various reference points,¹⁹ but since all reference points exist somewhere on the timeline, there is no semantic indeterminacy, at least not for well-formed, grammatical sentences. The B-theorist needn’t, however, deny objective possibility, unpredictability, or contingency. This is because a single actuality is compatible with many alternative possibilities. In particular: (P1) (P2)

The past up until time t is deductively compatible with many future histories; The past up until time t is inductively compatible with many future histories.

Assume that the history of the universe up to this point, combined with the laws of nature, is compatible with only one outcome: the sun must rise tomorrow; in other words, suppose that it is nomologically necessary that the sun will rise. Nevertheless, it is logically possible that the sun does not rise tomorrow, because it is not a contradiction to assume that the laws change between now and then (for any finite duration, no matter how small, between ‘now’ and ‘then’). Moreover, no matter how much evidence we gather at any time, the built-in limitations of our cognitive systems, as well as of time and effort, ¹⁸ Some prefer to make the utterance itself, rather than its time of occurrence, the relatum of the last class, i.e., the utterance is true if and only if the sea battle is later than the utterance (Dyke 2008); for present purposes the differences between a time analysis and token utterance analysis are immaterial. ¹⁹ See Reichenbach (1947: §§50–51) for a classic treatment along these lines.

ON MODELLING THE FUTURE

191

entail that we can never be a hundred per cent certain that we have in fact completely described the state of the universe or its laws, so predictions are always tempered with some inductive uncertainty. Hence, (P1) and (P2). None of this is incompatible with a single future; none of this demands concrete branches. Here is why. Since branching models of time locate all indeterminacy in future structure, there is no way, absent backward causation, for such indeterminacy to be the cause of present uncertainty. Branches are an explanation of the unpredictability of the future, but they are not the source of it. Accordingly, even the branching futurist must admit that the reason we have trouble making predictions is not that the future itself is ontologically indeterminate but, rather, because our state of knowledge at a given time underdetermines what will happen next. But that means that the future will be unpredictable even if fully determinate so long as either the laws are compatible with distinct outcomes or our knowledge is incomplete, and we have excellent reason to believe that one or both of these conditions obtain. Consider that the world is, at least in many respects, a chaotic system. What this means is that, for many natural processes, unobservable differences in initial conditions between two systems can lead to observable differences later. Two measurements giving the same reading might in fact be of materially different configurations, leading to differences down the line. Suppose, for example, that a certain natural system will evolve very differently depending on whether the value of a particular parameter is 2.45, 2.46, 2.47, 2.48, 2.49, 2.50, 2.51, 2.52, 2.53, or 2.54. Our instruments, however, are only accurate to the first decimal place so that a reading of ‘2.5’ will display no matter which of these ten values is in fact instantiated. Accordingly, we can expect even our best measurements to be right only one in ten times. Given our best measurements, even an actual future remains unpredictable. Or, imagine that a dozen six-sided dice are placed in a sealed, opaque box, which is shaken vigorously and then placed on a desk. Assume, further, that even the best equipment available can only reveal the state of six of the dice inside the box. We are then asked what the remaining six dice show. We may assume that the state of the dice is ontologically determinate: they are macrolevel objects not subject to significant quantum effects, for instance. Still, our deepest investigation into the interior of the box leaves us with 66 possible configurations, so any guess will have a 1 in 46,656 chance of being right. What this shows is that ontological indeterminacy is not a necessary condition for unpredictability. It is our partial knowledge that is responsible for our difficulties here. If the dice were, in the future, to turn into quantum systems in a state

192

M. JOSHUA MOZERSKY

of superposition, that could not be what makes it the case that we are unable to guess at the configuration of the dice inside the box today unless that future state is somehow reaching back to the present and preventing our measuring devices from peering into the box. In other words, without backward causation, future indeterminacy cannot be the source of current unpredictability. Is such indeterminacy, however, the best explanation of the fact that we repeatedly fail in predicting the future? Certainly not if ontological indeterminacy cannot be coherently described, in which case some other explanation must be sought. Luckily, partial knowledge will do the trick. It is worth pausing to reflect on the demand for coherence. The problems besetting branching time are the result, I suggest, of a conflict in perspectives, one that is particularly pressing in the attempt to construct a model of time. What I have argued is that the branching model of future indeterminacy collapses because what its proponents want to say about time as a whole cannot be combined with their intuitions about what is the case at a given time. It is, in other words, a result of the inability to integrate all the individual facts-at-atime into a coherent sum of temporal facts. The model leaves irresolvable the conflict between the perspective on time with the perspectives within time or, as Nagel puts, between the ‘perspective of a particular person inside the world with an objective view of the same world, the person and his [or her] viewpoint included’ (Nagel 1986: 3). It is the attempt to combine these perspectives into a whole that leads us to the conclusion that the B-theory is right about the future: it is actual. If we were to limit ourselves only to the internal perspective, then there would be no need to combine all times into a coherent model, for we are only ever looking at the world from a given moment and a model that is globally incoherent may be locally coherent. MacFarlane takes note of this: From the point of view of an observer at [t1 ]²⁰ the actual future at [t0 ] held a sea battle, while from the point of view of an observer at [t2 ] it did not. And qua semanticists, we do not speak from the perspective of any particular moment on the tree of branching histories; instead we take a God’s eye point of view, looking down on the tree from the outside, and try to say how the truth of sentences depends on features of the context of utterance. From this external point of view, there is no sense to saying that one of two histories passing through a moment is ‘going to be the actual one’. It is only if we blur our vision, taking up internal and external perspectives simultaneously, that ²⁰ Symbols changed for consistency with the current chapter.

ON MODELLING THE FUTURE

193

it can seem to make sense to mark out one of the histories in the tree (as seen from above) with a thin red line. (MacFarlane 2003: 326)

So, he thinks it is a mistake, an artefact of the way we operate as semantic theorists, that gives rise to the inclination to accept an actual future. However, there are two considerations that suggest it is not an error at all. The first is that even within a temporal perspective, we regularly preserve the contents of our thoughts and utterances simply by modifying their tenses, which is what the fixity-of-the-past hypothesis captures. It is a perfectly natural thought, when witnessing a sea battle, to think back to yesterday’s conversation with someone who uttered ‘there will be a sea battle tomorrow’ and think ‘they were right’. Michael Dummett (Dummett 1969) classifies a set of truth-value links that connect the contents of thought and language across time: (TV1)

If it is the case that φ, then it will be the case that it was the case that φ;²¹ (TV2) If it is the case that φ, then it was the case that it will be the case that φ.

The branching theorist disputes (TV2), wanting the contingency of the future to entail semantic indeterminacy, so that even if there is a sea battle now, it was not true earlier that there would be one. But it is (TV1) that does the damage, for if it is the case that what is future is indeterminate, then it will be the case that what was future was indeterminate. The desire to preserve content across time arises from the point of view of someone in time who wishes to talk about what was once the case, an inclination that is both natural and suggests the existence of time as a whole. The failure of the branching model to combine future indeterminacy with past future indeterminacy renders it inadequate for the semantic theorist. The second consideration is that, in the attempt to construct a model of time, the inclusion of both the internal and external perspective is, I would say, the point of the endeavour. This is because we know that we not only speak and think about time but are embedded within it. Our perspective on time is part of time, so must be included in the model and if we cannot coherently combine the two, then we cannot coherently understand either one entirely: our understanding of time would be incomplete because something within it— our thoughts about time—would be missing; our understanding of ourselves would be incomplete because we would not understand something we think and talk about regularly. Hence, a consistent model of both is the goal, which is ²¹ This is, of course, just a version of the fixity of the past hypothesis.

194

M. JOSHUA MOZERSKY

why we should be hesitant to accept globally incoherent models. The B-theory, with its eternalist ontology, relational passage, and tenseless semantics offers a locally and globally coherent model that does no substantive injustice to our understanding of temporal experience. It is, accordingly, the preferable alternative to branching.²²

7.8 Branching, supervaluations, and the direction of time The most straightforward model of branching would be one with no branch attrition, no thin red line and no niche-like path, just a single unchanging array of branches capturing the structure of reality as a whole. What the foregoing discussion indicates is that it is the attempt to understand changes in the structure of time that leads to philosophical problems. This is not surprising because the attempt to understand alterations of the model are bound to conflict with the fact that time is a parameter within the model; a model of time runs into trouble if it also attempts to be a model in time. So, it is best just to do without attrition or time slices, either of which implicitly demand a higher order time dimension, as MacFarlane notes: We’ve already represented time as one of the spatial dimensions of our tree. So what could possibly be represented by the motion of a point on this tree? Certainly not a process that takes place in time, since all such processes are already represented spatially on the tree. (MacFarlane 2008: 86)

Accordingly, in more recent work, MacFarlane considers the semantics of tense in an unchanging branching model, comparing the supervaluationist and context of assessment relative approaches. The canonical supervaluational approach is due to Thomason (1970). Returning to Figure 7.1 above, there are three histories contained in that model: h0 = {t 0 }; h1 = {t 0 , t 1 }; and h2 = {t 0 , t 2 }. Thomason’s approach assigns truth values to u as follows: Val(u, h1 ) = T; Val(u, h2 ) = F; Val(u, h0 ) = undefined (1970: 274). If both h1 and h2 were to contain a sea battle, then a sea battle would be necessary and Val(u, h0 ) = T; if neither h1 nor h2 were to contain a sea battle, then a sea battle would be impossible and Val(u, h0 ) = F. However, genuinely contingent utterances about the future are neither true nor false, ²² The point might be put as follows: the B-theory integrates semantics and meta-semantics in a way the branching future cannot.

ON MODELLING THE FUTURE

195

since their referents exist only in some branches. In all cases, there is neither a thin red line nor a niche-like path. What is the supervaluationist to make of retrospective assessments of past utterances? Thomason’s valuations agree with MacFarlane’s intuition that because a sea battle is (or is not) occurring today, what one said in uttering u yesterday is, it turns out, true (or false) and should be evaluated as such today. What, however, should be said now about its truth value at the time it was made? On the one hand, the branches from t0 are an unchanging part of the model, so it seems right, today, to say that u was neither true nor false when it was produced since at that time it did not single out either h1 or h2 as part of its content.²³ On the other hand, that appears to conflict with our current assessment of what the utterance expressed as, ultimately, true (or false). MacFarlane thinks the supervaluationist can reconcile these opposing forces in one of two ways. The first is by appeal to the device of determinate truth, in which case what u expressed can be assessed, today, as true (or false) and also not determinately true (or false) yesterday (MacFarlane 2008: 96–8).²⁴ Alternatively, the supervaluationist might admit that the fact that what was expressed by u was semantically undefined at the time is, today, ‘ineffable from the “internal” point of view’ and requires recourse to technical semantic notions, such as context-relative utterance truth, to be expressed; in other words, we cannot capture what the past was like using only object language expressions referring to content (e.g., ‘what was said’) but meta-language expressions that range over object language entities such as utterances and devices such as disquotation (MacFarlane 2008: 97). Regardless of the approach one takes here, difficulties remain, which can be seen when we consider what it means to work with a fixed structure in which all branch histories remain fixed and unchanged in the model. There are three points worth noting. First, the model is ontologically extravagant. Without branch attrition, a thin red line, or a niche-like path, we each stand before indefinitely many future selves, all equally existent on distinct branches. This view is much like David Lewis’s well-known modal realism (Lewis 1986), in which a distinct, concrete reality corresponds to every possible state of affairs, or the many worlds interpretation of quantum mechanics (Lewis 2016), in which a distinct, concrete world exists for each possible outcome of a given event. There are, to be sure, arguments for such views, but they ²³ On Thomason’s view, at the time of utterance, u is open to either history coming to pass. ²⁴ This sounds a bit awkward in natural language to my ears—‘it was true but not determinately so’—but it makes sense if we just interpret it as a summary of Thomason’s three semantic valuations mentioned above.

196

M. JOSHUA MOZERSKY

are controversial. Moreover, they conflict with (EA), for if all branches exist unchanged, then the answer to the question, ‘what will happen in the future’, is, simply, ‘everything’. If every eventuality occurs, then there is no genuine unpredictability after all. Secondly, note that the fixed branching structure presupposes a temporal orientation (i.e., direction). There is no reason, inherent to the structure, why the direction of branching must represent later than rather than vice versa. Much as an arrow points from its tail to its head by convention, so that the shape ‘→’ could be taken to signify the right to left direction rather than the reverse, so an array of branches ‘points’ in either direction so far as internal structure is concerned. Consider Figure 7.4. A

B C

Figure 7.4 A branching model

Which direction is represented here, A to B/C, or B/C to A? Only something added onto this simple structure will tell us. Thomason makes no mistake here, defining future as the direction of branching without falling into the error of supposing this is entailed by the structure itself. Thirdly, and as a result, the model does not explain temporal passage, but must take it as given. If the structure itself fails to entail a temporal orientation, then it fails to entail a direction for temporal passage, which means that passage is underdetermined by the model. Accordingly, something must be added to the model to capture the passage of time. If there is nothing inherent to a branching structure that determines temporal orientation and passage, then the external-to-branching factor that introduces these notions will be available to the B-theory and its single, unbranching timeline. Let us return to the simple model of Figure 7.4, and imagine that a temporal orientation is added to it as in Figure 7.5. A

B C

Figure 7.5 A branching model with temporal orientation

Next, suppose that branch C falls away, leaving the following, as in Figure 7.6.

ON MODELLING THE FUTURE A

197

B

Figure 7.6 A whittled down branching model with temporal orientation

This remains a model in which time has a direction, as indicated by the arrow, which is external to the branching structure itself, by hypothesis.²⁵ Figure 7.6 simply describes the B-theory, with relational passage. Might it be countered that temporal orientation requires branching, so that Figure 7.6 is not a possibility?²⁶ No, because there are well-defined accounts of temporal orientation that lack branching. Consider, for example, the model of Robert Weingard, in which a four-dimensional spacetime appropriate to Einstein’s Special Theory of Relativity is combined with a temporally oriented vector field, i.e., a field in which each spacetime point is associated with a set of unit vectors oriented towards one of its light cones, which is thereby determined to be its absolute future (Weingard 1977). Since a vector field can be defined without branching, it follows that temporal orientation is not logically wedded to such a structure. To put the point slightly differently, if a branching structure can be inherently directed then so can a vector. Accordingly, the B-theory can appeal to such a field, or something similar, in order to define a temporal direction for relational passage.²⁷ In short, the rather high ontological and epistemic cost of the fixed branching structure comes at no gain relative to the simpler B-theory, which is experientially appropriate because, recall, we do not experience future indeterminacy. What we experience is that the future is epistemically indeterminate, i.e., that, at any given time, there is more than one description of what will be that is compatible with all that we know at that moment. Any number of combinations of incomplete information, chaos, complexity, and probabilistic laws of nature will suffice for such unpredictability. Ontological indeterminacy, on the other hand, cannot make the future any more unpredictable because, as noted above, it is a given that the future is determinate once present and experienced, so unless concrete future timelines somehow causally diminish our current state of knowledge, they can have no additional impact on ²⁵ It might be thought that a deterministic universe would lack temporal direction since the future entails the past as much as vice versa, but that follows only if determinism implies a lack of direction, which in this context would be question-begging. ²⁶ McCall (1994: 22–4) suggests that branching is required to introduce temporal orientation, as opposed to mere anisotropy, into a model. ²⁷ Note, further, that if Figure 7.6 is not a possibility, then passage is impossible in deterministic worlds.

198

M. JOSHUA MOZERSKY

the predictability of the future.²⁸ On the other hand, abstract possibilities, i.e. distinct descriptions of the future that are nomologically and logically compatible with what we currently know, will do the work of explaining this aspect of our experience of time, for they explain why our current level of information will not entail a single future. There is nothing left for concrete branches to do; they constitute an unneeded and ontologically costly hypotheses.

7.9 Conclusion A model of time ought to include an account of the future. I have compared two kinds of temporal model, one in which the future is ontologically indeterminate, in virtue of containing many concrete timelines, and one in which it is not, in virtue of containing only a single concrete history. The primary strength of the latter is its ability to coherently combine perspectives in time with the perspective of time as a whole without recourse to multiple future selves, semantic indeterminacy, or a fragmented reality. It remains, nonetheless, compatible with three fundamental observations about time: that the future is unpredictable, the past is fixed, and time the dimension of change. It is, further, compatible with substantive temporal passage. Hence, some progress has been made towards understanding the structure of time by eliminating one family of contenders, branching. This does not settle the issue because there are alternative accounts of the future to consider, in particular those in which it is entirely empty, i.e., a realm of non-being (Tooley 2000; Prior 2003). It turns out that considerations similar to the foregoing can be brought to bear against such models but demonstrating this will have to await another occasion.²⁹

²⁸ One might argue that our inability to predict the future is the result of the fact that the laws are indeterministic and the explanation of this is that the future is ontologically indeterminate. I have suggested, however, that the laws can be indeterministic even if there is an actual future, so the suggested explanation is the worse of the two options because it is attached to a model that cannot reconcile the internal and external perspectives while the B-theory can. ²⁹ Thanks to Matt Farr and Katarzyna Jaszczolt for helpful comments and suggestions throughout.

8 Perceiving direction in directionless time Matt Farr

8.1 Introduction A couple of key questions we can ask about the nature of time are: A. Does time flow? B. Does time have a direction? Question A is famously debated by A-theorists and B-theorists about time. Whereas A-theorists take the various flow-like qualities of our experience of time to reflect an underlying property of the world, namely the flow of time itself, B-theorists reject that time really does flow, holding instead that our experience of time is either illusory in this respect or else includes no such appearance of time flow. This chapter focuses on question B, and the less well-studied distinction between the B- and C-theories of time. Whereas A- and B-theorists hold that time is directed, C-theorists reject this. What, then, can C-theorists say about our experience of time? Do we experience time as directed? And if so, how does this fit with a directionless ontology of time? Or can we legitimately deny that we experience time as directed? This chapter will (1) establish how the C-theorist can accommodate a directed experience of time within a directionless ontology of time, and (2) motivate the claim that we do not even experience time as directed. The plan is as follows. Section 8.2 sets out in detail what it means for time to be directionless. Section 8.3 distinguishes between two putative problems about time direction that our experience of time poses for the C-theorist: the ‘hard’ and ‘easy’ problems. And Section 8.4 asks whether the direction of time is a perceptual illusion, cognitive error, or neither.

Matt Farr, Perceiving direction in directionless time. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Matt Farr (2023). DOI: 10.1093/oso/9780192896445.003.0008

200

MATT FARR

8.2 What is directionless time? 8.2.1 Time from A to C As the terminology of a C-theory of time is non-standard,¹ let me introduce it. The more well-known A- and B-theories of time are based on John McTaggart Ellis McTaggart’s A-series and B-series. McTaggart (1908) introduces the ‘A-series’ as the series of events ‘running from past to future’ and the ‘B-series’ as the series of events ‘running from earlier to later’, with the key distinction being that only the former is dynamic: events in the A-series change from being ‘future’, to being ‘present’ and then finally ‘past’; whereas on the B-series, one event’s being ‘earlier than’ another doesn’t change in any equivalent way. While both have a direction built into them by way of their ordering relations, the C-series ‘has no direction of its own, though it has an order’ (p. 462), with McTaggart implying that the central ordering relation of the C-series is something like temporal betweenness, which provides an undirected ordering of events in time, one that is invariant regardless of whether one were to look at it from past to future or from future to past.² Whereas a B-theory takes things in time to form a directed ordering by means of the ‘earlier than’ relation, a C-theory of time³ takes things in time to form an undirected ordering. Imagine some series of events, such as the (rather temporally extended) writing of this chapter, W, the publication of the book, P, and your reading of this sentence, R. The B-series of these events is W, P, R, since W is the earliest and R latest. On the B-theory, the B-series R, P, W (where the reading is earlier than the publication and writing) would represent a different world to our own, since the events are in different directed ordering. But the C-series of the events is exhausted by the fact that P is temporally between W and R; on the C-theory, since the fundamental temporal structure of the world is determined by the C-series of things, there is no such thing as a world just like ours but run backwards; so long as P is temporally between R and W, then we’re in the same situation.

¹ It is introduced in Farr (2012b). ² McTaggart (1927) goes on to develop a very different understanding of the ordering relation of the C-series, holding it to be an ‘asymmetric’ relation. See Farr 2020a for details. This presentation will assume the C-series is based on something like ‘temporal betweenness’, fitting with McTaggart’s (1908) version. ³ Farr (2020a, 2022b, 2012b).

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

201

8.2.2 What is a direction of time? Since we’ve defined the C-theory in terms of its rejection of a direction of time, we can ask what exactly is a direction of time. It is first important to distinguish direction from asymmetry. Undoubtedly the world is full of time-asymmetric processes—processes that have different properties relative to the opposite directions in time. For instance, if we take the outward-radiating ripples on the surface of a pond after being hit by a stone, we can see that from future to past the ripples appear to move inwards, converging on the point of impact rather than spreading outwards. Moreover, the world is full of apparently irreversible processes; those that occur in only one direction in time and not the other, such as the dispersing of gases into the wider environment, the smashing of glasses into scattered shards across the kitchen floor, and so on. This class of irreversible processes is subsumed under the Second Law of Thermodynamics, which holds that the entropy of systems (where entropy corresponds roughly to a measure of the randomness of the order of systems) increases over time and does not decrease. Each process just mentioned, and the Second Law itself, were described in past-to-future terms. But it does not follow that any of these processes, or the Second Law, are directed in time. It is customary for us to describe and represent the world relative to our past-to-future direction, but it is not clear that we would be saying anything false if we were to adopt the opposite convention and describe the world from future to past. The future-to-past description would still describe the same world with which we are familiar, but from the opposite time-directed perspective. Now we’d have a world in which entropy tends to decrease rather than increase in time. The same asymmetries in time would be present, and so the same laws relating these asymmetries would hold, but their time-directed description would be different. As such, the irreversibility of processes described by the Second Law need not be understood in time-directed terms. Rather, the idea of time as directed is something over and above the existence of time asymmetric and irreversible processes in the world. A privileged direction of time would be equivalent to one set of timedirected descriptions of the world (e.g. past-to-future descriptions) being truer than their counterpart (future-to-past descriptions). Whereas some argue that there are good reasons to think that past-to-future descriptions of the world are more true than their converse (e.g. Maudlin 2007; Mellor 2009), others have argued that such a view is either unmotivated (Reichenbach 1956; Farr 2020b) or unintelligible (Price 2002, 2011).

202

MATT FARR

There are a range of different strengths of time direction that have been defended by philosophers, ranging from the metaphysically richest idea of a worldwide tide of ‘nowness’ that flows from past to future, to the most radical antirealist position of the very notion of time directionality being something to eradicate from our description of the world. Existing views can be broken down as follows, in order of decreasing richness. Directed Nowness. There’s an objective ‘now’ that moves from earlier times to later times. This is a position commonly held by A-theorists about time.⁴ But even those who reject the idea of a moving ‘now’ can hold a B-theoretic view of the direction of time. A couple of different versions have been defended. Directed Passage. Time is dynamic and objectively directed from past to future, but without a preferred ‘now’. (e.g. Maudlin 2002, 2007). Directed Block. Time is non-dynamic but objectively directed from past to future. (e.g. Earman 1974; Mellor 1998, 2009). Each of the above positions are realist views about time direction in that they take the direction of time to be a feature of the world, independently of our experience, language, and classificatory schemes. C-theories of time, conversely, are antirealist about time direction. There are at least two different ways for the C-theorist to go (for more details about the range of different possible C-theories, see Farr (2020a)): Direction Conventionalism. Time (and processes in time) are not directed; but the convention of describing processes from past to future is either useful or indispensable. (Farr 2020a, 2022a; Reichenbach 1956). Direction Eliminativism. Time is not directed; nothing is time-directed; all such talk is false and to be eliminated.⁵ Both of these approaches deny that there is anything fundamental or mindindependent about the directionality of time. Where the views differ is that conventionalist C-theories take the use of time-directed descriptions and models as well-motivated by our temporal perspective in the world.

8.2.3 C-theories: Directionless time C-theories of time reject that time has a direction. A useful way to understand this position is to think of the past-to-future and future-to-past directions as ⁴ See, for instance, Zimmerman (2005). ⁵ Price (1996, 2002) appears at times conventionalist and at others eliminativist about the direction of time. See Farr (2020a:10–11) for details.

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

203

providing different ways of describing the same world. Reichenbach (1956: 31–32) put it this way, holding that ‘positive and negative time supply equivalent descriptions, and it would be meaningless to ask which of the two descriptions is true’, his point being that there is no sense in which our conventional past-to-future description of everyday processes is any more true than future-to-past descriptions. So, in the case of the apparent arrow of thermodynamics, Reichenbach notes that ‘it has no meaning to say […] that […] entropy ‘really’ goes up, or that its time direction is ‘really’ positive’ (Reichenbach 1956: 128–129). This is not to disparage or undermine the Second Law, but rather to make clear that what the law describes is a set of time-asymmetric phenomena which only has the appearance of a time direction relative to our conventional past-to-future description. On this view, the C-theory holds that there is nothing about the nature of time that is in conflict between forwards and backwards descriptions of processes; the two sets of descriptions describe the same reality, with no further direction of time in the background. Similarly, the cosmologist and contemporary of Reichenbach, Thomas Gold (1966: 327) held that ‘the description of our universe in the opposite sense of time […] is not describing another universe, or how [our universe] might be but isn’t, but it is describing the very same thing’, adding that such a description may indeed ‘sound very strange’, but this oddness is due to the unfamiliarity of the future-to-past perspective rather than due to getting something about the world wrong. This way of thinking about C-theories fits particularly well with McTaggart’s (1908) presentation of the C-series, since what remains invariant when switching from past-to-future to future-to-past descriptions is the temporal betweenness relations holding between the events described. The B- and C-theories disagree as to whether there is, beyond our convention of describing things from past to future, a time-directionality belonging to processes in the world, or directionality of time itself. Whereas B-theorists take the direction of time to be a mind-independent feature of the world, C-theorists can take time-directed talk to be false in general (eliminativism) or simply a useful convention due to the human perspective within time (conventionalism). Conventionalism is of more interest in the present chapter as it seeks to accommodate both a directionless ontology of time and the apparently time-directed aspects of human time. Conventionalism about time direction holds that there are no time-directed facts, but that it is simply more convenient to describe the world from past to future.⁶ Since there is no underlying fact about the world that time goes from past to future, then it is not ‘false’ to describe things from future to past, but ⁶ See Farr (2022a) for a defence of conventionalism about time direction.

204

MATT FARR

such talk goes against our standard conventions. This is a position hinted at by Reichenbach, and consistent with ‘perspectivalists’ about time direction, such as Price (2007). Reichenbach (1956) holds that although past-to-future and future-to-past descriptions are equivalent in that they represent the same facts, the convention of using past-to-future time to describe processes is not arbitrary; there are good reasons for preferring to describe the world in a way that accords to the apparent direction of ourselves in time. Intriguingly, Reichenbach takes this to be because future-to-past descriptions of everyday processes ‘contradict […] the time direction of psychological experience’ (Reichenbach 1956: 154), implying that we do indeed experience time as directed from past to future.⁷ But if the aim of our program is to establish how C-theorists can accommodate the properties of human time, we must ascertain whether the C-theorist need even concede that human time is, or even appears to be, directed.

8.3 What problems does human time pose for the C-theorist? Now we’ve set out the metaphysics of the direction of time, we can turn to our experience of time and how the two relate. These can be categorized into ‘hard’ and ‘easy’ problems. First, hard problems take the form: how are direction-like experiences or beliefs possible in directionless time? Behind such questions lies a background disagreement between B- and C-theorists: some B-theorists may claim that a directionless world simply cannot give rise to the kinds of temporal experiences that we have; C-theorists conversely reject such claims, instead implicitly holding that there are no such hard problems of time direction. On the other hand, easy problems concern things such as: why do we have the ‘directed’ experiences of motion and change that we do? And what do motion and change illusions tell us about our perception and/or cognition of time direction? Solutions to easy problems would make reference to the specifics of neuroscience, psychophysics, and cognition of change and motion, among other things. Hard problems can be posed solely for the C-theory (whereas easy problems pose specific puzzles that are independent of which theory of time one adopts): they hold that the C-theory is missing something important, that we could not ⁷ Similarly, Price (1996, 2007) holds that the direction of causation (from cause to effect) largely depends upon the temporal asymmetries of human agency, and that nonetheless the direction of causation plays an important role in our understanding of the world.

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

205

in principle have the kinds of temporal experience we do, or moreover that there could not exist the kinds of time-directed processes that there appear to be in the world, without time itself being directed. There’s a lot to unpack in this idea. First, it is assumed that we have time-directed experiences; second, that there exist time-directed processes. A time-directed experience is one that represents time as being directed or things as being directed in time. Consider the B-theorist Tim Maudlin on this issue. Maudlin defends a version of the hard problem by holding that directionless time would not give rise to the kinds of processes and experiences that exist in our world: Nothing happens in [a directionless] world. […T]he state of [such a] world is so unlike the physical state of anything in our universe that to suppose that there are mental states at all is completely unfounded. (Maudlin 2007: 124)

As such, he argues, we need to assume that time is directed, contrary to the C-theory, in order to explain the very existence of experience at all, let alone a specific directed experience of time. It is not clear how the C-theorist can respond to such a worry other than simply to reject the key background assumption and take for granted that the C-theory is indeed consistent with the existence of beings like ourselves. Hard problems such as this set out a dialectical stalemate: to deny a hard problem outright, the C-theorist begs the question against the B-theorist by assuming we can have the experience we do in directionless time; and the B-theorist in defending hard problems begs the question against the C-theorist by denying this. But such problems have an a priori flavour; the C-theorist of course can’t simply rule out the coherence of the idea that directed time is a precondition for conscious experience, but the B-theorist equally cannot rule out the C-theorist’s converse view. This debate again mirrors part of the debate between A- and B-theorists, with Prosser (2012) noting an analogous stalemate exists between realists and antirealists about temporal passage: ‘both [passage realist and antirealist] theories, on their own terms, predict the same experiences[, …] experience does not favour one theory over the other’ (p. 71). Indeed, Williams (1951) refers to A-theorists who hold that block universe doesn’t do justice to temporal experience as ‘time snobs’, holding instead that the block universe leaves nothing out, and that the concept of temporal passage over and above the spread of events in time as being a misleading ‘myth’. Williams’ contention is that those who reject the A-theory do not seek to undermine the phenomenal qualities of temporal experience, but rather hold that no part of our temporal

206

MATT FARR

experience requires that time has an extra property of ‘passage’ to explain it. Similarly, I have argued (Farr 2022b) that C-theorists indeed seek to take the directional aspects of our temporal experience and representations seriously; they simply reject that these require, or would even be well explained by, the posit that time has an intrinsic direction.⁸ To focus the issue on the directed aspects of temporal experience, it’s easy enough to pose the following argument against the C-theory: 1. We have time-directed experiences; things appear to us to go from earlier to later. 2. We could not have such experiences unless time itself is directed, or processes are themselves directed in time (from earlier to later). 3. The C-theory rejects that time itself is directed and that processes are directed in time. 4. Therefore the C-theory is incompatible with our temporal experience, and hence false. The C-theorist has various options here. Primarily, they may reject either (or both) of the first two premises, and conventionalist C-theorists could quibble with the content of premise 3. Premise 2 presents a version of the hard problem; the idea that it is in principle impossible for a directionless time to give rise to the kind of experience of time that we have. As noted, I don’t intend here to offer a defence of the C-theory against hard problems, since the C-theory is predicated on the contention that all of our beliefs and (putative) experiences of the direction of time follow from a directionless ontology of time coupled with the various details of physics, chemistry, biology, psychology, sociolinguistics, etc. In this way, the C-theorist takes the relevant problems about time direction to be the ‘easy’ problems: how exactly it is that we come to have the beliefs, representations, and experiences about the direction of time that we do. In order to ascertain the size of this task, we must first establish in what sense our experience of time is directed; in other words, we’ll consider what the C-theorist can say in response to premise 1. This turns out to be a surprisingly understudied topic, to which we now turn.

8.4 Does human time have a direction? We’ve seen that, for the C-theorist, physical time is not directed. But to what extent is human time directed? There are a range of different positions on this ⁸ Paul (2014) considers the hard problem, offering a series of ways in which the ‘reductionist’ about the arrow of time (i.e. the C-theorist) can respond to the objection that temporal experience is inconsistent with their ontology.

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

207

that one may take, from the view that we literally experience time as though it is directed from past to future, to the view that it is a mistake to think that we can even have a coherent belief that we experience time as directed. In order to spell out the different possibilities here, we can first look to the positions that have been defended by B- and C-theorists with respect to the passage of time, before constructing analogous positions for the C-theorist with respect to the direction of time. First, a terminological note. The issue at the centre of debates about temporal experience is whether our experience has phenomenal qualities that correspond to certain metaphysical views about time. So, we can ask whether we experience things as though time is passing; and in our case, we are interested in whether we experience things as though time is directed. Paul (2010, 2014) refers to such experiences as ‘experiences as of ’ passage or direction; others have framed the issue in terms of whether we have ‘passage phenomenology’, and I’ve (Farr 2020c) called these ‘temporal qualia’. Where I use these terms in the following discussion they are interchangeable. However, one reason I prefer the latter terms (phenomenology or qualia) is that the term ‘experience as of ’ implies that the relevant experience has some particular representational content, such as experience as of passage representing time as passing. As I have argued elsewhere (Farr 2020c), this very issue is central to the debate—there are various reasons why B- and C-theorists may wish to reject that temporal qualia/phenomenology even represents time as passing, and as I argue in this section, there are plausible reasons for the C-theorist to reject that our experience represents time as being directed.

8.4.1 Degrees of temporal experience The question of whether the passage of time is something experienced by us has received plenty of attention in recent years, with a range of positions being defended. Those who reject that time really does pass, namely B- and C-theorists, can hold that we have illusory experience of time as passing, a false belief that we experience time as passing, or simply that there is no aspect of our temporal experience or cognition that aligns with the idea of time as passing. I’ll go through each in turn. Phenomenal passage illusionism. We have a phenomenology of time as passing that is illusory in that it represents something that does not exist. (e.g. Paul 2010; Prosser 2012) This corresponds to the idea that we experience time as though it passes. For example, there might be some unifying ‘passage quale’ that accompanies our

208

MATT FARR

experience of things in time, or instead that our experience of things like change and motion involves a faulty phenomenal representation of time as passing. For the illusionist B- or C-theorist, either of these constitute an illusory projection of time as passing. Paul (2010) argues that those who reject the A-theory may regard the appearance of flow or animation that accompanies a perception of change or motion as a creation of the brain in representing change or motion: ‘a stage of one’s brain creates the illusion of such flow, as the causal effect of prior stages on (this stage of ) one’s brain’ (Paul 2010: 352; my emphasis). However, many non-A-theorists have rejected the idea that time even appears to pass, being eliminativist about passage phenomenology: Phenomenal passage eliminativism. There is no aspect of our experience of things in time (e.g. change, motion, etc.) that corresponds to a phenomenology of time as passing (e.g. Deng 2013a, 2019; Hoerl 2014; Farr 2020c). We can subdivide eliminativism into cognitive and non-cognitive versions, both of which being eliminativist about passage phenomenology (denying that there we have experiences ‘as of ’ passage), but disagreeing as to whether our attitudes towards our temporal experience correspond to a false belief that we have passage phenomenology. Phenomenal passage cognitivism. Though we do not have phenomenology of time as passing, we are subject to a systematic and false set of beliefs that we do have such phenomenology. Cognitivism holds that we are in the grip of a cognitive error when we think that our experience is as though time is passing, and has elsewhere been referred to as ‘cognitive error theory’ (Miller et al. 2018) due to its similarities with the structure of moral error theory.⁹ The idea behind this position is that though we do not have a phenomenology of passage—we do not experience time as though it passes—we nonetheless believe that we do. Miller et al. (2018) set out a series of ways in which we can come to the false belief that we have passage phenomenology. However, this doesn’t exhaust the possible options, nor does it capture the views of those who simply reject talk of passage as ⁹ There is an unfortunate terminological ambiguity here. In the case of time, illusionism has come to refer to those who hold that we have passage phenomenology but that time does not really pass, and cognitivism is a distinct position denying that we do have passage phenomenology. However, cognitivism is closely related to the ‘illusionist’ theory of consciousness of Dennett (1993), Frankish (2016) and others, who hold that we falsely believe that we have certain kinds of phenomenology (like experiencing the ‘redness’ of seeing red things).

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

209

metaphorical in nature or even conceptually incoherent.¹⁰ To accommodate this even more minimal claim about temporal phenomenology, we need an extra option: Phenomenal passage non-cognitivism. We neither have a phenomenology of passage nor a systematic false belief that we do. Instead, our tendency to describe human time as having the appearance or feeling of passing or flowing is metaphorical in nature and does not ascribe such a property to our temporal experience. The idea behind non-cognitivism is that our ways of talking about our experience of time, or things in time, invoke evocative metaphors such as the flow of rivers or passage of objects through space but do not assert propositions about time capable of being true or false. In this way, passage talk is akin to how moral non-cognitivists regard moral talk—as expressing (for example) attitudes or expressions towards objects or behaviours. So, in the case of passage, the non-cognitivist would regard phrases like ‘time flows forever like a river’ as functioning as a kind of aesthetic attitude towards time rather than a statement of fact about some property possessed by time.¹¹ Before moving on to the case of time-direction perception, it’s worth elaborating on the distinction between cognitivism and non-cognitivism. The point at issue is whether the best way to philosophically categorize our talk and experience of time is in a cognitive or non-cognitive manner. Despite its lack of explicit defence in the literature, there are plausible motivations for non-cognitivism. While there is something it is like to see change and motion, which is commonly associated with the idea of passage, some have suggested that these aspects of experience do not give rise either a projection of time as passing or to a coherent belief that time passes. For instance, Hoerl (2014) suggests that because the very concept of time passing is selfcontradictory or otherwise incoherent, we simply cannot have a coherent belief that we experience such a thing. In Farr (2020c), I defend a position termed ‘reductionism about temporal qualia’, suggesting it to be an error-free account of our perception of things like change and motion. Though there is a specific phenomenology associated with perceiving real or illusory motion ¹⁰ For instance, I argue (Farr 2020c) that it is an unhelpful mistake in the philosophy of perception to relate the kinds of qualia associated with real or illusory motion and change perception to the idea of time as passing. ¹¹ One might think that cognitivism vs non-cognitivism presents a straightforward empirical question: do the folk have the belief that they have passage phenomenology? This question has been put to the test—see Latham et al. (2020) and Shardlow et al. (2021) for dissenting views on this issue, and Norton (2021) for an overview of this disagreement.

210

MATT FARR

or change—motion qualia and change qualia—I argue that these are simply what it’s like to perceive or project motion and change in the world, and that there is no reason to categorize such qualia as pertaining to the idea of time as passing (nor things as ‘passing’ through time). As such there is no false belief about temporal passage to be explained.

8.4.2 Do we ‘perceive’ the direction of time? With these positions in place, can we lay out an equivalent set of options in the case of the direction of time? Immediately there are two options here: one might hold that the only sense in which we can have time-direction phenomenology is through having passage phenomenology. This appears to be the view of Le Poidevin: [I]f, by ‘the direction of time’ we mean more than mere asymmetry, then this is due to our sense of time’s passage, and this, it is plausible to suggest, is something we do not perceive, but rather project. (Le Poidevin 2015: 469)

This suggests that there is nothing more to having time-direction phenomenology than having temporal passage phenomenology, and so if one rejects that we have passage phenomenology, this verdict carries over to the case of timedirection phenomenology. This view is attractive insofar as the prime candidates for time-direction phenomenology are also those commonly appealed to as temporal passage phenomenology, namely experiences of change, motion, and succession. Directed change. We typically experience the change of some object (e.g. a change in the brightness of an old lamp) as directed—things change from some earlier state to some later state. Directed motion. We typically experience the motion of some object (e.g. a car driving down the street) in the direction from earlier to later times, such that its spatial motion is in the direction from the earlier location to the later location. Directed succession. We typically experience sequences of events as later events succeeding earlier ones, and not vice versa (e.g. when hearing a descending musical scale). While experiences of change, motion, and succession are typically associated with some flow-like or dynamic qualia that is often related by philosophers to

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

211

the idea of time as passing, there is nonetheless an element of directionality in each experience that is conceptually distinct from the dynamic, passagy part.¹² The alternative option is to hold that there is a distinct sense in which time might appear to us as directed, as implied here by Simon Prosser: We experience time as having a direction. […T]his […] strikes me as easier to explain than most of the other features of experience that have seemed to be in tension with the B-theory. For although passage entails directionality, directionality does not entail passage. (Prosser 2016: 204)

On this latter view, the idea that we have direction-of-time qualia is distinct from, and less controversial than, the idea that we have passage qualia. It is this idea that I’ll explore further in this section: Is there a specific sense in which we have either a phenomenology of time direction or a systematic false belief that we do that the C-theorist is required to make sense of ? As with the case of passage, we can delineate a series of options for the C-theorist with regards to our purported experience of the direction of time. Firstly, illusionism: Phenomenal time-direction illusionism. We have a phenomenology of time as being directed from past to future that is illusory in that it represents something (the direction of time, or the directedness of processes in time) that does not exist. And for phenomenal time-direction eliminativism, the C-theorist has the following options: Phenomenal time-direction cognitivism. Though we do not have a phenomenology of time direction, our experience and cognition of time, change, motion, succession, and other related aspects of experience leads us to the false belief that we do have such phenomenology. Phenomenal time-direction non-cognitivism. Our experience of change, motion, etc. do not give rise to a phenomenology of time as directed nor a coherent belief that we have such phenomenology. Time-direction talk is at best non-cognitive (e.g. metaphorical) in nature. In what follows, my aim is simply to demonstrate that in this range of options, the C-theorist has ample resources to make sense of the various features of ¹² For instance, Deng (2013b) makes the case against Le Poidevin that experiences of succession can be considered veridical on the B-theory, since succession—the sense that one thing appears to another—is not an essentially A-theoretic notion but fits with the time-directed ontology of the Btheory.

212

MATT FARR

human time within the context of an adirectional metaphysics. In principle, each of these is a viable option for the C-theorist. I’ll consider each in turn, starting with illusionism.

8.4.3 Phenomenal time-direction illusionism The phenomenal time-direction illusionist holds that we have some kind of phenomenology of the direction of time, such that we are subject to a phenomenal illusion regarding the direction of time. There are two different ways in which this might be the case: first, that there is a time direction quale that typically accompanies temporal experiences—a kind of pure experience as of time being directed; second, that more mundane kinds of perception, such as motion perception, involve a representation of processes as being directed in time. Is there a ‘direction quale’, namely a sense of something being directed that is independent of motion and change qualia? One approach to this is to look at cases in which awareness of direction is dissociated with awareness of motion. While ordinarily motion illusions involve an unambiguous sensation of directed motion, Seno et al. (2012) outline a purported case of ‘directionless vection’. Vection is a ‘conscious perception of self-motion’ which is ordinarily accompanied with an apparent direction, ‘perceived heading’, and speed, ‘egospeed’ (Palmisano et al. 2015). In the study of Seno et al. subjects were presented with a combination of stimuli that resulted in a reported sense of motion without perceived heading or egospeed. Interestingly, this is suggestive of the idea that one can have an apparent perception of motion without perception of direction, and so, abstracting away from this, we could entertain the idea of direction qualia and motion qualia being distinct things. However, this is not to suggest that there is anything in the way of empirical grounds for supposing that there is unique ‘direction-of-time’ qualia that we are in any sense subject to. In general, the relationship between direction qualia and direction-of-time qualia is unclear, and there is certainly a metaphorical jump from the former to latter when appealing to experiences of motion as putative experiences as of the direction of time. In the case of directionless vection, rather than taking it to show that direction-of-time qualia could be distinct from motion qualia, one could instead read the result as evidence that there can be awareness of motion with an ambiguous spatial direction with respect to positive time. If we give up the idea of a pure quale of time direction, the illusionist can instead hold that things like motion, change or succession qualia (what it is like

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

213

to see real or illusory motion/change/succession) involve a conscious representation of time as directed. What does it mean to experience a motion or change in such a way that it involves the representation of time as directed? One way to think of this is that we invariably take our experiences of such things to be directed from past to future. For instance, when I experience a car moving towards me, it is taken for granted that my experience is of the motion of the car relative to the past-to-future direction of time: the perception of the car as directed in space (i.e. towards me) encodes a direction of time (i.e. the direction in time relative to which the car is moving towards me). With change and succession, the situation is slightly different. When I see a coloured light change from red to green, there is clearly a sense in which I experience it as changing in a directed way from red to green. Once again, this corresponds to direction of change relative to the past-to-future direction and not the future-to-past direction. And with succession, there is once more a clear sense in which we take an experience of one event succeeding another as directed from the earlier event to the later event. So, in each such case, insofar as we appear to experience processes as directed, either with motion, change or succession, the direction of the process itself implies a direction of time. In this way, the illusionist might hold that a sense of time direction is a part of experiences of direction. Illusionism is a plausible option for the C-theorist. There is nothing obviously contradictory in holding time to be undirected, and processes being undirected in time, whilst also holding our experience of processes in the world to involve a representation of them as directed in time. Paul (2014) entertains such a view, firstly making the case that there is a clear sense in which we have experience ‘as of ’ a direction of time: [W]e have phenomenal features of our experiences involving causal impressions and causal direction, and phenomenal features involving temporal anticipation. Putting them together, we can see the beginnings of an account of how our cognitive system could construct our experience so that it presents the world to us as an evolving, causally governed, productive universe. (p. 187)

For instance, Paul cites examples of ‘temporal anticipation’ as putative experiences ‘as of ’ the direction of time, where we have the ‘phenomenological feel of having an experience with an anticipatory or predictive unfolding character’ (p. 186), such as seeing a rock precariously teetering over the edge of a cliff and feeling as though the rock is about to fall. Paul’s point is that there is

214

MATT FARR

a class of experiences related to anticipation, prediction, and causal inference that are directed from past to future. Secondly, Paul suggests that the ‘reductionist’ about time direction (those that hold facts about the direction of time to reduce to facts about the time-asymmetric entropy gradient; i.e. one way in which one could be a C-theorist) could potentially hold that ‘our experience as of temporal direction is merely our visual and other cognitive systems’ response to entropy gradients’ (p. 190). One can relax the details about entropy gradients and simply understand illusionism as holding that there can be an experience ‘as of ’ time direction on the C-theory insofar as a time-asymmetric processes in the world can collectively give rise to a phenomenal projection of processes as being directed in time. This is certainly a plausible enough position since processes in the world are standardly thought of as experienced from earlier to later times.

8.4.4 Phenomenal time-direction cognitivism The C-theorist may wish to reject illusionism. It is true that when experiencing a direction of motion or change there is an implied direction of time—the direction of time relative to which one’s judgements about the direction of motion or change are correct. However, the C-theorist can deny that this amounts to a phenomenological representation of time being directed. There are a couple of phenomenal eliminativist options here that correspond to cognitivism and non-cognitivism respectively. First, the cognitivist can hold that we merely have the background belief that the experience we have is relative to the past-to-future direction and not the future-to-past direction. And second, the non-cognitivist can hold that it is simply more convenient to make use of the past-to-future direction rather than the opposite time direction to describe or represent our experiences, but that this does not correspond to a belief that our experiences are directed from past to future. I’ll go through these options in this and the next subsection respectively. If the C-theorist rejects illusionism, and so rejects that we have a phenomenology of time direction, can they offer an explanation as to how we can come to falsely believe that we do? One place to look for such a view is the ‘enduring self ’ conception of time defended in varying ways by Velleman (2006), Ismael (2006), Prosser (2012), and Callender (2017). The central idea is that the appearance of time as directed and flowing owes in part to our conception of ourselves as temporally extended beings coupled with the nature of memory formation and recall. As Velleman summarizes,

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

215

I exist in my entirety at successive moments in time, thereby moving in my entirety with respect to events. As I move through time, future events draw nearer to me and past events recede. Time truly passes, in the sense that it passes me. (Velleman 2006: 12–13)

Prosser (2012: 107) notes that it is ‘this notion of a single entity passing “through” a change that captures at least a very important element of the experience of temporal passage’. While both of these authors focus on the notion of the extended self as providing an explanation for the belief in the passage of time, it is clear that the asymmetry of memory formation—the fact that we have memories only of earlier events, and subsequently that memories generally accumulate over time—brings also a sense of time directionality that is conceptually distinct from the sense of passage or flow. Reichenbach’s own conventionalist account of time direction draws on a similar-sounding explanation for our conventional preference for describing things from past to future: That [a language of increasing entropy] appears to us as a more natural language, that we are so strongly disposed toward the identification of the direction from interaction to order with positive time, has its basis in the nature of the human organism. (Reichenbach 1956: 154–5) Why is the flow of psychological time identical with the direction of increasing entropy? […] The answer is simple: Man is a part of nature, and [their] memory is a registering instrument subject to the laws of information theory. The increase of information defines the direction of subjective time. (Reichenbach 1956: 269–70)

In both cases here, what we see is a defence of the usefulness of the past-tofuture mode of describing the world that is based on the structure of human memory formation. This account is based on the assumption that the timeasymmetry of memory—why it is that we have memories of the past but not of the future—is due to the time-asymmetry of entropy. In Reichenbach’s (1956) wider programme, the existence of a monotonic entropy gradient in our region of the universe (where entropy steadily increases in one temporal direction but decreases in the other) underpins other modal asymmetries, such as the asymmetry of cause and effect (why it is that causes precede their effects in time and never vice versa) and the

216

MATT FARR

asymmetry of intervention (why it is that interventions have a significant effect on the future of systems but leave their past unchanged). This general programme of deriving the modal asymmetries of causation and intervention from facts about the entropy gradient has been developed by Albert (2000), Kutach (2002, 2013), and Loewer (2007, 2012) by analysing such asymmetries in terms of the existence of a low-entropy constraint at one temporal end of the universe—the so-called ‘past hypothesis’. Following this programme of grounding the time-asymmetry of knowledge (that we have specific kinds of knowledge of local contingent facts about the past but not the same kind of knowledge of the future) in terms of the entropy gradient, others such as Hartle (2005), Ismael (2017), and Callender (2017, ch. 11) have developed accounts of why beings like us take themselves to operate as agents towards the future, making predictions and decisions about the temporal future on the basis of information about the temporal past, creating a sense of themselves as future-oriented decision-makers. Such accounts, if adopted by the C-theorist, naturally fit with cognitivism, as they can be understood as giving rise to the belief that we experience time as directed from past to future since this is the direction in which we function as agents. However, even this claim can be resisted by the C-theorist in favour of non-cognitivism.

8.4.5 Phenomenal time-direction non-cognitivism Can the C-theorist plausibly reject that we even have a belief that we experience time as directed? One way to do so is to hold that it is simply more convenient to represent or describe our experience from past to future than from future to past, and that this convention falls short of a belief that our experience is directed from past to future. This view amounts to a conventionalist account of human time direction, fitting with the conventionalist account of physical time direction outlined in Section 8.2.3. 8.4.5.1 Conventionalism about human time direction? Interestingly, conventionalism about human time direction is perhaps what is meant in the above Reichenbach quote when he notes ‘[t]he increase of information defines the direction of subjective time’.¹³ The non-cognitivist can ultimately define the direction of human or subjective time to be that which ¹³ Reichenbach’s position on the direction of human time is ultimately unknown. It would have been the topic of the final part of The Direction of Time, but he died before it had been written. In her editing of the book, Maria Reichenbach notes this, and includes in the appendix notes from Reichenbach’s public lectures on human time.

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

217

matches the direction of growing memories, etc., and so matching the direction of increasing entropy. It is clearly useful to do so since this matches up with the time-directionality of agency. But the C-theorist is not required to hold that human time is ‘really’ directed from past to future, merely that this is a more useful way to describe or understand it. On such an account, as with the directionality of entropy, it is not strictly false to say that human time ‘goes’ from future to past, but rather that we do not use that particular convention for a variety of reasons. On this reading, when holding it to be ‘inconvenient’ to describe the world from future to past since it would ‘contradict[…] the time direction of psychological experience’ Reichenbach (1956: 154) is not referring to time direction as an element of human experience, nor even human time as intrinsically directed, but rather taking it to be unusual and ultimately pointless to describe humans relative to the future-to-past direction. Building from these lines of reasoning, the C-theorist can take human time to be directed from past to future insofar as it is useful to stipulate that human lives are oriented from past to future. It is useful because this direction makes sense of the directionality of agency, that humans are agents that deliberate about an unknown future based on the knowledge of the past and present. Given this stipulation, it follows that our experience of things like change, motion, and succession are time-directed in the sense that experience itself is best described and understood from past to future and not future to past. 8.4.5.2 Veridicalism about C-time An alternative option for the non-cognitivist C-theorist is veridicalism. The veridicalist C-theorist holds that all temporal experience and beliefs about temporal experience accord to what it would be like for the C-theory to be true. What is meant by this? The C-theorist can regard each of the putative kinds of directed experience—e.g. change, motion and succession—to involve experience of objects’ changing properties across time, and regard them as directed only relative to a choice of time direction (i.e. the past-to-future direction). Consider the case of perceiving the spatial motion of an object, such as the car moving towards me. In such a case, the veridicalist C-theorist holds that what we are perceiving is the change in position over time. From past-to-future, this accords to the car moving towards me, and from future-to-past, this accords to the car moving away from me. For the veridicalist, what is being perceived is the car moving in space from point X to point Y relative to the same time direction in which the perceiver is growing older, accumulating memories, and so on, and equivalently the car moving from Y to X relative to the same direction of time in which one is growing younger. What is perceived here, in other

218

MATT FARR

words, is the relative temporal ordering of the motion of the car and the ageing of the perceiver, where ‘temporal order’ concerns which events are temporally between which other events; in other words the C-series rather than B-series. As such, the veridicalist C-theorist can hold that our temporal experience is exactly how it would be were the C-theory true, and that in order to reject this view one would need a substantial argument to the effect that a C-theoretic world would somehow ‘seem’ different to ours.¹⁴ 8.4.5.3 Time direction as conceptually incoherent? As with the case of passage, a motivation for phenomenal non-cognitivism is the questionable conceptual coherence of a direction of time. In the case of passage, one might object that we can’t literally think of time as appearing to flow or pass, since flow and passage are merely metaphors: it is clear what it means for a river to flow or a moving object to pass by oneself, since these are processes that take place over time. But time itself cannot have the same literal properties since time cannot flow or move relative to itself.¹⁵ At face value, the idea of time as being directed is less conceptually problematic than time passing or flowing, but there is nonetheless reason to think it is not wholly unproblematic. Imagine a car driving down the street, and suppose we consider the car as moving in the direction of the distant end of the street, away from us. The car is ‘directed’ insofar as it moves towards a particular point in space (the far end of the road) relative to ‘positive’ time (the direction towards what we call the ‘future’). If we picture this four-dimensionally, the trajectory of the car is denoted by a vector pointing in the direction of positive time. And if we were to view this trajectory relative to reverse time—from future to past—the car would appear to be moving backwards, away from the far end of the road and towards the observer. It is only relative to a choice of time direction that we can hold the car to be headed in one spatial direction rather than the other. Similarly, one can hold that experiences of change and succession are only directed relative to a choice of positive time. As such, it is unclear what it can mean for time to be directed, since each definable direction ¹⁴ On this point, it is hard not to be reminded of G. E. M. Anscombe’s (1959: 151) story about her conversation with Ludwig Wittgenstein about geocentricism. When Anscombe notes to Wittgenstein that geocentricism was intuitive because it just ‘looked as if the sun went round the earth’, Wittgenstein responds ‘what would it have looked like if it had looked as if the earth turned on its axis?’ Of course, the answer is that it would look exactly the same. Analogously, the C-theorist can defend the claim that everything is exactly as it would appear were their theory true, and that to think otherwise would be a misconstrual of what the C-theory predicts about our experience. ¹⁵ Some, e.g. Skow (2012), have bitten the bullet here and simply postulated existence of a second time relative to which ordinary time passes or flows.

PERCEIVING DIRECTION IN DIRECTIONLESS TIME

219

of time (past-to-future and future-to-past) are ‘positive’ relative to themselves. Can we say that time moves forward over time, or that time is a vector in time? Both are problematic ways of talking since they appear either tautologous, or to make reference to a second-order time variable.¹⁶ As with passage, the noncognitivist could in principle say that because the concept of time direction is incoherent, it follows that it is not coherent to believe that we experience time as directed.

8.5 In sum… To sum up, while C-theorists deny that time is directed, there is a wealth of plausible options available to the C-theorist for making sense of our experience of time. With respect to the ontology of time, I have argued that the C-theory need not be understood as strictly eliminativist about the direction of time, but rather can be understood in conventionalist terms, such that under certain conditions it is overwhelmingly preferable to describe the world from past to future. With respect to human time, we have seen how the C-theorist can hold that: (1) our experience of time constitutes the illusion that time is directed; or (2) our temporal experience coupled with background beliefs give rise to the false belief that time appears to be directed; or (3) it is simply useful to talk as though our experience of things is directed from past to future; or (4) the world appears to us exactly as though the C-theory were true. In each case we see plausible strategies for accommodating the allegedly ‘directed’ aspects of human time within a C-theoretic picture of time.¹⁷

¹⁶ A popular mathematical means of representing time as directed is to understand the direction of time in terms of a temporal orientation—a continuous, non-vanishing timelike vector field on a relativistic spacetime (Earman, 1969, 1970, 1974; Weingard, 1977), but such a view amounts to a conventional stipulation of one of the two definable temporal directions as the ‘positive’ direction (Price, 2011). ¹⁷ Many thanks to Kasia Jaszczolt, Natalja Deng, Casey McCoy, Milena Ivanova, and the participants and audience at the ‘Understanding Human Time’ workshop (Cambridge/online, April 2021) for valuable comments on an earlier draft, and Zach Ottati for engaging discussions of non-cognitivism about passage phenomenology.

9 Temporal transparency and the flow of time Giuliano Torrengo

9.1 Introductory remarks Perceptual experience is, at least partly, a way to ‘be in contact’ with what goes on in the world around us. Philosophers have speculated on the nature of this contact. One way to think of it is in terms of mental representations. Roughly speaking, representationalists (as they are sometimes called) think that to perceive the world is to represent features of the world in the mind. Perceptual experiences, on this picture, are individuated by their representational content, which is a feature of the experience that in turn is individuated by the accuracy conditions of the mental representation involved. At least in its contemporary form, representationalism does not violate the central idea of transparency, that is, the thesis that in perception, primarily at least, we are aware of the worldly entity that we perceive. To visually experience a seagull flying past the church bell is to represent the seagull as flying past the church bell. And to represent the seagull in perception is to be aware of, or be presented with, the seagull passing by, rather than being aware of one’s own experience.¹ Note that it is crucial that the awareness in question be of a perceptual nature. ‘Awareness’ is a loose term since we can be aware also through reflection and overt inferences of things involved in our perceptual activities. In what follows, I will use expressions such as ‘being aware of ’ and ‘being presented with’ (and cognates) to refer to this kind of perceptual awareness achieved through phenomenally conscious states initiated by external stimuli.² ¹ This is at least true of representationalists with ‘externalist’ inclinations (cf. Tye 2002). Others with more ‘internalist’ inclinations (cf. Gow 2016, 2017b) would say that we are aware of the content of our mental state, which is what represents the seagull, rather than being presented with the seagull. However, they would also add that it seems to us as if there were a seagull out there, and that’s enough not to violate transparency. I will be more precise on this in what follows, when introducing the distinction between metaphysical and phenomenal transparency (which is Gow’s). For now I just want to stress that I am not questioning that the representationalist understanding of perceptual content as a feature of a mental state is compatible with transparency. ² I will also often talk of features (of the world/of experience itself ) as what we are aware of (or what we are presented with). This expression is meant to cover various theoretical options such as properties, state of affairs, and also (with respect to the transparent cases at least) objects, events, and processes. Giuliano Torrengo, Temporal transparency and the flow of time. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Giuliano Torrengo (2023). DOI: 10.1093/oso/9780192896445.003.0009

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

221

Representationalism is a popular view, but it is not universally accepted. There are two common ways of departing from it. The first, more radical, way is to deny that perceptual experience involves representing the world around us at all. Naïve realists think that to perceive the world is to be in some relation with it—a relation that is explanatorily primitive, or at least that is not to be characterized as representing. Perceptual experiences, in this picture, are (at least partially) constituted by the external object that we are presented with in having them. Whether transparency is among the motivations for adopting such a view of perception or not, naïve realism seems particularly congenial to it. After all, if perceptual experiences are individuated not by their representational content, but by a relation to external features, then there is no risk of being presented, while perceiving something, with features of the perceptual experience itself (rather than with features of what we are perceiving). The second, sensationalist, way does not need to deny that when we are aware of what we perceive, we are representing entities in the world around us. The sensationalist departs only from representationalism as an all-encompassing claim: perception does not only (and does not always) present us with features of the world around us, it also presents us (at least sometimes) with features that belong to the experience itself.³ Sensationalism (as opposed to naive realism and representationalism) seems to be in direct contrast with transparency. If we are perceptually presented with features of our own experience, then the central idea of transparency has to be weakened, if not abandoned. A great deal of the literature on transparency concerns phenomena such as afterimages, phosphenes (the ‘spot of lights’ that appears when our closed eyes are pressured), and other ‘eerie’ experiences that usually come with a sense of irreality or at least a lack of objectivity of some sort.⁴ A more recent strand focuses on temporal experience, in particular experiences that involve cognitive contact with temporally extended entities, e.g., experience of successions of events and their properties, such as their durations.⁵ In this contribution, after some reflections on temporal transparency in general (Section 9.2), I propose an account of experience of duration and succession that includes a radical form of transparency (Sections 9.3 and 9.4). In the last part (Section 9.5), I will tackle the ³ An exception is Papineau (2021), who denies that we are ever perceptually aware of features of the world around us, although our sensations contingently represent them. ⁴ Cf. Boghossian and Velleman (1989); Kind (2008). For a critical discussion, Phillips (2012b). ⁵ Cf. Phillips (2014a); Soteriou (2013). There is also another debate involving transparency, which is relevant for temporal experience, and it is the debate about presentness (namely the awareness that our perceptions occur in the present). I will touch upon it only marginally in this contribution. See Hoerl (2018).

222

GIULIANO TORRENGO

problem of temporal transparency in the experience of temporal flow itself— a feature of our experience over which there is disagreement on whether it is to be understood as a perceptual aspect or not. I will argue that my temporal modifier theory of the felling of time passing—roughly, the thesis that we are aware of an internal flow in virtue of being presented in a ‘flowy’ manner with successions of events—has certain explanatory advantages over rival views in explaining how the flow of time that we experience in the external succession of events, and that which we experience within us can be unified.

9.2 Perceptual transparency and temporal transparency Although the central idea of transparency is, as I said above, that in perception we are aware of an external world, I think its broader formulation should be in terms of metaphysical priority. Perceptual transparency. Awareness of features that appear to us as externally located is prior to the awareness of features of our own experiences. Transparency can be construed along at least two pairs of distinctions. The first is the distinction between positive transparency and negative transparency (Martin 2002). Positive transparency is the thesis that when we inspect our perceptions we are aware of properties of external objects. Negative transparency is the thesis that we are never aware of the properties of our experiences themselves. Full transparency is the conjunction of the previous two theses, namely the tenet that external objects and their properties are the only things we are aware of in perception.⁶ Note that the negative reading excludes that we are presented with intrinsic properties of our own perceptions, while positive transparency is compatible with the claim that we are sometimes aware of the intrinsic properties of experience, but only while perceiving external objects. Negative transparency thus requires that the priority claim above is trivially true (as in ‘auditory experiences of vibrations within the audible spectrum are prior to auditory experiences of vibration outside the audible ⁶ Although in the literature negative transparency is usually understood to entail positive transparency, I prefer to keep them separate. A position that upholds negative, but not full, transparency is, for instance, one according to which we are never presented with external objects, but also never with features of our own mental states. Such a position is strange, but not unheard of; think of Berkeleyan idealism.

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

223

spectrum’), but positive transparency allows for non-trivial readings of the priority claim, in which awareness of intrinsic properties of experience is in some sense dependent on awareness of mind-independent properties of objects. There are various ways to capture this idea. For instance, we can be aware of intrinsic properties of experience only within our peripheral attention, while the focus remains the external objects;⁷ or perhaps awareness of properties of our experience is only obtained through our awareness of properties of external objects.⁸ The second distinction is between the metaphysical and the phenomenal readings of transparency (Gow 2017a). On the metaphysical reading, transparency is the thesis that we are de facto aware primarily of properties of external objects (regardless of what it seems to us). Thus the first part of the general thesis, about our awareness of features that ‘appear to us as externally located’, has to be read as implicitly stating that those very features also are externally located. On the phenomenal reading, transparency is the thesis that it seems to us that we are primarily aware of properties of external objects (regardless of whether we actually are). Thus the first part of the general thesis is read without further implication. The metaphysical reading is compatible with the claim that people nonetheless can sometimes make the internalizing content mistake (Millikan 1991), that is they attribute features of the world out there to the experience itself. For instance, one can maintain that afterimages are (illusory) perceptual presentations of light phenomena, but given the presence of certain defeaters of their objectivity (they ‘move’ with us, they cannot be inspected by going around them, etc.), we tend to experience them as ‘internal’ (cf. Phillips 2013a). The phenomenal reading is compatible with the claim that people nonetheless can sometimes make the externalizing content mistake (Millikan 1991), that is they attribute intrinsic properties of experience to the world out there. For instance, one can maintain that the phenomenal character of our perceptual experience is internally constituted, but it comes with a presentational phenomenology ‘by which we experience ourselves as creatures existing in a mind-independent world’ (Gow 2017a: 413).

⁷ Cf. Richardson (2014), who argues that this is the case for what she calls ‘structural properties of perception’, such as the boundaries of the visual field: ‘… when we turn attention from the mindindependent objects of perception, to the experience we have of those objects, the objects remain the focus of attention … But in thus attending … we find those phenomenological differences that can, we have argued, be understood as a matter of the form or structure of the experience’ (p. 10). ⁸ Cf. Richardson (2018), commenting on Hoerl (2018)’s ‘eliminativist’ take on tensed properties in perception, suggests this for ‘perspectival’ temporal features such as presentness.

224

GIULIANO TORRENGO

What I call temporal transparency is the application to temporal features— such as duration, temporal order, flow, or passage—of the ideas that I have just discussed. In its general form, thus, it is the following thesis. Temporal transparency. Awareness of temporal features that appear to us as externally located is prior to the awareness of the temporal features of our own experiences.⁹ The events that we perceptually experience have temporal properties, such as having a certain duration and being constituted by possibly qualitatively distinct parts. But experiences are also events that happen in time and have temporal properties. Full temporal transparency in the metaphysical reading tells us that whatever the temporal properties of our own experiences are, they are not the object of our awareness; we can only be presented with temporal properties of the events that we perceive, not of our own perceptions themselves. However, if only the positive reading holds, it may be that we are derivatively aware of the duration and order of our own experiences, by being aware of the duration and order of the events that we perceptually experience. Full temporal transparency in the phenomenal reading rules out the view that it seems to us as if our experiences have temporal properties. The phenomenal reading is prima facie more problematic than the metaphysical one. Experience seems somehow to contain information about both the temporal properties of what it presents to us and of itself. For instance, we seem to know by perceiving a brief event that we had an experience of some duration.¹⁰ Weakening the full claim as to keep only the positive horn may help us out, since positive temporal transparency in the phenomenal reading is compatible with having a secondary or derivative awareness of the temporal properties of our own experiences. In the rest of the chapter, I will discuss two problematic cases for temporal transparency—that of temporal structure (duration and order) and that of the experience of passage—and argue that we may need to treat them differently. ⁹ Notice that Temporal transparency, so defined, it is not merely an instance of Perceptual transparency, since it requires that the externally located features and those of our experiences are of the same kind (namely, temporal), while Perceptual transparency is more liberal on this. ¹⁰ Perhaps even more, we know that our experience must have had the same duration. Cf., for instance, Phillips (2014b: 143): ‘However, we can all agree that the naïve view is committed to the following conditional: if you experience an event as lasting two seconds, your experience of it must itself last two seconds.’

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

225

9.3 Duration and order As already pointed out, transparency is threatened by our experiences of order and succession, which seem to involve not only events around us but also those events that are our own perceptions. More precisely, there is a phenomenological datum concerning perceptual experiences of duration and succession that is at odds with temporal transparency. We can formulate it as the conjunction of the two following pairs of theses. (i.a) We are aware of the durations of the events that we perceive as well as of the durations of the corresponding experiences. (i.b) We are aware of the order of succession of the events that we perceive as well as of the order of succession of the corresponding experiences. (ii.a) Discerning the duration of our own experiences is as easy as discerning the duration of the events that we experience. (ii.b) Discerning the order of our own experiences is as easy as discerning the order of the events that we experience. If (i.a)–(ii.b) reflect a natural way of characterizing perceptions of duration and order, it is obvious why phenomenal temporal transparency is problematic. (i.a)–(i.b) entail, and (ii.a)–(ii.b) presuppose, that we are aware of temporal properties of our own experience, and thus are in conflict with at least the negative horn of full transparency. If we put on the side the possibility of flat-footedly denying (i.a)–(ii.b), we are now faced with three options. First, we could ditch transparency altogether, and embrace the datum expressed by (i.a)–(ii.b) at face value. Second, we could weaken it to positive transparency only (and cook up a story about awareness of the temporal properties of our own experiences being parasitic on awareness of temporal properties of the content of perception). Third, we could stick to full transparency and give a deflationary account of the datum, according to which (i.a)–(i.b) only seem to entail, and (ii.a)–(ii.b) only seem to presuppose, that we are presented with the temporal properties of our own experiences. Roughly, I will follow and defend the third strategy. The idea is that when we perceive an external event, we thereby have the tendency to describe our own experience as possessing the corresponding temporal profile, that is as having the same duration and presenting the same order. But before going into some more detail, let me say something about the two other strategies, in order to show their weakness and indirectly support my claim that the third strategy is the best. The first option, abandoning transparency, goes hand in hand with the idea that we represent the time of our own mental states through

226

GIULIANO TORRENGO

time markers. According to the time marker view, a perceptual content that has a certain temporal profile, for instance that of a succession of two short events e1 and then e2 , does not need to be neurally realized in the brain by a process that has the same temporal profile, that is the realization of a perception of e1 followed by the realization of a perception of e2 . This view entails, roughly speaking, that perceptual contents come with temporal information about their occurrence that is not encoded through their own duration (or temporal order).¹¹ The thesis is compatible with the claim that experience presents us, independently, with the temporal properties of the events that we perceive; but it is not compatible with the idea that our awareness of the timing of our mental life is somehow derivative on perception of temporal properties. The time marker model has been defended by some psychologists.¹² However, the rival brain time model seems to have more empirical corroboration.¹³ According to the latter, which is sometimes also called the TOR (time as its own representation) model, there is a correspondence between the temporal profile of the content of our perception and that of the neural realizers in which the content is processed. Therefore, if the content is that e1 follows e2 in a certain amount of time n, such a content is realized by a perception of e1 followed by a perception of e2 , and this has overall duration n. Although the brain time model is compatible with having some form of derivative awareness of the temporal properties of our own experiences, I will argue that it sits better in a framework where full phenomenal transparency is adopted. If I am right, and it is also true that the brain time model is to be preferred to the temporal markers model, it follows that we should at least try to find an alternative to the first strategy. I will argue for the association between the brain time model and temporal transparency in the next section, in the context of discussing certain temporal illusions; in the rest of this section I provide independent objections against the second, weakening strategy: retreating from full to positive transparency is not a good move. In order to see why a restriction to positive transparency is problematic, let us take a step back and ask about the relationship between the temporal profile of the content C of a perception E, and the temporal profile of E itself. Unless those temporal profiles are independent from each other, there seem to be two possible options, which correspond to the theses that in the literature are ¹¹ Cf. Johnston and Nishida (2001). Strictly speaking, the time marker view (and the rival brain time view, see below in the main text) are theses about the relation between the neural realizers of our experiences and the perceived contents, but they have consequences for the issue of perceptual transparency. ¹² Cf. Kiverstein and Arstila (2013). ¹³ Cf. Arstila (2015a).

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

227

sometimes labelled the Inheritance thesis and the Projection thesis and which I give in their metaphysical readings (hence the subscript) below. InheritanceM An experience E has a certain temporal profile because its content has a certain temporal profile.¹⁴ ProjectionM The content of an experience E has a certain temporal profile because E has a certain temporal profile.¹⁵ The formulations above reflect the metaphysical readings of the explanatory relations between content and experience (or the content and vehicle of experience, as it is sometimes put), because they are silent on the relationship between the awareness of the temporal properties of the content and the awareness of the temporal properties of the experience itself. Therefore, both the claim that the relation of priority goes from the content to the vehicle (the Inheritance thesis), and the opposite one (the Projection thesis) are compatible with positive phenomenal transparency. One can maintain that it seems to us that we are directly aware of the temporal features of the events that we perceive and conjoin that claim either with (i) the thesis that such features also determine the temporal profile of our own experiences, or with (ii) the thesis that the temporal features of which it seems to us we are directly aware are determined by the temporal features of our own experience. Neither combination is particularly appealing though. If we employ option (i), that is the coupling of phenomenal transparency and the inheritance thesis in its metaphysical construal, we leave the datum (i.a)–(ii.b) unaccounted for. Assume that we have a complex experience E of a three note arpeggio, which is a succession of three shorter experiences E1 -E2 -E3 of the three individual notes do, mi, so. From InheritanceM it follows that E’s temporal profile can be explained in terms of the temporal profile of the arpeggio as it is presented to us. However, from InheritanceM it does not follow that our awareness of E’s temporal profile can be so explained, indeed it does not follow that we are aware of E’s temporal profile in the first place—the thesis can be true regardless of whether we are aware of it. The hypothesis that the temporal profile of our perception is determined by that of its content is explanatorily idle with respect to accounting for why we report being aware of the temporal profile of our own experiences. Appealing to such a hypothesis is like trying to explain our awareness of an emergent feature of a percept, let us say a shape that we see in a constellation of dots, simply by saying that the shape is determined by the ¹⁴ Cf. Soteriou (2013); Phillips (2014a). ¹⁵ Cf. Lee (2014).

228

GIULIANO TORRENGO

constellation of dots. That just does not seem to be enough, if only because other shapes are determined by the same constellation, if we do not take into consideration also our visual system. If we employ option (ii), that is the coupling of phenomenal transparency and the projection thesis in its metaphysical construal, we are claiming that something analogous to an externalizing mistake happens. Although it seems to us that the short arpeggio that we are listening to has a certain duration, say, it has that apparent duration only in virtue of the fact that it is presented to us in an experience that possesses such a duration. Would this be a violation of metaphysical transparency? It depends on how we construe the dependency relation in ProjectionM . If the temporal profile of the content is reduced to that of the experience, then it is a violation after all. Be that as it may, the point is, similarly to what we have seen for the previous option, one of explanatory deficiency. ProjectionM entails that the temporal profile of an experience E metaphysically explains the corresponding temporal profile of its content C, but it is silent with respect to what explains our awareness of E to begin with. We are presented with the arpeggio as having a certain duration and showing a certain order because our experience of it has this very same temporal profile: E has a certain duration and it is constituted by E1 followed by E2 , followed by E3 . But ProjectionM can be true also if we are not aware of the temporal profile of E. And adding that we are aware of the temporal profile of E exposes us to the risk of abandoning positive phenomenal transparency, since it is difficult to see how such an awareness could be derivative on the awareness of the temporal profile of the content, given ProjectionM . One may think that rather than looking at the metaphysical formulations of the two theses, we should see whether there is an available phenomenal (hence the subscript ‘P’) formulation. One way of doing this would be as follows. InheritanceP . We are aware of what appears to be the temporal profile of an experience E in virtue of being aware of what appears to be the temporal profile of its content C. ProjectionP . We are aware of what appears to be the temporal profile of the content C of an experience E in virtue of being aware of what appears to be the temporal profile of E. Inheritance in its phenomenal formulation is tailor-made for preserving positive phenomenal transparency of duration and succession. But how independently plausible is it? I will try to answer this question not by appeal to intuitions, but by investigating whether the thesis in question is explanatory felicitous vis à vis explananda such as (i.a)–(ii.b). The question is muddled

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

229

by the fact that principles in the ballpark of inheritance and projection are discussed in the literature about the temporal structure of experience. In particular, the debate concerns whether experience comes in ‘units’ that are temporally extended, as according to extensionalism, or it comes in virtually instantaneous atoms with a temporally extended content, as according to retensionalism, or, rather, it is structured as a series of instantaneous mental events with an instantaneous content, as in the snapshot view.¹⁶ However, the problem of accounting for the datum (i.a)–(ii.b) is not trivially solved by solving the problem of temporal structure of perception, and vice versa. When we ask whether reports such as (i.a)–(ii.b) have to be taken at face value, we are asking whether perceptual awareness comes with an awareness of how long the experience lasted and of the order of its qualitative parts. The idea of the three different strategies is that of three possible answers: (i) it does and we are aware of it by introspective the temporal profile of experience itself; (ii) it does and we are aware of it indirectly, in virtue of being aware of the temporal profile of what we are experiencing; or (iii) we are not, although we are led to describe our experience in those terms (more on the third one below). What we are not asking is whether by introspection we can know the temporal structure of our experiences. Think of what a snapshot theorist could say to account (i.a)–(ii.b). They can grant that there is a sense in which our experience of a small arpeggio lasts a second and a half (say), and we are somehow aware of this. And they can provide an explanation of it in terms of their theory, according to which the auditory experience of a short arpeggio is a short sequence of atomic experiences. Now, perhaps in the context of discussing the temporal structure of experience InheritanceP is plausible (as the extensionalist thinks), but it does not follow that it also plausible as an explanans for data such as (i.a)–(ii.b). It remains thus at best unclear that we are justify in restricting transparency on the basis of InheritanceP . ¹⁶ In particular, the Principle of Presentational Concurrence (PPC), according to which ‘[t]he time interval occupied by a content which is before the mind is the very same time interval which is occupied by the act of presenting that very content before the mind’ (Miller 1984: 107). PPC is silent with respect to the order of dependence (if any) between the two time intervals ‘before the mind’. That is, it is compatible with either InheritanceP or ProjectionP , as it is with the negation of both (although, arguably, it is not compatible with the truth of both). However, PPC is not silent with respect to the temporal profile of experiences. Indeed, it is a principle that extensionalists usually defend, since it entails that if we are aware of temporally extended events, as it seems we are, then our experiences of them are temporally extended too. In contrast, the retentionalists reject it in favour of the Principle of Simultaneous Awareness (PAS), according to which ‘If one is aware of a succession or duration, one is necessarily aware of it at some one moment’ (Phillips 2014a: 140). PSA is in tension with extensionalism, since it suggests that experiences of successions are themselves virtually instantaneous, and it is not compatible with InheritanceP , if we read the latter in a theoretically loaded way, that is as entailing that the temporal profile of E is the same as that of C. On extensionalism, retensionalism, and snapshot view, see Dainton (2000). For a recent defence of the snapshot view, see Arstila (2018).

230

GIULIANO TORRENGO

What about ProjectionP ? On the face of it, it is incompatible with positive transparency. If being aware of the temporal profile of our own experience is what grounds the awareness of the temporal profile of their content, then the latter cannot be prior to the former. However, whether the two theses are incompatible depends on whether the awareness of the temporal profile of our own perceptions is gained through introspection of our perceptions. If the awareness that our perception of the arpeggio E has a certain duration and it is constituted by the succession of E1 , E2 , and E3 is not to be understood as introspective knowledge of the temporal structure of our own experiences, but rather derivative on having such experiences, then maybe there is a sense in which we can use ProjectionP to explain the datum, thus abandoning negative transparency, but not positive transparency. The problem with this manoeuvre is that a succession of experiences is not, in and of itself, an experience of succession.¹⁷ Less elliptically, it is difficult to see how we can be aware of the temporal profile of our own experiences simply by having them. Certain extensionalists appeal to a holistic conception of experience, according to which the temporally extended experiences are more fundamental than their successive parts. And they do so in the context of vindicating some form of transparency.¹⁸ One could think that to have a temporally extended experience E in which a short arpeggio is presented is to be aware of the temporal profile of E itself. The projection of the temporal profile of E onto its content does not require that the temporal profile of E is presented to us. Now, on the one hand, I fail to see how the fact that the extended experience is metaphysically more fundamental than its parts helps us do without the need to be presented with its extension, and for reasons analogous to the ones we have seen with respect to the metaphysical readings of the two thesis: we cannot pull a phenomenological rabbit out of a metaphysical hat. Whatever exactly the ‘projection’ amounts to here, it seems to me that it entails at least that we are aware of both temporal profiles. But maybe I am wrong, and I have read uncharitably the extensionalist manoeuvre. However, if the idea is to appeal to the brain time model (that is, TOR) without assuming that we are aware (directly or indirectly) of the temporal profile of our own perceptions, then ProjectionP starts looking very similar to the idea of the deflationary strategy I defend. Our reports about being aware of the temporal profile of our ¹⁷ As for the famous dictum by James (1890, Vol. I: 629). Many examples of analogous points made in the literature are given in Hoerl (2013b: 374). ¹⁸ Cf. Phillips (2011, 2014a), but also Hoerl (2009). Other accounts explicitly reject positive phenomenal transparency here, for instance by appealing to an ‘inward’ higher order experience in which the succession of the first order experiences is represented. See Sattig (2019). He is providing an account of the experience of time passing, but the point can be adapted for the case of duration and succession.

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

231

own experiences should not be taken at face value, as reports on a piece of phenomenology. They are rather the best way to describe situations in which extended contents seem to be presented to us. To put it in terms of an inferential reasoning, the idea is that from the fact that what is presented to us has temporal extension we infer that our experiences also unfold in the same stretch of time. Note two things here. First, it looks like we are using a heuristic based on InheritanceM , rather than ProjectionM. This is not surprising, since if we appeal to the brain time model, but without assuming that we are aware of the temporal profile of our own experiences, their content is the only element that comes with an experienced temporal profile. Second, the appeal to extensionalism and holism is explanatorily idle; the heuristic does the whole job. If we are not aware of the temporal profile or our own experiences, then the explanation works perfectly even if our experiences are virtually instantaneous, as the retentionalist and the snapshot theorists maintain. Let us then have a closer look at this proposal.

9.4 Illusions Many cases of illusory experiences involving duration and order are discussed in the literature. Some of them are cases of temporal illusions, in the strict sense that the property that is illusorily experienced is a temporal property. For instance, in the so-called oddball effect, it is the duration of the stimulus that is misperceived (or at least misreported).¹⁹ Although we are presented with a succession of stimuli each of the same temporal length, in certain circumstances we report one of them as lasting longer than the others. Other cases involve an illusory experience of the distributional qualitative profile of an event, and thus strictly speaking are not temporal illusions (although they are sometimes so labelled). Rather than being inaccurately presented with the temporal features of the events we perceive, we are inaccurately presented with how an event qualitatively develops through time. An example is apparent motion (the phi-motion phenomenon). While the stimulus is constituted by a dot flashing for a moment on the left, followed by a ‘blank’ short period, followed by a dot flashing for a moment on the right, it is reported as being experienced as a continuous motion of the dot from left to right. (Other examples are the so-called ‘cutaneous rabbit’, and in general all posdiction effects.²⁰) ¹⁹ See Tse et al. (2004). ²⁰ In the cutaneous rabbit illusion, a sequence of groups of taps delivered at the wrist and in other locations of the arm ‘seem to the subjects to travel in regular sequence over equidistant points up the

232

GIULIANO TORRENGO

The first case of illusion is not particularly problematic for temporal transparency. To see why, consider the relation between the phenomenon of perceptual inaccuracy, which in principle seems to involve any kind of perceptual experience, and the perceptual transparency thesis. Perhaps illusions are problematic for the idea that perception presents us features of the world around us.²¹ But it seems too quick to consider cases of inaccurate perception of a feature F as conclusive evidence for the claim that F is not transparent to us, namely that in order to explain the inaccuracy we have to appeal to awareness of a feature of the (inaccurate) experience itself. The second case, postdiction effects, is more interesting for the issue of transparency, because they constitute a challenge to the brain time thesis, which is crucial to the understanding of the deflationary strategy, as I mentioned above. Consider again the case of illusory motion. We have a succession of three stimuli: (S1 ) A dot flashing for a moment on the left (S2 ) A ‘blank’ short period (S3 ) A dot flashing for a moment on the right The report is of an experience of a continuous motion of the dot from left to right. What the report suggests is that the experience of the third stimulus S3 (when it comes after the previous two) influences the experience of the previous two. We do not experience the succession of S1 and S2 as a stationary dot followed by a ‘blank’ experience, but rather as parts of a continuous movement of the dot from the left towards the right. There are two main interpretative frameworks of this situation: Orwellian rewriting and Stalinesque delay.²² According to the Orwellian rewriting, there is never a moment in which we experience movement. The experience of the succession of S1 , S2 , S3 can be seen as having three parts, E1 , E2 , E3 respectively, each with a veridical content. That is, the content of E1 is a stationary dot on the left, the content of E2 is a ‘blank’ moment, and the content of E3 is a stationary dot on the right. However, right after E3 an inaccurate memory of what happened imposes on us, and we thus report our experience having been that of seeing a continuous arm—as if a little animal were hopping along the arm’ (Dennett and Kinsbourne 1992: 186). Roughly, cases of postdiction are situations in which there is a mismatch between the how the stimulus develops through tine and how it is reported to have developed. Cf. Gepshtein and Kubovy (2007) for a general discussion. ²¹ Papineau (2021) seems to defend an analogous point. ²² See Dennett and Kinsbourne (1992), who criticize both models in favour of the Multiple Drafts model (see next footnote). Todd (2009) argues, against them, that the distinction is significant. See Grush (2007) on the Orwellian strategy, and Dainton (2008) on the Stalinesque one.

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

233

movement of the dot from left to right. (The brain in this account is like Big Brother in Orwell’s 1984, and has the capacity to delete the past and write a new version.) The Orwellian rewriting scenario is not compatible with the brain time model. Remember, the brain time model entails that the very occurrence of an experience provides us with the information about its occurring at a certain time, relative to other experiences. But if Orwellian rewriting happens, an experience occurring after E1 provides us information about the temporal location of E1 . This picture requires that temporal markers enter the scene.²³ I will then put it aside. Appealing to Stalinesque delay is more promising. The idea is that the elaboration of S1 takes time, and the time at which it becomes conscious has been influenced by processes initiated by stimulus S3 . So the content of E1 is that of a spot that begins to move towards the right. According to the model, therefore, we do experience movement (although not veridically), and there is no misremembering. (The brain in this account is like Stalin’s secret police that creates bogus evidence in trials.) If the Stalinesque delay model is understood linearly, it predicts a delay in the elaboration of the stimulus that seems to be in contrast with the hypothesis that there is always the minimum delay necessary for a stimulus to be elaborated consciously. This is implausible for evolutionary reasons. However, Arstila (2015b) has elaborated a non-linear version of the theory, which is supported by empirical evidence and does not require an extra delay. This concludes my discussion of postdiction illusions. I can now explain how I intend to use the brain time model to motivate the third strategy in the case of experience of duration and successions. According to the brain time thesis, there is no temporal information about the temporal profile of our own experiences other than what we get from the fact that we experience certain events in certain order and with certain durations. No phenomenal awareness of the temporal location of our own mental episodes is available in introspection. However, since we retrospectively can easily infer that of which we were aware during the experiences in question (otherwise how could have we been presented with anything?), it is not surprising that we accept (i.a)–(i.b) as a good way to characterize our experience

²³ This is true also of the more complex Multiple Drafts model. I agree with Artsila (2015b), who argues that the Multiple Drafts model is Orwellian in the crucial sense that according to it, too, there is never a moment in which we experience movement. The difference between the Multiple Drafts model and the other two models is that the first rejects the ‘Cartesian’ assumption of a point of entrance of the experiences to consciousness.

234

GIULIANO TORRENGO

of temporal properties.²⁴ Consequently, it is not surprising that we show a tendency to describe the temporal properties of our own experiences as identical with those of the events that we have been presented with. Something similar happens with (ii.a)–(ii.b). Assuming that by being presented with the temporal properties of an event we are in a position to discern their duration and order of parts, it follows from the above consideration that we take ourselves to be in the same position with respect to the temporal properties of the experiences in which we are presented with them. The heuristics that we use to evaluate the length of our own experiences, according to the hypothesis that I am defending here, bears analogy with the inheritance principle for the case of duration and succession. Notice that the heuristic is specific to the temporal case. There is no reason to think that we should use it for the case of colours and shapes (assuming phenomenal transparency holds for them), and other non-temporal properties. The reason is that simply assuming that one must ‘be there’ to experience a colour or a shape does not entail further similarity between the experience and what is experienced. Contrariwise, in the case of temporal properties, it is very plausible to assume that the default position is that our experience and the event that we experience ‘live in the same temporal dimension’, as it were. Finally, notice that one can interpret this strategy as the claim that (i.a)–(ii.b) are false, or at least that they are false if we understand ‘being aware’ and ‘discern’ in sensory terms. This concludes my account of transparency in the case of experiencing duration and order of events. Although I do not have knock-down arguments against all alternative proposals, my conclusion is that we should embrace the full thesis of phenomenal transparency and ‘soften’ the transparencythreatening implications of the datum (i.a)–(ii.b). I now move to the case of the experience of the flow.

9.5 The experience of time passing Recall the three options we saw with respect to the clash between the fact that we report being aware of properties of our own experiences such as duration and order, and the idea of transparency. First, we can abandon both conjuncts ²⁴ See also Tye (2003: 97): ‘Continuity, change, and succession are experienced as features of the items experienced, not as features of experience.’ Notice that here I am talking about the way we characterize our phenomenology, and not about the fact that we may evaluate the duration of definitely longer portions of our stream of consciousness through memory. The well-studied kind of inaccuracy of those situations is irrelevant here.

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

235

of full phenomenal transparency. Second, we can try to make sense of the positive conjunct and reject the negative one. Third, we can embrace full transparency and ‘revise’ the problematic datum. In this last section, we face another datum that is in tension with the idea of temporal transparency. We can formulate it as follows. (iii) We are aware of the passage of time with respect to the world around us, and with respect to our own mental life. An important qualification: (iii) does not entail that we experience two flows (although it does not rule out it either). What (iii) says, taken as a datum, is that people tend to describe the flow of time both as involving the event that they perceive and the internal state that they possess.²⁵ The experience of flow is, in a sense, undetermined with respect to the outward and inward aspects of our conscious experience. However, as long as our own ‘mental life’ also includes perceptual awareness, (iii) clashes with the thesis that perception is always outward oriented, and does not reveal in introspection features of itself as a vehicle. Notice that there are two elements in tension with transparency here. The first is that (iii) suggests that we are presented not only with time flowing in the external world, but also with an ‘inner flow’, which involves not only thoughts, desires, and cognitive states more generally, but—crucially for the failure of full transparency—also our own perceptions. The inner flow is part of our awareness also when we are focused on perceptual activities, and not merely when we are lost in the stream of our thoughts. Although I do not take the phenomenological datum to be that there are two flows, any theory of this aspect of temporal experience has to specify if this aspect of experience entails a failure of transparency or not, since an all but obvious interpretation is to take it as an awareness of the inner succession of mental states, including perceptual ones.²⁶ The second is that even if the inner flow has nothing to do with perception (and thus it does not impinge on transparency), our awareness of the external flow may require that we are aware of features of our own experience. The second worry is particularly pressing if we do not ²⁵ Jaszczolt (2020) gives an account of the relation between the two flows at the level of belief, rather than phenomenal character. See also next footnote. ²⁶ In this volume, Jaszczolt [p. 17, where (3) is] analyses the awareness of our internal flow in terms of a subjective overt qualifier (SOQ), and our awareness of the external flow in terms of an objective covert qualifier (OCQ). She then identifies the belief in an internal flow as the application of the overt subjective awareness to the objective covert awareness of external time. As I read it, her proposal relies on the idea that the awareness of an internal flow is parasitic on that of an external one, and it is thus compatible with positive epistemic transparency (as mine is too, although for a different reason, see below).

236

GIULIANO TORRENGO

think that there is an objective flow of which we can be aware, but I will try to set metaphysical considerations to one side as much as possible. What is important here is that the second worry is part and parcel of my phenomenal modifier account of the experience of passage, as shall become clear in a moment. As mentioned, I will defend a form of the second strategy for the case of the passage of time (abandoning negative, but retaining positive phenomenal transparency). Before proceeding, let me say something about the third strategy. The idea of embracing full transparency and reinterpreting (iii) accordingly is similar to what elsewhere I have called the deflationist take on the experience of passage, viz. the idea that our awareness of movement and change is mistaken for an experience of flow, the mistake being a cognitive, rather than perceptual one.²⁷ I say that it is ‘similar to’ rather than the same because deflationism (as intended here) is silent with respect to the inner flow: the experience that is ‘deflated’ is the passage of time out there in the world around us. However, it is possible to adopt a more general version of deflationism, according to which we are aware neither (a) of a flow in which the perceived events unfold, nor (b) of a flow in which our own experiences unfold. Given that I have argued against (a) elsewhere,²⁸ I will not consider this expanded deflationism here. However, one may adapt deflationism rather than expand it. This means rejecting (a) and keeping (b), that is giving a nondeflationist account of the perceptual experience of passage, at the same time treating our report about an inner flow as spurious—at least when the inner flow is that of our perceptions, rather than thoughts. I will say something about this midway deflationist strategy later on. Suffice for the time being to notice that accounting for being aware of a flow in which the perceived events unfold in a way that is both non-deflationist and compatible with full transparency is not trivial (remember our second worry: it may be that being aware of time passing by outside us requires that we are aware of features of our own experience). Let us see then whether we can implement the second strategy (trying to save positive transparency), or whether we are forced to retreat to the first one (abandoning transparency altogether), which, as I pointed out above, seems to be a quite natural framework to interpret the idea of an experience of an inner flow of experiences. The second strategy requires that we are presented with a dynamic aspect of our own perception through being presented with an ²⁷ Hoerl (2014); Huggett (2014). ²⁸ See my Torrengo (2017a).

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

237

external flow of events; in addition we may still be aware in some more direct way of an inner flow with respect to non-perceptual experiences. At first sight at least, the view seems in line or compatible with my theory of the phenomenal modifier. Roughly, the idea is that we could be presented with the internal flow of our perceptions in virtue of being presented in a ‘flowy’ manner with successions of events. If so, our experience of an inner flow is compatible with positive transparency, although not with negative transparency. Unfortunately, there is an immediate problem with this project. The phenomenal modifier theory (PMT) of the experience of passage is a form of sensationalism, according to which (in a nutshell) the dynamic character of our phenomenology is an intrinsic feature of experience, a modification of its content that is phenomenally apparent. Therefore, the flow of time is not something that we are presented with in perception, as we are presented with colours, shape, movements, and changes. However, we are aware of such a feature of our perceptions by being presented in a ‘flowy’ manner with the content of our perceptions and our mental life in general. Crucially, in my account, not only perceptions, but any mental state that presents us with a content (imagining, remembering, desiring and the like), in such a flowy manner, invites the thought that time passes. Insofar as our experience is outward directed, the thought is that the events around us flow in time, and insofar as our experience is inward directed, the thought is that our own experience flow in the same temporal flux. It is important to realize that PMT is not detrimental to the idea that perceptual experience in general presents us with a variety of worldly features,²⁹ and thus it is not detrimental to the thesis that transparency in most of the cases holds. Indeed, as long as there are no other reasons to doubt that transparency is in good standing, the fact per se that the experience of the flow of time requires a local failure of transparency is not a reason to abandon transparency in other cases. After all, the experience of the passage of time is a sui generis experience, and it should not be too shocking to discover that it requires to be treated as an exception-like case. Even more importantly, with respect to the external flow, such a ‘local’ failure involves only metaphysical transparency. From the point of view of what seems to us to be the case, the perceptual experience of the passage of time is an experience that comes with

²⁹ Besides, such a form of sensationalism with respect to the experience of passage is neutral with respect to whether other experiences are explained in terms of mental representations or in relational terms (see Section 9.1).

238

GIULIANO TORRENGO

a presentational aspect as perception in general (Gow 2017a). Perceiving is perceiving a world in flux.³⁰ The situation is quite different to the previous case of experience of duration and order. There we needed to appeal to time markers or temporal modes of presentation in order to make sense of the idea that we are aware of the temporal profile of our own perceptions, without construing it as derivative of the awareness of the temporal profile of their contents. Applying the first strategy in the case of perception of duration and order would have meant to admit that the failure of transparency is indeed pervasive. Dialectically, on the one hand, for certain temporal features, such as duration and order, which potentially threat transparency extensively, there is a viable ‘deflationist’ alternative; and on the other hand, if we drop full transparency for the case of the experience of time passing, the failure is restricted to an exception-like feature of experience anyway. But do we have to drop positive transparency too for the experience of time passing, if we explain the latter in terms of the PMT? Here is a rationale in favour of a compatibility between the phenomenal modifier theory and full phenomenal transparency. If PMT is correct, our awareness of the modification of the contents of experience is the basis of the belief that time passes. The best way to understand such awareness, given its role as basis of the belief in the passage of time, is in terms of how the events that we perceive seem to us. But then phenomenal positive transparency is not violated, although metaphysical positive transparency is, as we noted above. We are aware of an intrinsic feature of our own experience, but we systematically make the externalizing mistake with respect to it. Therefore, it seems to us that time passes; it does not seem to us that our own perceptions have a certain intrinsic feature that ‘tricks’ us into thinking so. Unfortunately the rationale leaves something out. If (iii) above is a good prima facie characterization of our phenomenology, we are aware not only of the events around us happening in time, but also of our own mental episodes, including perceptions, happening in time. Compare the difference between the experience of time passing and experiences that are usually taken to be fully transparent, such as visual experience of shape and colour. If I perceive a tree in front of me, I don’t gain more insight about my own mental states by introspecting: it is still the tree that experience presents me with.³¹ But in the case of the passage of time, the situation seems to be slightly different.

³⁰ Cf. O’Shaughnessy (2000: 49 and ff ). ³¹ Cf. the ‘original’ discussion of transparency, presented as a datum rather than as a thesis, by Moore (1903).

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

239

In line with (iii), by shifting my attention from the perceived world to my inner mental life it seems to me that the flow of time is still somehow presented to me. Or it doesn’t? Perhaps. I do not intend to dismiss entirely the rationale above by appealing to a possibly dubious phenomenological datum. I reckon that the rationale is good within its own boundary, but that is not enough. To see why, think first that when we have experiences that are not perceptual, such as memories or imaginings, we do not commit the externalizing mistake—unless for some ‘unusual’ circumstances they come with a presentational phenomenology. Or, more precisely, we do not think that the events that are presented to us in memory or imaginations are happening at the time in which we are presented with them. We are, however, still aware of the passage of time, or at least we are aware of our inner flow.³² Second, given that perception is virtually always part of our mental life, it is doubtful that it should be exceptional in this respect. It is exceptional in the fact that it (as opposed to imagination, desiring, etc.) is characterized by a presentational phenomenology, as already noted, and this is part of the explanation of why we make the externalizing mistake with respect to the passage of time. But it is not an exception with respect to presenting to us also with an internal flow. If so, the PMT requires that metaphysical negative transparency fails, since we are aware of an intrinsic feature of our perception (the flowy manner in which we experience events around us). Phenomenal negative transparency holds ‘locally’, since it seems to us that time flows in the world, but fails with respect to our inner flow, because it seems to us that our mental life (including perceptions) is flowy. I do not want to put too much weight here on introspective data. I take it that it is not obvious that the best reading of (iii) entails that we are aware of the flowy nature of our own perceptions. However, in what follows I aim at providing broader theoretical reasons to give this account of the phenomenology, rather than one in which perceptual and non-perceptual states are treated differently. But let us consider such an alternative, asymmetric view first. Although, as I have just argued, I think that there are good reasons for maintaining that negative phenomenal transparency is violated, the midway deflationist third strategy which I sketched above could come in handy here. The idea would be the following: we are aware of the flowy manner in which perceptions, memories, and any phenomenally charged mental episodes present contents to us (full metaphysical transparency fails), but it seems to us as if time ‘out ³² Cf. Phillips (2012). The claim that in cases in which perception does not play any significant role in our mental life, as when one is immersed in their own thoughts, it seems that we are aware of our internal time passing can be challenged. What is important to me here is that it is compatible with the fact that in certain cases we ‘lose track’ of the external time, that we are aware of the inner flow.

240

GIULIANO TORRENGO

there’ passes (full phenomenal transparency holds), and yet when we report on our being aware of an internal flow, this does not involve perceptions, but only other types of mental episodes. If we think otherwise, this is because we confuse the flowy way perception presents to us the external events with the way we are aware of our own mental life. When we diverge our attention from the passage of time out there to our own mental life, we either keep on ‘tracking’ the external flow and misattribute it to our own experiences, or we inadvertently switch to the inner flow of our non-perceptual mental life and misattribute it to the external world too. This position is problematic. Take again the example of someone visually perceiving the seagull passing by the church bell. If they report that it seems as if they are aware of the passage of time with respect to both what they see and their seeing it, this position predicts that they are somehow mistaken. I think this is puzzling, not as a description of the phenomenology, I have granted that the issue of the introspective data here is as delicate as it could be, but with respect to the contrast with perceptions of duration and succession. The idea is that if the midway deflationist explains awareness of the inner flow as an interiorizing of awareness of the external flow, then they leave the non-perceptual case unexplained. And if they explain the non-perceptual case differently, they introduce a further complication in the account which can be avoided. My aim of the last part of this final section is to show how we can have a more ‘uniform’ theory, one based on the PMT and restricted to the second strategy. We have to be careful here and not read the idea of a derivative experience of inner flow as a form of what I have called reductionism with respect to the experience of the passage of time. Reductionism is the thesis that we are presented with the passage of time in the world through being presented with qualitative temporal features such as movement, change, duration, and the like.³³ The idea that I am exploring here is rather that we become aware of our own perceptions flowing in time in the same way that we become aware of any of our experiences as flowing in time, namely through being aware of the flowy manner in which we are aware of our conscious contents in general. The first step is to clarify what it is to be presented in perception in a ‘flowy manner’ with an event. On the face of it, it sounds like a temporal mode of presentation, and thus not something compatible with positive transparency after all. Now, this criticism is not totally off the tracks, methinks. But let me dispel how we should not understand it. The idea is not to construe perception as having a temporally extended and tensed content as in the so-called modal view ³³ Torrengo (2017a). Defenders of versions of the view are e.g. Paul (2010); Prosser (2016).

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

241

of the specious present.³⁴ Following certain ideas of Husserl, some construe perceptual experience as one in which we are presented with extended events. For simplicity, let us identify an event e with which we are presented with a succession of moments . Different moments in e are presented in different ways, m3 is presented as present (say), m2 as just past, m1 as a bit more past, and possibly still others as future. Perception comes with a primal impression focus, its retention tail (and possibly a protention front)—to use Husserl’s original terminology. Whether such modes of presentation are compatible or not with positive transparency is controversial,³⁵ and indeed it is also a matter of debate as to whether or not perception requires them to begin with.³⁶ Be that as it may, we have to carefully distinguish the merely perspectival elements involved in this picture, from the dynamic element. If m1 is presented as present, it is presented as occurring roughly at the same moment in which our own experience is, and if m2 is presented as past, it is presented as just ‘over there’ in the temporal dimension. As I have argued elsewhere, the perspectival elements are insufficient for accounting for the modification that gives rise to our belief in the passage of time.³⁷ Of course we can ‘read more’ into tenses and understand them as contributing some primitive dynamic ingredient. But then, given that the whole idea of perceptual contents being tensed is problematic, why not address this primitive dynamic element directly? So what could this flowy mode of presenting events in perception be? Let us come back to the idea of saving phenomenal positive transparency by dropping metaphysical positive transparency, and keep in mind that the aim is to have an account that is not limited to the perceptual case. If we were to provide an account that works only for the perceptual case, then we would not be better off than midway deflationism. The analogy, which inspires the PMT more generally, is with seeing through a blurred glass. The visual experience presents us with objects that have determinate boundaries in a blurred way. In certain conditions, we attribute the lack of determinacy to the boundaries of the perceived objects. It is also possible to be in a situation in which we

³⁴ Cf. Dainton 2008, p. 375. ³⁵ Richardson 2014 seems to think that the spatial analogue, perspectival features, are structural features that are compatible with positive transparency, and suggests (in 2018) that perhaps also the temporal case is analogous. ³⁶ I myself (with Hoerl 2018) do not believe that perceptual contents are tensed. ³⁷ Torrengo 2018. In a nutshell, the idea is that clearly our perception of space is perspectival, but we do not come to believe that space is dynamic. Hence the reason we think time is dynamic must be based on something different than the mere perspectival element (if it is there to begin with). Jaszczolt 2020 argues that our increasing awareness of Einstein’s relativity may contribute to the explanation of how come that we can experience things in spacetime as dynamic, while knowing that neither space nor time are so.

242

GIULIANO TORRENGO

both see the objects as lacking determined boundaries, and we are aware of the blurriness of the glass itself. The suggestion is to consider the awareness of our inner flow as analogous to the latter situation. We are aware of the modification of our perceptual contents in a way that it seems to us both that time passes and that our perceptions happen in the very same flow—the phenomenal modifier is the great binder. Think of the inferential consequences of this hypothesis. According to it, we come to believe that time passes, because of the flowy character of our experiences, and such a belief does not distinguish between an internal and an external flow. It is important to notice that, on this picture, our awareness of the inner flow is derivative of the awareness of the flowy manner in which the perceptual contents are presented to us, and yet the account of the non-perceptual cases is not treated differently, as it would have been if we had followed the midway deflationist strategy. An account of the inner flow based on the PMT seems to sit better with the thought that the experience of passage is not confined to perception, and lead to a belief that the inner flow is the same as the outer flow. When we do not attend to perceptual content, we may possibly ‘lose track’ of the passage of time; but this does not mean that we do not experience time flowing when we are absorbed in our imagination, memories, mental imageries, thoughts, since they all present us contents within the flow of time. Losing track of the flow of time simply means no longer being in a position to make accurate judgements about durations, but this is hardly a reason to take (iii) as a mischaracterization of our temporal phenomenology. Now, on the one hand non-perceptual experiences are not a threat to temporal transparency, but on the other maintaining that we are presented with a stream of thoughts or memories independently of our awareness of what we think, remember (and perceive) does not necessarily capture the phenomenology of our inner flow better. According to PMT non-perceptual experiences are characterized by a flowy mode of presentation as much as perceptual ones. The theory allows us to have a uniform explanation of the perceptual and nonperceptual cases, because it entails that whatever we are presented with in experience (perceptual or not), we are presented with it in a flowy manner. Therefore, even if there is no presentational phenomenology in the case of imagining or memories, and we are not presented with events as happening ‘out there’, around us, we are still presented with contents as if they happened in time. And even if we are aware that the contents are not happening as our own experiences unfold, we are aware that our experiences unfold while we are presented with them as happening. My hypothesis, in accordance with PMT, is that the explanation is the same as in the perceptual case. We are aware of

TEMPORAL TRANSPARENCY AND THE FLOW OF TIME

243

our own experiences unfolding in virtue of being aware of their content being presented in a flowy manner. Therefore, even though supposing that we are non-perceptually presented with our non-perceptual experiences themselves as unfolding in time would not be a violation of negative transparency, I do not think we should assume that we are. In both cases the awareness of the passage of time is mediated by the awareness of how our contents are modified. The PMT allows us to have a more general account of our experience of the inner flow (regardless of whether it involves perceptions), and one that seems to have at least as much explanatory power as the midway deflationist one. Again, negative phenomenal transparency fails because it seems to us as if our experiences unfolded in time also in the case of perception. However, the positive thesis, in its phenomenal reading, holds: this feature of our own perceptions is nothing over and above a flowy content being presented in awareness.

Acknowledgements This research was funded by the Department of Philosophy ‘Piero Martinetti’ of the University of Milan under the Project ‘Departments of Excellence 2018–2022’ awarded by the Ministry of Education, University and Research (MIUR), and the projects CHRONOS (PID2019108762GB-I00) and RYC-2017-22480, both at the Autonomous University of Barcelona and funded by the Spanish Ministry of Science and Innovation (MCIN/AEI/10.13039/501100011033).

10 Does human time really flow? Metaindexicality, metarepresentation, and basic concepts Kasia M. Jaszczolt

‘Sometimes I feel that life is passing me by, not slowly either, but with ropes of steam and spark-spattered wheels and a hoarse roar of power or terror.’ Martin Amis, Money, 1984, London: Vintage Books, p. 112

10.1 Introduction: From timeM and timeL to timeE Research on time can mean many things and address many different questions.¹ It can concern ‘real’ time, as in time ontology and metaphysics. It can also concern human experience of time and its properties, with its associated questions such as whether we experience the passage of time. Or it can concern the human concept(s) of time and the ways humans represent time. As a source of evidence for the latter, it can also mean research into how we express temporal reference and thoughts about time in natural language discourse. I shall refer to these three domains of inquiry as timeM (‘metaphysical time’), timeE (‘epistemological time’), and timeL (‘linguistic time’) respectively.² In what follows I focus on timeE and specifically on the concept of time humans employ in timeM/E/L -permeated discourse, in order to contribute a new perspective (and a new hypothesis) to the discussion of the dynamic nature of timeE . On the account proposed here, timeE does not flow, and as such has properties that are compatible with those of static timeM . I argue that it has been a regrettable ¹ I thank the audiences of the following events for their comments and suggestions: American Philosophical Association Central Division, Philosophy of Time Society; Centre for the Philosophy of Time, University of Milan; University of Genoa; Faculty of Philosophy, University of Łódź; Beijing Language and Culture University; Understanding Human Time workshop, University of Cambridge; Serious Metaphysics Group, University of Cambridge; Chronos Colloquium, University of Neuchˆatel; and First International Congress on Logic, Epistemology and Methodology, University of Costa Rica. I also thank Giuliano Torrengo and an anonymous referee for their comments. ² After Jaszczolt 2018a and 2020.

Kasia M. Jaszczolt, Does human time really flow?. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Kasia M. Jaszczolt (2023). DOI: 10.1093/oso/9780192896445.003.0010

DOES HUMAN TIME RE ALLY FLOW?

245

mistake to separate (i) linguistic discussions about temporal representation in discourse and the temporal concepts they externalize from (ii) philosophical discussions about the metaphysics of time and the passage of time. While it is true that linguists concern themselves with the human concept of time and its externalization in natural languages, while philosophers address the metaphysics (and ontology) of time itself (at best in relation to questions about temporal experience and representing time to oneself ), bringing together the three domains of timeM/E/L can open up new exciting perspectives. In what follows, I aim to address both types of readership, so I provide some relevant background as I move along. I explore two lines of inquiry that arise in response to the above concerns. The first one starts with the question of indexicality of time vis-à-vis the indexicality of the first-person perspective. I address it in order to ask the question of the origin of the dynamicity we attribute to timeE and sometimes also to timeM . The second question concerns the very property of dynamicity. I ask what it means for timeE to be dynamic and where the representation of time as flowing comes from. My main objective is to attempt to reconcile the B-theoretic, static conception of the universe as multi-dimensional spacetime with the concept of the flow of time acquired by humans through still little understood experience. It will be my methodological choice to focus on beliefs and associated concepts in order to shed more light on the object they pertain to, namely timeE . I make use of the idea of the indexing of the flow to the emergent ego, calling it metaindexicality, to explain further what the perspectival nature of the flow consists of and how it can be used to reconcile this ‘metaindexically dynamic’ timeE with static, B-theoretic timeM . I then propose that beliefs about timeE be analysed as metarepresentational, semi-propositional (Sperber 1985, 1996, 2000) beliefs about of timeM , and as such as not fully understood beliefs that are, nevertheless, faithful to the theory of static universe (‘static’ in a sense that there is no flowing time. Note that quantum worlds are the opposite of static.). Next, I move to asking about the properties of the concept of timeE (TIMEE ).³ Using evidence from the way temporal reference is expressed in natural languages, I bring in the idea of temporality as epistemic modality developed in Jaszczolt 2009, 2013, 2016b. ‘Dynamic timeE ’ proves not to be essentially dynamic after all when looked at from the level of conceptual building blocks. The building blocks are static: on that level, timeE does not flow. Putting together

³ As is the custom in semantic and cognitive linguistic studies, I use small capitals to stand for a concept.

246

KASIA M. JASZCZOLT

metarepresentational beliefs, metaindexicality, and inherently modal character of timeE helps me explain the apparent flow of time through a combination of interrelated explanantia: (i) the emergent, metaindexical, ego-perspective, (ii) the metarepresentational nature of beliefs about timeM and timeE , and (iii) evidence from timeL that timeE is itself a complex concept. On the level of its conceptual building blocks, all we find is degrees of commitment to events. The more detailed structure of the chapter is as follows. In Section 10.2 I assess what would constitute the most promising line of inquiry: (i) the experience, the feeling, of the flow or (ii) the concept of the flow, perhaps as gleaned from timeL . Setting aside (i) as tangential (in spite of the importance of the topic in its own right), I conclude that pursuing (ii) will allow me to approach timeE from two important vantage points: knowledge of contemporary theories of spacetime on the one hand (that is, the direction timeM → timeE ) and linguistic evidence on the other (timeL → timeE ). In Section 10.3 I develop the first direction by placing the indexicality of timeE in the context of the indexicality of the ego-perspective. I develop a nested ‘doubly indexical’, or ‘metaindexical’ view of timeE , with its apparently inherent ‘objective’ flow and variable, subjective ‘overlay’. (I deliberately stay away here from calling it ‘experience’ or ‘perception’ of time flow in that it is debatable whether it is time itself or rather something else we experience.) Throughout this section, I address the question of the possible reconciliation of the B-theoretic, static conception of timeM as a dimension of reality propounded in modern physics with the dynamicity of timeE . The compatibility that I derive from an emergentist picture prepares the ground for Section 10.4, where I move to the role of the increasing public awareness of static timeM painted by modern physics in how timeE is conceptualized. I rely here on the concept of metarepresentation as an incomplete representation—representation of reality, physical laws, and theories of spacetime. Next, Section 10.5 moves to the other direction, namely that from timeL to timeE . Having asked how and why timeE can be dynamic in the world of static timeM , I now ask in what sense time is dynamic. As previewed above, I offer an explanation whereby dynamicity is absent on the level of conceptual building blocks of thoughts and corresponding semantic building blocks of linguistic expressions; it is only present on the level of language-specific combinations, that is on the level of conceptual and semantic ‘molecules’ used in thinking and speaking. The flow of timeE on the level of explanation pertaining to the emergent ego is ultimately the human attitude to situations (and as such, events and states if one accepts them in one’s ontology)—an attitude that I link to gradable epistemic modality. On the level

DOES HUMAN TIME RE ALLY FLOW?

247

of conceptual and semantic building blocks, timeE does not flow: it is as static as timeM of the block universe. So, since TIMEE is shown to be static on the level of its conceptually primitive components, there is no mismatch between timeE and timeM that would require an explanation. In passing, I show briefly how the essentially static concepts associated with timeE can be represented in the contextualist theory of Default Semantics.

10.2 The passage of timeM/E/L : In pursuit of beliefs and concepts There is a noble tradition of explaining time duration as the way our experience reproduces moments (Kant) or retains them (Husserl). There is also an alternative tradition going back to William James according to which we experience intervals. For Husserl (1991 [1928]), time, at least in one sense of ‘time’ he distinguished, is constituted by the flow of consciousness: it is a construct of the human mind (timeE ).⁴ The internal flow has also been variously defended by a vast number of philosophers, famously from Zeno, through Thomas Aquinas, to Husserl, Heidegger, and other phenomenologists, and currently by a number of B-theorists.⁵ Current proponents of the B-theory of timeM point out the dissonance between timeM that does not pass and timeE , with its experience of flow—a dissonance that is variously explained as part of the representational character of experience, as phenomenal experience, as an error in reporting the B-theoretic phenomenology, with further variations in the camps of representationalists and non-representationalists regarding properties of timeE or objects of experience. I will assume from the start the static ontology of B-theory and search for a way to reconcile the apparently dynamic timeE with this ontology. But I will not have much to say about the ⁴ For intensionalist and extensionalist interpretations see Hoerl (2013c). ⁵ This might be helpful to non-philosopher readers of this volume. First, the terms ‘A-theory’ and ‘B-theory’ draw on the idea of the A-series (events move from the (real) future, through the (real) present, into the (real) past) and B-series (events are ordered by the earlier-than, later-than relation, there is no real future, present or past) proposed in the seminal paper by McTaggart (1908). Now, although lumping together all extant approaches as A-theoretic or B-theoretic tout court borders on obliviousness, I am going to risk it here in that intra-theoretic differences do not affect my current proposal. So, here is my subjectively selected sample of references. For a defence of various versions of A-theory see e.g. Prior (1967, 1968, 2003); Tallant (2007, 2013); Tallant and Ingram (2012); Smith (1993, 2002); Ludlow (1999, 2013); Bourne (2006); Tooley (1997, 1999); Parsons (2002, 2003); Deasy (2017); and Baron (2017). For arguments in defence of the B-theoretic perspective, see e.g. Mellor (1998); Oaklander and White (2007); Le Poidevin (2007, 2011); Prosser (2012, 2013a, 2016); Price (2011); Rasmussen (2012); Mozersky (2001, 2013, 2015); Sider (2001); and Torrengo (2014, 2017a, b). For arguments from modern physics see Price (1996); Rovelli (2018); Ismael (2016); and Callender (2017). See also Farr (this volume), for a defence of the (much lesser discussed) C-series: time with no direction.

248

KASIA M. JASZCZOLT

feeling, or experience, of timeE . It appears to me more methodologically prudent in this study to proceed through distilling the concept of timeE (TIMEE ) and the associated concepts pertaining to its flow from discussions that tend to focus on the experience, or the feeling, of the flow.⁶ Once we have moved from the domain of feelings to the domain of beliefs and concepts utilized in these beliefs, the path will be open to employing linguistic evidence and linguistic semantic analysis of natural language expressions that externalize these beliefs. So, I will first justify this methodological move. A lot of attention has been devoted in the philosophy of time to the question as to whether the passage of timeE can be experienced. For my current (methodological) purpose it suffices to focus on one example. Prosser (e.g. 2012, 2013a, 2016) sides with the view that there is no motion or flow; it is an illusion. He argues that there is nothing that could count as perception of the flow of time. He proposes that the passage of time has the status of a possible ‘enabling condition’ for the experience of passing but equally this enabling condition could be something else (Prosser 2013a). Neither is passage represented in the mind separately from the Kantian enduring objects (Prosser 2012). Objects endure through the changes because we foreground their permanence (just as we foreground the enduring ego—a topic upon which I touch in Section 10.3). Prosser entertains here a functionalist explanation from implicated negation: if objects were perceived to merely perdure, we would have to represent a series of objects in a sequence rather than the more economical alternative of a change in one object. Functional economy makes us see the world, including the ego, in this manner.⁷ On the other hand, functionalist explanations are often found wanting. Arguably, while sacrificing ‘computational economy’ in the domain of personal identity, we would gain a more direct access to the passage of time if there were any, and as such we would in fact gain computational economy. Our perspective would simply foreground time rather than ego. In short, if one were to ditch the experience of timeE passing, one ought to do it for better reasons.

⁶ On the question as to whether the concept of the passage of time can be dissociated from the experience of passage see e.g. Prosser (2016, esp. pp. 57–58 and this volume); Piata (this volume). ⁷ On perdurance and the ‘stage view’ see also Sider (2001) and Hawley (2001). On the passage of time vis-à-vis ‘correct meta-metaphysics’ see Sider (2011). To compare, Mozersky (2015: 170) takes here a middle stance, arguing that the lack of motion or flow does not entail that time does not pass: passage is a summary concept for the temporally ordered sequence of events in the universe: change entails B-series ordering and the latter entails the passage of time; so the existence of change suffices for the existence of passage.

DOES HUMAN TIME RE ALLY FLOW?

249

The problem is that we do not have neuroscientific evidence that would give us reliable answers. Debating the status of the experience or feeling of the passage is probably better viewed as a question in its own right, and addressed in the domain of neuroscience rather than as a subsidiary question in an inquiry into the timeE –timeM link. I believe that this question is best resolved through empirical, experimental methodology that takes current cutting-edge philosophical theories as its starting point for testing. Second, humans’ affair with time is inherently dual: we live and think in time, where there appears to be one, objective now, but we can also think and speak about time, fully aware that what counts as the present moment changes. Hoerl and McCormack (2019) present this as a dual system view (à la Kahneman 2011). Intricacies of their ‘naïve theory of time’ notwithstanding, it is evident that temporality (their ‘thinking in time’), as well as what I would call ‘metatemporality’ (their ‘thinking about time’), are aspects of the human affair with time. Here, to reiterate, we opt to focus on the concept of time that humans employ in thinking in, as well as about, time rather than on experiences since it will be more useful for the purpose of this inquiry.⁸ Crucially, concepts are the stuff that linguistic meanings give us a window on—and we have access to a well of empirical evidence on temporal reference across very diverse languages. I am therefore opting for a method that will let me utilize timeL as a source of evidence for timeE , and as such I opt for focusing on concepts and beliefs. On Torrengo’s (2017a) deflationist account, there is no sensation of the passage of time, merely the belief. The sensation of the passage is a so-called ‘primitive phenomenal modifier’ that influences (modifies) our experience. As such, the modifier influences the representational content of experience.⁹ Although the jury is out on the latter, what I need to borrow and adapt will be orthogonal to that debate.¹⁰ It takes me close to where I want to be, that is beliefs about time—as well as their linguistic externalizations. I will leave aside ⁸ Hoerl and McCormack make some controversial claims there, for example pertaining to the ability of animals to reach the ‘metatemporal’ level: they claim that animals do not represent time. They also point out that the two levels end up in a contradiction in the naïve theory in that people believe, and at the same time disbelieve, that there is objective present. For a response see Prosser (2019). These arguments will not concern me further: what matters for my story is that the two levels coexist. ⁹ See also Hoerl (2014) on an explanation of the phenomenology of passage through movement and change. ¹⁰ Phillips (2013b), for example, develops a partially empirically supported argument that defends the phenomenal immediacy of the experience of duration by making it relative, occurring in the context of other mental activity that can be influenced by such factors as fear, body temperature, or being under the influence of certain drugs. See also Phillips (2014a) and Boccardi and Perelda (2018) on the inheritance of any temporal property of the objects of perceptual experience by the experience itself— arguably attributable to Husserl (1991 [1928], see Hoerl 2013c). In addition to the naïve view and Torrengo’s deflationism, other stands include reductionism (experience of the flow is the experience of change), the view that reality itself is tensed (versions of A-theory), and the view that the flow comes

250

KASIA M. JASZCZOLT

the debates concerning the feeling, the experience of the flow, and focus on the belief and its expression in natural language discourse. Beliefs will be taken to be propositional, in the sense of propositions of the cognitive-semantics level of representing meaning, or, in other words, contextualist, pragmaticsrich propositions where a multitude of sources of information (in addition to the sentence’s logical form) can contribute to the truth-conditional content.¹¹ It is the epistemology of belief and knowledge about timeE and timeM , combined with the semantics of temporal reference, that will give me the necessary explanantia.

10.3 ‘My time’: Metaindexicality ‘Those who kill time commit temporal suicide’ Flaherty (2011: 29).

10.3.1 An appeal to emergentism According to the theory of general relativity, there is no one single relativistic spacetime—there are many, differing by how, and how much, they are curved. So, any theory of timeM has to anchor the latter to points in such spacetimes.¹² It goes without saying that in order to address timeM in a way that is accessible to human thinkers, we have to focus not on the perspectival quality (indexicality) of timeE but on the perspectival quality of the ego as the thinking agent. Based on an even perfunctory understanding of relativity, or what I analyse in Section 10.4 as a metarepresentation (an incomplete representation of a representation of reality that non-physicists laypersons arrive at in their minds), the first question that arises is ‘where am I in these relative spacetimes?’. Do I endure, persist as an ego, on this level of description of the universe? Or, perhaps, all there is, is events, situations, pieces of information, and the like, perduring, scattered through the different spacetimes? Unorthodox as it may seem, it is not unassailable that we think that ‘real time’ (timeM ) passes, in that human popular assumptions embrace, even if only from the attitude towards content (see Torrengo 2017a). The latter view can be taken in the context of attitudinalism towards the ego across episodic memories that I discuss in fn 25. ¹¹ I give the semantics for such contextualist propositions (semantics of ‘merger representations’) in the theory of Default Semantics in e.g. Jaszczolt (2005, 2010). The origins of the idea can be traced to the post-Gricean, ‘radical’ pragmatics whose landmarks include the Atlas-Kempson thesis (Atlas 1977, 1979, 1989; Kempson 1975, 1979, 1986), followed by Relevance theory (Sperber and Wilson 1995 [1986]). ¹² See e.g. the discussion in Skow (2015, Chapter 9).

DOES HUMAN TIME RE ALLY FLOW?

251

partially, the basic tenets of general relativity. After all, Einstein’s ideas have been percolating into daily concepts for close to a century and through various channels, such as popular science publications, audio-visual and more recently social media, or classroom teaching. As long as we afford these assumptions an attenuated, ‘metarepresentational’ status, they are undoubtedly there. So, our TIMEE is a naturalized concept. Slowly but surely, the commensurability of timeM with timeE permeates our everyday thinking and may even be beginning to take shape as static timeM/E —arguably as a result of this increasing familiarity with scientific theories that percolate to general knowledge. This increasing awareness of the core idea of Einstein’s special, and even general, theory of relativity may prove to be a game changer in the philosophy of time. Arguably, just as we move through space without conceptualizing space as flowing or passing, we also move through timeE and, with the increasing awareness of the theory of relativity, we reflect on the subjectivity of the timeE flow, or even on the sheer metaphoricalness of expressions such as ‘Christmas in approaching’ or ‘We are entering a period of recession’. After all, we don’t observe any absolute, non-deictic direction of timeE flow, neither do we observe any absolute, non-deictic rate.¹³ Directions such as ‘the past is behind us’ and ‘the future is ahead’ are culturally imprinted and vary;¹⁴ the experienced length of the interval is clearly subjective and, more importantly, is recognized as such: timeE ‘flies’ when you are happy, but it ‘drags on’ unbearably when you wait for the results of an examination. Now, notwithstanding this slow percolation of the static universe view into our everyday thinking, we behave as if timeE were really passing—as if the passage itself were objectively there and as if only the rate of this passage were subjective, reflecting how we feel about this passage.¹⁵ In view of my faith in the popular awareness (even if grossly incomplete) of the static conception of spacetime, we now obtain a summary, from the ego perspective, as in (1). (1) (a) I believe that timeM does not pass: it is a dimension of spacetime. (b) I believe timeE passes. (c) I believe that this passage can be objectively measured. (d) I believe that the objective measurement of timeE can differ from the gauged duration reported by the experiencing subjects. ¹³ On the rate of the passage of time see e.g. Smart (1949); Prior (2003); Olson (2009); Mozersky (2013). ¹⁴ References are ample; see e.g. Moore (2011) on the mixing of two frames of reference in Aymara that explains how ego faces the past. ¹⁵ See here Galton (2011) on the limited applicability of space metaphors for time in human languages. But travelling back in time notwithstanding, the analogy is still rife.

252

KASIA M. JASZCZOLT

The following sections will address the question of the inherent consistency of this set of beliefs. One way in which the question of the flow of timeE is commonly addressed is by asking what it is that really moves: time, events, or the subject. This question is widely discussed in cognitive semantics, mainly due to the abundance of pertinent metaphors that employ the MOVING EGO or the MOVING OBJECT conceptualization.¹⁶ Does time carry us along with it or do events slide into and away from our reach? More importantly, does the change have to be explained as the ‘passage’ of timeM or a ‘passage’ of something else—for example one’s own youth or life, as in my epigraph from Martin Amis?¹⁷ Or perhaps change is not a ‘passage’ at all?¹⁸ It is feasible to adopt a construal on which it is the ego that changes, endures, persists through time,¹⁹ and as such constitutes a construct without which moral laws and many other life-constituting ideas and values could not be proposed. My argument in what follows will exploit this importance of the indexing to the ego, whereby we index the dynamicity to ourselves. I will now pick up the inquiry where others left it, and will begin with two captivating accounts: Ismael’s (2016) How Physics Makes Us Free and Callender’s (2017) What Makes Time Special?. Ismael’s leading idea is that ‘the micro-laws create the space for emergent systems with robust capabilities for self-governance’ (Ismael 2016: xi). The ‘I’ emerges in the world as a complex, self-governing system: self-governance involves the creation of an internal point of view on the world, and so it opens up the psychological space for the growth of the self, from the early kindling of sensorimotor awareness to a fully developed autobiographical subject. Ismael (2016: 39)

On this view, there is no mismatch between our freedom to act and make choices and the deterministic universe. On the level of units that are of interest to physicists, the universe is governed by symmetrical laws: there is no difference between the past and the future, or, as Price (1996, 2011) puts it, what is the past and what the future depends on where we stand— on the ‘Archimedes’ point’; there are ‘time-symmetric, deterministic laws’ ¹⁶ ¹⁷ ¹⁸ ¹⁹

See e.g. Lakoff and Johnson (1999); Evans (2003); Gibbs (2017); Piata (2018 and this volume). On this topic see e.g. Rasmussen (2012). See e.g. Mozersky’s (2013, 2015) proposal along these lines. See e.g. Price (2011); Prosser (2012, 2016).

DOES HUMAN TIME RE ALLY FLOW?

253

(Ismael 2016: 216). When you add complex, self-governing systems, they add their perspective. Humans are such self-governing complex systems and the past, the present, the future, and time flow are their products. These products are not an illusion as Einstein claimed, neither do they create incongruity: they are real, and they also include causation: The laws of your everyday world are the laws that describe the emergent structures you interact with. At that level, indeterminism reigns, irreversibility is the norm, and causal relations are handmaids to effective action. Ismael (2016: 221)

Callender’s (2017: 142) argument is not dissimilar in its outcome in that he points out that flowing timeE is the most sensible explanation of reality that humans came up with: ‘[T]ime is that direction on the manifold of events in which we can tell the strongest or most informative stories’. As he says, time is ‘the great informer’, it provides an explanation. Put differently, ‘time’ is a kind of honorific applied to the directions on which the best system settles to make its principal algorithms. Its origin lay in whatever asymmetries in the distribution of events makes this algorithm best. Callender (2017: 150)

Like Ismael, he embraces emergentism and the association of the flowing timeE with the enduring ego—an ego that is biased in its attitudes towards the past and the future, unreliable in its estimate of the passing timeE , or, even, if one goes the whole hog to the Dennettian story, an ego that may itself be a fiction. There is much more to be said on the question of the compatibility of timeE of complex systems like us and timeM as a dimension of the static universe. It is a matter of asking the right questions, such as (i) how exactly the de se perspective explains timeE , how and why time is dynamic, or (ii) where exactly the dynamicity comes from and how crucial it is to the human beliefs about time. I discuss (i) in the remainder of Section 10.3 and in Section 10.4, and (ii) in Section 10.5.

10.3.2 De se, IEM, and inheritance I can’t help being perplexed by the questions addressed by A-theorists such as ‘Do past and present exist, or is it only the present?’ or ‘Do things begin

254

KASIA M. JASZCZOLT

and cease to exist?’²⁰ Taken in the light of perspectivalism on reality, such as Ismael’s evolving frame approach discussed above (or, indeed, any other B-theoretic outlook that is still favourably disposed to flowing time as more than an illusion), they fall prey to the identification of the transcendental with the indexical: there is no metaphysical ‘it’ that they can address without vagueness or underdetermination. There is no equivocation though: ‘time’ is timeM —just there are two different levels from which we can describe it, the ontological and the human. We are interested here in the human, assuming the B-theoretic ontology. In view of the latter assumption, we do not have to address the question of the relation between the indexicality of the ‘I’ perspective and the indexicality of time in the sense of ‘here and now’: the latter is not among the characteristics of timeM . Does timeE appear to flow because it is indexical? Yes and no. It appears to flow because we construe it as perspectival, indexical, but indexed to ourselves rather than to some fixed, objective now. As Torrengo (2018: 1056) points out, it is perfectly possible to conceptually dissociate dynamic qua perspectival timeE from inherently, transcendentally dynamic, qua A-series dynamic, timeM : one can uphold the first while denying the latter, just as we have done here. Inferring more from indexicality can lead to an equivocation. However, the really interesting question is not whether timeE can be dynamic qua indexical while timeM is static, but rather whether B-theorists are obliged to accept that such a dissociation is theoretically beneficial. Arguably, the onus of proof is on the ‘dissociationists’ rather than the ‘convergentists’: after all, taking timeM and timeE to be both static or both dynamic pre-empts the need to explain the incompatibility. To reiterate, my ultimate answer is going to be that they are both inherently static, so my proposal will score well on this argument. Now, while associating the dynamicity of timeE with the indexicality of the ego perspective, it has to be taken into account that there are different categories of de se thought. When I think ‘I have a toothache’, I cannot misidentify someone else as myself and misattribute someone else’s toothache to myself. On the other hand, when I am confident that I am looking at my own reflection in the mirror and think to myself ‘I have slimmed’, I may be mistakenly attributing someone’s slim silhouette to myself if I am, unbeknownst to me, looking through a clear glass panel at someone who from a distance looks ²⁰ For the first question, see e.g. Bourne (2006) on presentism (only the present exists); Tooley (1997) on the Growing Block theory (only the past and the present exist); Cameron (2015) on the Moving Spotlight theory (eternalism: the past, the present, and the future exist) and Skow (2015) for a critical discussion. For the second question, see Deasy’s (2017) proposal of transientism, pastism, and permanentism (analogous to the above three-way distinction) and a rejoinder from Tallant (2019). Among many others of course.

DOES HUMAN TIME RE ALLY FLOW?

255

like me. After Shoemaker (1968), the first, ‘unmistakably de se’ category is said to be characterized by immunity to error through misidentification (IEM).²¹ In other words, one can entertain thoughts about oneself from an ‘internal’ or from an ‘external’ perspective, where only the first one has the property of IEM.²² Arawak languages of South America, for example, have markers of visual evidentiality for talking about one’s own experiences—experiences that are unmistakably ‘mine’ (see Aikhenvald 2018)—a little like the distinction in English between remembering doing something (the ‘V-ing’ construction) and remembering that I did it (the ‘I V-ed’ construction), where, according to the common view in the semantics of English, the latter is open to misidentification of the agent (I may ‘remember’ that I suggested going on holiday to Sicily while it was my husband who did so).²³ The relevance of the distinction is this. TimeE appears to be dynamic because I make sense of static timeM in that way. As such, it is dynamic when (and only when on this construal) it is subsumed under the first-person indexical. Analogous to the first-person indexical, beliefs can be immune to misidentification as regards the present, the past or the future, or they can lack this immunity. I call this temporal IEM. Now, since, as I have shown elsewhere (Jaszczolt and Witek 2018), IEM correlates better with situations and speech acts than with particular linguistic items or grammatical constructions, ‘my time’ and ‘my place’ will likewise correlate with situations, eventualities, speech acts, or goals—and ‘my time’ and ‘my place’ will be equally immune, or not immune, to misidentification errors. The flowing time, being founded on the ‘stringing together’ of events that could be misidentified as already past while still in the future, or still present while long gone, will ‘inherit’ this property of having IEM as a binary (+/–) parameter. This construal shows that we can have a plausible theory on which timeE is, so to speak, ‘doubly indexical’: the past, the present and the future are such not only with respect to the thinking agent as assessed from the external perspective (say, I know that my graduation from Oxford is in the past), but also with respect to the worldview of the agent, as assessed from the agent’s internal perspective, where the two perspectives can come apart (say, I may wake up and strongly believe, under the influence of my dream, that I haven’t yet finished my DPhil at Oxford and that graduation is still in the future). We can say that an event is externally, objectively (but covertly) in the past and internally,

²¹ See contributions to Prosser and Recanati (2012) for recent discussions. ²² See e.g. Recanati (2007, 2012). ²³ See Higginbotham (2003).

256

KASIA M. JASZCZOLT

subjectively (but overtly) in the future. This distinction will become important in the next section. ‘Metaindexicality’ of timeE (qua indexicality of the de se thought) has further support in generalizing the IEM phenomenon beyond the ‘I’ thought. Cappelen and Dever’s (2013) have argued that (i) IEM is not inherently associated with indexicality, and (ii) indexicality is not inherently associated with the SELF: rather, as a rule, humans represent the world in a perspectival way because they have no choice but to do so. Recanati (2012) calls such thoughts ‘implicitly de se’: there is no reference to the self in their content but the thought is evaluated with respect to the thinking agent. On Cappelen and Dever’s construal, our thoughts about the world are such implicitly de se thoughts: timeE flows there because it flows in my imperfect representation of the static universe—or, as I call it in Section 10.4, in my skewed, imperfect, not fully propositional metarepresentational beliefs about timeM . This double indexicality of timeE , namely the first-person indexicality and the indexing to the now, are on this construal inextricably linked in an embedding relation, and thus as metaindexicality: the ‘metaindexing’ of the dynamic timeE is always to the SELF. It is not traceable through any linguistic analyses of grammatical or lexical categories but rather relies on the functional category of indexicality, derived from the constraint on what sense we can make from the transcendental universe of the physics of spacetime with its symmetrical, static timeM (Price 1996; Ismael 2016).²⁴, ²⁵

10.3.3 Metaindexicality and the subjective flow The ability to gauge the flow of time and the time distance between events has evolved in humans and in other species to become fairly reliable for their ²⁴ I develop the idea of functional indexicals in Jaszczolt (2016a, 2018b). ²⁵ It is worth remembering in this context that in the endurance-perdurance disputes, discussed in passing in Section 10.2.1, the continuity of the ego is questioned. Chadha (2019), for example, argues that remembering one’s actions or thoughts does not unequivocally mean that we experience the identity of our ‘present self ’ with that past agent. He points out that according to the Buddhist tradition, ‘[m]emory does not require an ontologically distinct persisting single self; it can be explained by manifold causal relations between discrete conscious events’ (p. 123). The identification with the ‘past selves’ proceeds through the, Dennettian in spirit, construction of a coherent narrative: ‘the episodic sense of the self depends on the ability to (re)construct our narrative self-conception’ (p. 132), not on the experience of this identity. So, it is our attitude to reality, not reality itself, that makes our time flow. Although attitudinalism is perfectly compatible with the B-theoretic ontology I adopt here (see here e.g. Torrengo 2018), exorcising the ego is not: metaindexicality rests on the assumption of the enduring ego. Like in Section 10.2.1, however, I will have little to say about Dennettian purge of the self, finding the emergentist view much more palpable. But it has to be pointed out that both still remain largely in the realm of hypotheses, necessary as ‘pegs’ for ‘hanging’ the theories that support them, but hypotheses nevertheless.

DOES HUMAN TIME RE ALLY FLOW?

257

respective purposes. They are far from constant and objective but not entirely unwieldy. According to Flaherty’s (2018) empirical findings, there is a regularity to the distortions in the felt passage of time. (I refrain from using ‘experienced’ or, as he does, ‘perceived’, although, to reiterate, my discussion will remain orthogonal to that debate.) By analysing 740 relevant narratives, he concluded that distortions correlate with the ‘density of experience’ per some objectively measured unit of time. ‘Density of experience’ means the density of stimuli when these are affected by the subject’s effort to control one’s own experience. ‘Temporal compression’ means that time is felt to pass quickly. ‘Protracted duration’ means that time is felt to pass slowly. In the following quotation I have replaced ‘perceive’ with ‘feel’: We [feel] the passage of time as protracted duration when the density of experience per standard temporal unit is high. We [feel] the passage of time as synchronicity when the density of experience per standard temporal unit is moderate. And we [feel] the passage of time as temporal compression when the density of experience per standard temporal unit is low. Flaherty (2018: 19)

There is also the so-called Kappa effect: when the physical distance between two stimuli increases, the felt time lapse also increases. Along the same lines, watching a scene in different spatial scales almost directly proportionally affects the time interval it appears to occupy. Next, emotions and other mental activity can affect it too: boredom or fear slow the flow, while enjoying oneself speeds it up.²⁶ Finally, a state of extreme concentration associated with a high level of skill produces effortless action where both timeE and the sense of the self disappear. In Csikszentmihalyi’s (1990) celebrated psychological theory of optimal experience, it is called a state of flow. ‘Flow’ used in this sense is associated with happiness derived from the total immersion in what one is doing. Such activities are always autotelic activities: they are performed for their own sake, without any external concerns about the ego; self-consciousness is considered an enemy of flow. The flow is characterized as a state opposite to psychic entropy (disorder): it brings about harmony and makes the ego ‘grow’ to be more complex. I believe a way to account for such distortions lurks in Torrengo’s (2017a) idea of the ‘primitive phenomenal modifier’ that I have made brief use of in Section 10.2.2, as long as we adapt that operator to serve a somewhat different ²⁶ For an accessible discussion in popular science see e.g. Cossins (2019). See also Prosser (this volume).

258

KASIA M. JASZCZOLT

purpose from that originally intended. What we have to account for is this: even if we know, as assumed here, that timeM does not pass, and we know that our feeling of passage we owe to our phenomenology, we also know that our feelings can be distorted as compared with the norm. Where does this norm fit in? Torrengo (2017a: 182) writes, the influence of the feeling of the passage of time on the representational content may not be invariant. For instance, if the felt ‘pace’ of the passage of time can be influenced or can vary through experiences, then there will also be a variation in the way the concurring mental episodes feel to us.

To recall, on his construal, experience has representational content but flowing time is not directly part of it: the flow is its phenomenal aspect. But if the flow is ‘part of the experiencer’, so to speak, then we need a more complex structure. We represent the world to ourselves and the phenomenal aspect can distort these representations, but we can also ‘go back and check’, so to speak, our representations against new representations and find them wanting. For example, we can notice and register temporal compression or protracted duration. At this point, we form beliefs about the content of our experience and introduce the concept TIMEE as the dimension on which we compare that content. I don’t know if the acquired awareness of the distortion is part of the new experience or a belief —perhaps it is experience whose representational content consists of metarepresentations (representations of earlier representations). Be that as it may, to be faithful to my methodology, I am not going to pursue questions to do with properties of experience but focus instead on the concept: TIMEE that appears in our beliefs at such a stage of the above revelation. Let us now consider the summary in (1) again, repeated below. There is no online awareness of the dissonance between (1a) and (1b) on the level of beliefs and associated concepts, where by ‘online awareness’ I mean awareness during thought processes rather than in autoreflection. But there is an online awareness of the subjectivity of the measurement expressed in (1d). (1) (a) (b) (c) (d)

I believe that timeM does not pass: it is a dimension of spacetime. I believe timeE passes. I believe that this passage can be objectively measured. I believe that the objective measurement of timeE can differ from the gauged duration reported by the experiencing subjects.

DOES HUMAN TIME RE ALLY FLOW?

259

I will now need an operator, or a ‘qualifier’, that will capture the relation between timeE and timeM that I want to express. For this purpose I will loosely follow Torrengo’s (2017a) idea of a primitive phenomenal modifier introduced earlier but I will also ‘misuse’ it in some essential ways. First, my qualifier does not magically ‘transform’, so to speak, static timeM into the phenomenon of dynamic timeE . Yes, it turns symmetrical timeM into a ‘human timeline’: stringing experiences that result in the SELF. But ‘human timeline’ does not yet explain the apparent dynamicity of timeE . We will have to leave the question as to where dynamicity comes from until Section 10.5. First things first. I grant that our experience may not contain the flowing timeM and in order to get going, we need some magic wand. Using my terminology, Torrengo’s proposal can be summarized as in (2), where Φ stands for his primitive phenomenal modifier. (2)

Φ (timeM ) = timeE

But, as was said earlier, a ‘phenomenal modifier’ is not exactly what we need. On one hand, it is compatible with my metaindexical account because we can unpack Φ as my ‘double indexing to the ego’ (stripping off the label ‘primitive’). On the other, Torrengo’s theory is about how time figures in experience, while I will be concerned with the level of temporal beliefs.²⁷ To digress somewhat, his modifier seems to show that proposing to ‘kick up’ the study of time experience to the domain of neuroscience may be a promising move; while the modifier is a useful theoretical construct, it is also a ‘black box’: it covers up the fact that we have no direct insight into our feelings and experiences. But we could make use of a similar ‘equalizer’ concept, a function, by moving to the level of beliefs and associated concepts and addressing the relation between timeE and timeM there—which is exactly where Torrengo leaves it. What we need instead is a qualifier—a two-tier one, with first- and secondorder operators: the ‘objective’ for the passage of time and the ‘subjective’ for the felt rate of this passage. I call the first the ‘objective covert qualifier’ (OCQ) and the latter the ‘subjective overt qualifier’ (SOQ), where ‘covert’ and ‘overt’ refer to the online awareness as assessed above for (1c). My operators are no longer ‘phenomenal qualifiers’. For the reasons that will become apparent in Section 10.5.1, they are theoretical constructs that allow us to associate timeM with timeE , reflecting the beliefs in (1)—constructs that capture my explanation of this association developed throughout Sections 10.3–10.5. We obtain the nesting representation in (3). ²⁷ See my disclaimer in Section 10.2.1.

260

KASIA M. JASZCZOLT

(3)

SOQ (OCQ (timeM )) = timeE

A strong argument for the analytical utility of the SOQ operator comes from the presence of its clear correlates in natural language, such as adverbials (‘time passes slowly’), motion verbs that incorporate manner (‘time flies’), or evidential disclaimers (‘time seems to fly’). They either map onto SOQ biuniquely (‘slowly’) or at least their core semantic characteristics map onto SOQ bi-uniquely (manner of movement inherent in ‘flies’). Crucially, they all introduce epistemic detachment of the type that is different from the one that can be captured by the modifier for the passage itself: I believe that timeE passes, but I also know that my experience of the length of the interval may be distorted. My next steps, utilizing the schema in (3), are as follows. In Section 10.4 I develop a proposal of timeE as an imperfect representation of timeM that will pave the way to conceiving of both of them as static. In Section 10.5, I present TIMEE as a complex concept, composed of conceptual building blocks that are themselves static. This will complete the picture: timeE , ‘dynamic’ in virtue of being indexed to the emergent ego (Section 10.3), is a metarepresentation of static timeM (Section 10.4) and retains the property of being static precisely because it is indexed to the ego and, as such, reflects not a flow or passage but rather human attitudes to events and states (Section 10.5) that range from strong commitment (such as well-remembered or well-planned events) to low commitment (such as events that might happen or might have happened).

10.4 Metarepresentational beliefs about timeM While presenting her emergentist account of time flow, Ismael (2017: 35) points out the following commonly shared conflict in our assumptions: The apparent conflict between the familiar, flowing time of everyday experience and the static time of the Block Universe has a stubborn way of reasserting itself as a substantive and all-important metaphysical disagreement, even in my own mind. It is a reminder of the constant tension in the human between the transcendent and embedded viewpoints, which is in its turn the product of the peculiarly human form of mindedness.

As she says, understanding the content of our experience, ‘the logic of temporal perspectives’ (p. 24) will lead us to better understand the link between the concept and its object. She defends the so-called Temporally Evolving Point of View, or the ‘evolving frame’, whereby a string of representations from the

DOES HUMAN TIME RE ALLY FLOW?

261

agent’s deictic centre (the ‘here and now’ perspective) draws on her memories about the past and inferences to do with possible future. Put simply, we focus on the content and on how it is put together—analogies with a Facebook timeline or a photo album with spaces reserved for future snapshots spring to mind. While partially subscribing to this picture, I think it does not offer the full story on the ‘apparent conflict’. It is possible that there isn’t even an ‘apparent conflict’ in that, as proposed in Section 10.3 and followed up in this and the next section, human agents do embrace, to a sufficient degree, the conception of the static universe, and the only relativization that there is, is the relativization to the ‘I’ perspective. Once we adopt the ‘I’ perspective, or what I called earlier metaindexing, we can talk and think about the motion in space or the motion in time, just as we can talk and think about space spinning or moving out of one’s field of view or time moving from the future, into the present, and then into the past. These are more than just decorative metaphors; they are ways of conceptualizing experience. I will propose that asking ‘Does time flow?’ or ‘Why does time appear to flow?’ is in fact posing a rather naïve question, analogous to ‘Does space flow?’ or ‘Why does space appear to flow?’ It is not normally seen as naïve only because the conceptualization of time as flowing is so much more entrenched in running our life cycle and the comprehension of relevant scientific facts so much harder.²⁸ But what exactly does it mean that the scientific conception of spacetime has percolated to common knowledge ‘to a sufficient degree’? I suggest that the conception of a static, block universe has the status of what Sperber (1985, 1996) calls a semi-propositional representational belief. In his anthropological research, he proposes a distinction between factual and representational, and orthogonally between propositional and semi-propositional beliefs for the purpose of explaining cross-cultural differences, and rationalizing what may look like irrational beliefs when the latter are taken out of their cultural context. A belief is semi-propositional when it is held in spite of not being fully understood. So, for example, ‘Red giants become white dwarfs’ held by a person with little familiarity with astrophysics, or, for our purposes, ‘Time is a dimension of static spacetime’ or ‘The universe is governed by symmetrical laws’ are for most people who hold them semi-propositional beliefs. Such beliefs are reflective (as opposed to intuitive, spontaneous), in the sense that people are aware of holding them, and people hold them in virtue of holding some second-order beliefs

²⁸ See here Heidegger (1996 [1927]) and my discussion in Jaszczolt (2009).

262

KASIA M. JASZCZOLT

about them.²⁹ They are often popular representations of a scientific representation of reality and as such are metarepresentational (Sperber 2000). It is hardly controvertible that popular understanding of the concept of a static universe falls within this category of metarepresenting (qua a popular representation of a scientific representation) that involves only partial understanding. By the same token, the concept PASSAGE OF TIMEE employed in such beliefs is not a fully understood concept: it is a metarepresentational concept, akin to concepts employed in scientific hypotheses, held as ‘promissory notes’, so to speak, marking a need for further theoretical or experimental research. Using our terminology, we can say that PASSAGE OF TIMEE equals a metarepresentational PASSAGE OF TIMEM , in virtue of the metarepresentational link, as summarized in (4). Quotation marks stand for the metarepresented content. (4)

timeE = ‘timeM ’

According to Sperber, metarepresenting is an important evolutionary achievement that fosters understanding, cooperation and communication. It is more than plausible to fit the representation of timeM precisely there; it is a partially understood embrace of a theory of the universe, held because factual and fully propositional beliefs about timeM are not available to us laypersons. Viewed in this way, the belief that timeE flows arises out of the awareness of the semipropositional character of the belief in the properties of timeM , as summarized in (5). (5)

TimeE flows = ‘TimeM doesn’t flow’

It then follows that the belief that timeE flows must itself be representational and semi-propositional: by definition, it will never become factual and it will never become fully propositional because it is founded on the assumed definitional difference between timeM and timeE —see here my earlier summary in (1) and the beliefs in (1a) and (1b). It will never become factual because it makes use of complex, rather than primitive, concepts, and the building blocks of these complex concepts that are not themselves temporal remain below the level of awareness. Put simply, the staticM /dynamicE clash remains unexplained to us, which makes timeE metarepresentational as well. This is the topic I take up in Section 10.5. To summarize, in this section I have addressed PASSAGE OF TIMEE , starting with drawing attention to the all-pervasive sense of conflict between what modern science tells us about timeM and what experience, and thinking about ²⁹ See Sperber (1996: 89).

DOES HUMAN TIME RE ALLY FLOW?

263

everyday experience, tell us about timeE . I invoked the idea of metaindexing as a partial explanation of this conflict and have now supported it with the observation that our representations of timeM have the status of partially understood (semi-propositional) representational beliefs, that is popular representations of scientific representations. I also concluded that both the belief in the static nature of timeM and the belief in the dynamic nature of timeE are representational, semi-propositional beliefs. This creates a conceptual bridge between the metaphysics of time and our thoughts about time, explaining the apparent incompatibility between static timeM and what now looks like its metarepresentational dynamic overlay. But one part of the explanation is still missing: I talked about why and how, in virtue of what, timeE is dynamic. Now I want to address the final question: in what sense is it dynamic? This will lead us to the question as to whether timeE , as represented in TIMEE , is essentially dynamic. Needless to say, a negative answer will clinch the argument for the compatibility of timeM and timeE , at the same time fitting snugly with the earlier arguments from metaindexicality and metarepresentation.

10.5 TIMEE as a complex concept 10.5.1 Complex concepts and their building blocks I now move to the questions as to what exactly the dynamicity of timeE is and how crucial it is to human beliefs. Faithful to my adopted methodology, this takes us to beliefs and associated concepts, but also further into their reflections in natural language where these concepts are lexicalized, grammaticalized, or left to pragmatic inference in the context of discourse. The linguistic turn included in my method requires a further explanation. One of my methodological assumptions is that the forms of temporal reference employed in natural languages give us a window on TIMEE . However, they only give us a window on online thinking, not on the ultimate properties of temporal concepts per se. In other words, how we speak tells us about what Slobin (1996) calls ‘thinking for speaking’: the way concepts are employed for the purpose of externalizing, sharing thoughts in discourse. But the concepts we employ for the purpose of speaking are often complex concepts rather than primitive, ‘atomic’ building blocks—or, even, ‘primitive subatomic’ if a lexical item or grammatical category count as an ‘atom’.³⁰ I entertain the ³⁰ Wierzbicka (e.g. 1996) offers an influential, empirically supported defence of such ‘building blocks’, called semantic primes, in her theory of Natural Semantic Metalanguage.

264

KASIA M. JASZCZOLT

possibility, based on analogous analyses in other semantic domains, that timeE may be conceptualized using precisely such complex concepts—concepts such as those of the temporal location in the past, the present, and the future, or of time flow and its ratio. The complexity of TIMEE FLOW, and RATIO OF TIMEE FLOW might reasonably be thought of as universal, as well as accountable, on the level of conceptual semantics, by the operators pertaining to the two-tiered modifier proposed in (3) in Section 10.3.3 and repeated below. (3)

SOQ (OCQ (timeM )) = timeE

But the complexity of TEMPORALE LOCATION and TEMPORALE INTERVAL call for deconstructing (3). We can begin as in (6), where the construction in (3), taken on the level of beliefs, leads to the concept: TIMEE . (6)

(SOQ (OCQ (timeM )) = timeE )belief = TIMEE

To reiterate, I have assigned to my SOQ and OCQ operators a vague role of theoretical constructs, instead of following Torrengo and awarding them the status of phenomenal modifiers, in order to be able to address the question as to whether TIMEE (and TIMEM , for that matter) is a complex concept. If so, then the experience of time will likely be more complex as well. Now, if TIMEE is a complex concept, then we want to know its composition. Once we know it, the composition of concepts such as PASTE , PRESENTE , FUTUREE will follow suit. Temporal reference is conveyed in natural languages in a variety of ways. In many languages, it is commonly conveyed by grammatical tenses in combination with grammatical aspect, like in English. In tenseless languages such as Yukatek Maya, Mandarin Chinese, Paraguayan Guaranı´, Burmese, Dyirbal (an Australian Aboriginal language), West Greenlandic (Kalaallisut, an Eskimo language), or Hopi, it can be conveyed by such devices as aspect and mood markers, modality and evidentiality markers, different inventories of temporal adverbials, or pragmatic inference from context, sometimes making use of default interpretations of constructions without overt temporal markers.³¹ In Yukatek Maya, markers can convey information about aspect, mood, and temporal distance (such as proximate relative future, immediate relative past, recent relative past, or remote relative past). Immediate relative past is ³¹ A language can be called ‘tensed’ when it has grammaticalized expressions for temporal reference. These expressions have to be absolute rather than relative (the coding time normally constitutes the deictic centre). See e.g. Tonhauser (2011). The theory of so-called covert tenses can muddy the picture but we will not concern ourselves with deep syntax and with the categories of covert, unarticulated components proposed there in order to ‘save’ universal grammar (see e.g. Matthewson 2006). See also Caudal (this volume) and Everett (this volume).

DOES HUMAN TIME RE ALLY FLOW?

265

exemplified in (7). Notice that the temporal markers are relative; they are not markers of absolute temporal location. (7)

Táant in xok-ik le periyòodiko-o’. IMM 1SG read-INC(3SG) DEF newspaper ‘I have/had/will have just read the paper.’³² (adapted from Bohnemeyer 2002: 9)

In addition to not marking overtly temporal reference, Yukatek Maya has very few expressions of temporal ordering such as ‘after’ or ‘while’. Temporal adverbials are optional. So, overt marking of time is almost lacking, while other related concepts take preference in lexicalization and grammaticalization. The combination of aspect, mood and context is also used, for example, in Mandarin (Lin 2005), West Greenlandic (Bittner 2005), or Hausa (Chadic, Afro-Asiatic, Mucha 2013).³³ There is ample evidence from natural languages that it is not FLOWING TIMEE or TEMPORALE LOCATION that are foregrounded but rather some other, apparently more peripheral concepts such as the status of an eventuality as completed or in progress, the attitude (necessity, obligation, wish), or the kind and strength of evidence. All these kinds of information are directly related to temporal reference but only when we think of timeE in the ‘metaindexical’ way, as inextricably linked to, and subsumed under, the SELF, with its attitudes of wish, want, obligation, its desire to express the degree of commitment based on the source of evidence, and so forth. These categories are modal-epistemic in nature. Hausa provides an example of this metaindexicality in the form of what Mucha (2013: 392) calls a ‘hierarchy of simplicity’. First, present-time reading of a tenseless construction is preferred over the past because it comes with no displacement. Next, the past-time reading takes precedence over the future because, while they both involve temporal displacement, the latter also comes with modal displacement. Then, explicit marking of temporal reference can override such defaults. Now, the question to ask is whether this variation of means suggests differences in the conceptualization of timeE across cultures. Arguably, it does not. The diversity of means does not have to lead to the conclusion of linguistic relativity. Rather, it is explicable in terms of what Levinson (2003) calls neoWhorfianism while discussing the relativity of spatial reference. While, on the ³² ‘IMM’ stands for the aspect-mood marker ‘immediate’; ‘1SG’ and ‘3SG’ for first and third person singular; ‘INC’ for the marker of incompletive status; and ‘DEF’ for definiteness marker. ³³ On a more in-depth analysis of the evidence from tenseless languages for the complex but universal concept of time see Jaszczolt (2009, 2020). For more examples of timeL in different languages that suggest that temporality is founded on epistemic modality see Jaszczolt (2009).

266

KASIA M. JASZCZOLT

surface, spatial (and for us here, temporal) reference displays significant crosslinguistic variation, this variation only affects complex, molecular concepts. On the level of their building blocks, we can still discern universals. And we do. Metaindexicality, the centredness on the ego and its attitudes to eventualities signal that the conceptualization of time follows general universal principles. The main principle is that timeE revolves around epistemic modality: when we express propositions in natural language discourse, we are normally committed to the truth of these propositions. But we are committed to a variable degree, signalled by the lexical, grammatical, intonational, or other means. Markers of epistemic modality such as ‘possibly, ‘might’, and so forth, as well as markers of evidentiality that signal, for example, whether the event was observed or inferred, are examples of such variations in commitment. In Jaszczolt 2009, I proposed that TIMEE rests on building blocks that mark such degrees of commitment—the degrees to which we are prepared to endorse statements about the past, the future, and the now. This is the proposal that I pursue a little further below, in Section 10.5.2. In Default Semantics, the building blocks are derived from the concept of epistemic modality (MODALITYE , the degree to which a speaker is committed to a proposition).³⁴ They testify to the universality of core human beliefs about time.³⁵ On the other hand, the ‘molecules’, such as language-specific temporal constructions with the relevant temporal adverbials, grammatical tenses, or language-specific pragmatic inferences, display ample cross-linguistic variation. I can now fit this semantic account into the broader picture of timeE endorsed earlier in this chapter, where I subscribed to the idea that timeE can be seen as an emergent property of real, static spacetime and as such of timeM (Section 10.3.1). Analysing the linguistic representations of relevant concepts, I claim that the complex TIMEE itself appears as an emergent property on the higher, molecular level of human concepts. In a slogan, timeM does not flow, and neither does timeE , when looked at through the conceptual building blocks of TIMEE . TimeE does not flow on the level of such building blocks: these building blocks are just concepts standing for the agent’s epistemic commitment to states and events. But it flows on the level of their molecular, language-specific combinations. So, we have universal, species-specific thoughts and their culture- and language-specific expressions.

³⁴ For analyses of what I now call timeE using temporalL reference see e.g. Jaszczolt (2009, 2013, 2016b, 2017, 2018a, 2020). However, none of these address the question of timeE flow that is discussed here. ³⁵ Remember that in virtue of our earlier metarepresentational analysis, this means ‘beliefs about timeE/M ’.

DOES HUMAN TIME RE ALLY FLOW?

267

A fortiori, we have universal, species-specific timeE (and TIMEE ) on the one hand, and culture- and language-specific expressions of it on the other. It has been extensively documented that a language can facilitate certain thoughts when a concept is lexicalized or grammaticalized there and as such provides a ‘shortcut’ in the form of a molecular concept. (Note that we follow Wierzbicka (e.g. 1996) in stipulating that universal, primitive concepts already have, by definition, their realizations in the lexicon or grammar of all natural languages.) A language then foregrounds certain distinctions in preference to others—for example, absolute time reference over evidentiality or relative time; exact numbers over the ‘not many/many’ distinction; material or shape of an object over function; manner of motion over trajectory; and so forth.³⁶ But this only provides a shortcut, not evidence of relativity: Whorfian relativity stops at the level of such language- or culture-specific combinations, without affecting the pool of their building blocks. In other words, languages employ lexical, grammatical, and pragmatic means to achieve communicational goals, and they all do it equally successfully, reflecting the universality of the human mental architecture—a topic to which I move in the following section. To take stock, so far I have (i) endorsed the views that (i.a) timeE emerges from timeM through indexing to the ego that I called metaindexicality (Section 10.3); (i.b) timeE is a metarepresentation of timeM , related to it via semi-propositional beliefs (Section 10.4); and, against this background, (ii) began the construal of the concept of dynamic timeE (TIMEE ) as reanalysable into static conceptual building blocks pertaining to MODALITYE (Section 10.5.1), using the argument from temporal reference in language (timeL ). What remains is to complete the picture by (iii.a) proposing a sketch of the semantics of temporal expressions that express TIMEE and (iii.b) speculating on the consequences of the proposal in which TIMEE turns out to be dynamic only on the level of the composition of basic concepts. I do so in Sections 10.5.2 and 10.5.3 respectively.

10.5.2 TIMEE as epistemic modality: Towards a contextualist semantics of timeL My representation of temporal reference in discourse relies on what I called elsewhere the ‘lexicon-grammar-pragmatics trade-offs’ (Jaszczolt 2012). Temporal reference can be conveyed by the lexicon, by grammatical distinctions, or left to pragmatic inference, with variations among languages as to ³⁶ See e.g. articles in Part I of Gumperz and Levinson (1996), and Levinson (2003).

268

KASIA M. JASZCZOLT

which of these vehicles (or their weighted combinations) are prioritized and employed. What is important is that languages are all equally well suited for representing timeE : what a language lacks in grammar, it can make up in the lexical inventory or in conventions used in inferring information. So, we can’t look into one linguistic category, such as grammatical tense or temporal adverbials, for information on TIMEE . We have to look at discourse at large, into its various forms and aspects. We can see there that speakers employ an array of forms to signal different degrees of commitment—for example, a past event can be described using the simple past, past of narration (‘vivid present’, grammatically present tense), as in (8), or modal forms such as must/may/might have + Past Participle. (8)

This is what happened to me last Monday: I sit in my office and work on my chapter and then this terrible noise starts. I look out of the window…

This is principally so because speakers can remember events more, or less, faithfully or possess different amounts of information. Analogous explanations pertain to the present and the future: the speaker can be more, or less, committed to what she is saying. First, events can be understood or remembered to different degrees; they can also be anticipated more, or less, strongly. Inference about events can be monotonic or non-monotonic, and as such be more, or less, trustworthy. To represent this modal nature (that I defended elsewhere as modal supervenience, see e.g. Jaszczolt 2009), I had adapted Grice’s (2001) Acceptability operator that he proposed in his quest for the underlying common core of different kinds of modality.³⁷ For alethic modality, Acceptability produces Acc ⊢p (‘It is acceptable that it is the case that p’). To adapt it to the framework of the contextualist theory of Default Semantics, I replaced Acc ⊢p with ACCΔ n ⊢Σ, where the symbol Σ stands for a truth-conditional representation that draws on all sources of information, both linguistic and non-linguistic, available in relevant discourse—to reflect the ‘lexicon-grammar-pragmatics trade-offs’ just mentioned. The replacement of p with Σ also reflects a commitment to a special kind of proposition: a proposition that pertains to the content obtained through such trade-offs and amounting to the primary communicated message rather than to the content of the sentence. I call it elsewhere a ‘functional proposition’ (Jaszczolt 2021). Next, I index ACC with Δ that symbolizes the degree of this acceptance of ³⁷ In this little known, unfinished and posthumously published work Aspects of Reason, Grice discusses practical and alethic modality. But the argumentation equally applies to epistemic and deontic modality.

DOES HUMAN TIME RE ALLY FLOW?

269

the proposition and as such stands for the degree of epistemic commitment. (9) gives the formula for the standard use of ‘regular past’ (‘rp’, such as English Simple Past). (9)

ACCΔ rp ⊢ Σ (‘It is acceptable to the degree Δ pertaining to the regular past that it is the case that Σ’)

In the case of a tense–time mismatch as in (10), the source of information with which the representations of the relevant chunks of the sentence are indexed in Σ (details are not relevant for the current purpose) signals how this mismatch is resolved. (10)

I go to Lisbon next Monday.

For example, it can be resolved in the process of conscious pragmatic inference (CPI), when it has to be inferred from the broadly understood context—the temporal adverbial in (10) or even inference from the conversation itself should the adverbial be left out. ACCΔ is then indexed with ‘tf ’ standing for tenseless future: ACCΔ tf ⊢ Σ.³⁸ We can now unpack the formula proposed earlier in (6) as in (11), substituting ‘ACCΔ n ⊢ Σ’ for ‘TIMEE ’. (6) (11)

(SOQ (OCQ (timeM )) = timeE )belief = TIMEE (SOQ (OCQ (timeM )) = timeE )belief = ACCΔ n ⊢ Σ

The semantics is contextualist and results in representing a functional proposition; the truth-conditional representation draws on information that goes beyond the logical form of the sentence, to capture, say, effects of linguistic conventions, such as using simple present tense for future-time reference when events are planned and as such fairly certain, or effects of pragmatic inference from current situational context. Needless to say, this contextualist construal departs significantly from the treatment of temporal reference in traditional formal semantics. But it is not incompatible with it; it builds on it and pushes it in the direction of linguistic pragmatics. Let us take here Partee’s (1973) proposal to represent grammatical tenses as variables rather than operators. This allows for temporal reference marked by the grammatical tense to be anchored to the speaker. As she says, in (12), the speaker is not communicating that there is no past time at which the speaker turned off the stove, or that there is a time at which she didn’t turn ³⁸ It is tangential to my current purpose to discuss the actual formal representations of Default Semantics (Σs). For an accessible exposition see Jaszczolt (2009).

270

KASIA M. JASZCZOLT

off the stove, but rather that during a certain interval that is relevant for this particular discourse, the speaker did not turn off the stove. (12)

I didn’t turn off the stove. (after Partee 1973: 602)

What I have done now is take the idea of indexing further in pointing out the relativization of the indexicality of time to the metaindexicality of the ego: the event is in the past because the speaker conceptualized it as such in virtue of being a ‘self-governing complex system’ that sees reality in this dynamic way.³⁹ To reiterate, we have to remember that linguistic time (timeL ) is not grammatical tense (tenseL ⁴⁰): temporal reference can be conveyed through temporal adverbials, grammatical tenses, or pragmatic inference from context and from the defaults or preferences that languages use to signal temporality, for instance when aspect, modality, or evidential markers are employed for the purpose. Even if, like Partee, we stipulate that the proposed semantics is specific to the English language, we have to account for tenseL -timeL mismatches, where, for example, a past eventuality is reported in the present tense as in (8) or a future eventuality is stated in the present tense as in (10). So, the solution is a contextualist semantics where tenseL is still indexical but not strictly in the sense of Partee: it is indexical as it contributes to the value assigned to the delta index in ACCΔ that stands for the degree of acceptability of a proposition, that is the degree to which the speaker is committed to the eventuality the statement is about. This value of epistemic acceptability, and as such epistemic modality, is what temporal reference is decomposed into on the level of conceptual building blocks.

10.5.3 From static timeM to (apparently) dynamic timeE : The denouement Now, there is nothing dynamic about the value of the delta index: it is just the degree of commitment of the agent to the eventuality (proposition). But the ‘self-governing complex system’ that performs this epistemic assessment ³⁹ It is ‘metaindexicality’, as introduced in Section 10.3, because the conceptual scoping is governed by a different principle than semantic scoping: ‘She didn’t turn off the stove’ does not anchor the temporal index to the pronoun ‘she’ but to the speaker. Also, the timeline as such can be constructed with reference to different subjects, as anaphoric dependencies demonstrate. Cf. ‘She turned off the stove and went for a walk’ vs ‘She turned off the stove and the stew stopped simmering.’ ⁴⁰ TenseL is to be distinguished from tenseM (on A-theory, tensism, reality is tensed) and tenseE (our thoughts are tensed).

DOES HUMAN TIME RE ALLY FLOW?

271

bestows it with dynamicity through constructing timeE as flowing (or, on some accounts, by experiencing the flow of timeE ).⁴¹, ⁴² And it is precisely because dynamic timeE is metaindexed (and metarepresented) static timeM that TIMEE emerges as dynamic from its static building blocks. Put simply, humans create dynamic representations in virtue of the inescapability of the ego perspective and in virtue of their metarepresentational abilities. My modal reductionism does not entail a clear stance on the phenomenology of the passage because it admits it only on the ‘molecular’ level—that of complex concepts and their externalization in linguistic constructions. My account of the PASTE , PRESENTE , and FUTUREE does not allow for the flow of timeE at the level of basic concepts. But it is a form of reductionism in which the passage is not an illusion. It is present in the emergent ego-perspective through metaindexicality, even if only on the molecular tier. As was argued, since we have subscribed to a view where a temporal property can emerge on a higher level, modal supervenience fits the bill as far as the compatibility of TIMEE (within timeE ) with static timeM is concerned.

10.5.4 Loose ends and further directions for semantics and philosophy of timeM/E/L There are concepts associated with the flow of timeE that still require attention in that they appear on the higher, molecular level regardless of the proposed reduction to epistemic modality. One of them is the direction of the flow. And this direction can differ. In some cultures, the past is conceptualized as being ‘in front’ of the observer. In Maori, ‘ngārā mua’ means ‘the days in front’ and refers to the past (Thornton 1987); in Aymara spoken in the Andean region in South America, the word for ‘past’ comes from the word ‘nayra’, meaning ‘eye’ or ‘front’ (Evans 2003; Moore 2011); Ancient Greek used the stem ‘opi-’, meaning ‘behind’, for the future (Morpurgo Davies 1983) and we know that the experiencer ‘faced the past’. This shows that the relative conceptualization of timeE differs between cultures:⁴³ the experiencer faces the future, as the unknown, or alternatively faces the past, the known. This may suggest one ⁴¹ Following my disclaimer in Section 10.2.1, I will have nothing to say here about the latter. ⁴² As was remarked earlier, I am not engaging in this chapter with the discussion as to whether reality could be dynamic: I am assuming, espousing common wisdom in modern physics, that it is not. See Baron (2017) on a good defence of the dissociation between timeM and timeE even where both are dynamic (Baron is a defender of A-theory). The dissociation is even easier to construct with the assumption of static universe, so emergentism is not in trouble here. ⁴³ That is, relative to the orientation of the speaker rather than relative to another eventuality as in relative tenses.

272

KASIA M. JASZCZOLT

of two explanations: either (i) that human conceptualization of timeE varies between absolute and relative, analogous to that of space (e.g. west/east vs. left/right; see Levinson 2003), where the absolute corresponds to the commensurate direction of the arrow of time and human ‘sight’, or (ii) that both types of conceptualization are relative, one focusing on the known and the other on the unknown. But, more importantly, the differences in the direction of facing may also provide an additional argument for the modal supervenience view: if we face the past, we face what is better known and that for which we have more empirical support. Alternatively, if we face the future, we put behind what is already experienced and move on to face challenges. Either way, TIMEE is interwoven with (gradable) epistemic commitment. Next, the very organization of timeE into a timeline is also culture-specific. Not only do cultures differ in adopting the right-to-left or left-to-right conceptualization of the flow, or horizontal, vertical, or even circular (where the cyclic occurrence of events is foregrounded), but the very importance of a timeline also differs from language to language. Languages appear to conform to the Husserlian concepts of retention and protention but further evidence is needed to assess how the events are ordered in memory or anticipation. The fact that they are ‘behind’, ‘in front’, ‘above’, or ‘below’ tells us too little about the ordering; the ordering can follow the principle of earlier/later, but it can also follow the principles of or more/less salient, more/less credible, and so forth. In the latter cases, modality qua remoteness on the atomic level of concepts can give us remoteness in timeE on the molecular level.⁴⁴ The question now arises whether this molecular/complex level of conceptualization requires the arrow of time. It seems that now the flow can be explained as degrees of epistemic detachment from certainty tout court, analogous to the temporal reference itself: experiences are classified along the cline of acceptability. So, the cline has a modal, quantitative interpretation. Correlated degrees of commitment then give us the cline of perceived metaphysical probability, from, say, somewhere near zero to one. When we juxtapose the metaphysical probability with the theoretical construct of possible worlds, time reduces to the accessibility of worlds.⁴⁵ What is represented as a flow of timeE may in fact be the gradation of epistemic commitment, translatable

⁴⁴ ‘Can’ in that supervenience does not come with bi-uniqueness. Nota bene, Ludlow (1999) derives both remoteness in time and remoteness in possibility from the idea of evidentiality. But this would be using a linguistic concept that has a limited distribution and constructing an analogous epistemic concept of an equally limited occurrence. ⁴⁵ See also Parsons (2003) on relativizing truth to worlds indexed with time-at-this-world.

DOES HUMAN TIME RE ALLY FLOW?

273

into degrees of metaphysical probability that produces on the level of complex concepts asymmetric TIMEE , PASTE , PRESENTE , and FUTUREE —in virtue of the properties of the expanding universe in which life can exist (Hawking 1988; Price 1996) or, perhaps, just in virtue of the emergent order that humans impose on the world to comprehend it (Ismael 2016; Callender 2017)—or both. Put differently, as was said, the proposal boils down to a model on which FLOW OF TIMEE comes from the arrangement of the building blocks of TIMEE along a cline of epistemic modality. Now, if dynamism comes from degrees of modality, there has to be a ‘converter’, so to speak, that takes us from the basic level without the flow to the complex level with the flow. The passage and the vagaries of the rate of the felt passage require the ‘operator’ that converts the primitive modal concepts into the temporal complex concept, itself in the scope of an operator that allows this dynamic time to be skewed. In Jaszczolt (2013) I gave this idea a label: a Modal-Contextualist view, but did not have a fully developed proposal. Now I have come closer, by proposing the nested operators in (3), modelled on Torrengo’s phenomenal modifier but used for a somewhat different purpose. While his modifier accounted for flowing time in experience that itself does not represent flowing time, my qualifiers convert static timeM into apparently dynamic TIMEE —that is into the concept of timeE in beliefs about timeE , now demonstrated to be the concept of timeM/E in beliefs about timeM/E when metaindexicality, metarepresentation and concept decomposition are all taken into account. The qualifiers in (3) are the ‘converter’ we need. This, naturally, opens up myriad further questions and myriad further speculative hypotheses that will have to wait for future research in philosophy, semantics, and language documentation. In particular, conceptual semantics of temporal reference associated with (3) that I sketched in (11) is still in its relatively early stages. It is a radically contextualist outlook according to which pragmatic clues, such as contextually salient or conventionally conveyed information, are considered to be valid indicators of temporal reference for the purpose of semantic representation— the latter conceived of as in Default Semantics, as truth-conditional representation but construed on the level of conceptual, not linguistic, structures.⁴⁶ On such a construal, unlike in the case of semantic representations that simply model grammatical tenses, the mapping from the uttered sentence to the representation is complex indeed. To reiterate, the value of delta in ACCΔ , that ⁴⁶ For arguments in favour of conducting analyses of meaning on the level of conceptual structures rather than (natural-language-) syntactic structures see Evans and Levinson (2009); Everett (2008, 2012a); or Jaszczolt (2016a).

274

KASIA M. JASZCZOLT

is the degree of epistemic commitment assigned to the proposition, draws on many different sources of information in the context of a relevant discourse. In order to be able to assign to real utterances real semantic values for delta, one would have to have extensive empirical data—either from a corpus of conversation or possibly from neuroimaging. We are far off this stage in corpus-based pragmatics and neuropragmatics, and I am not even certain how important it is to assign such ‘real values’. The testing of proposals such as the one developed here may be much simpler through other avenues of research such as an inquiry into the formation and behaviour of complex systems vis-à-vis basic, symmetrical, transcendental, ‘naked’, and still largely poorly understood reality. Time will show (no pun on static time intended). Metaphysics, psychology, and linguistics cannot successfully anchor to the unknown.

10.6 Conclusion ‘Time emerges from a world without time’ (Rovelli 2018: 117)

The main objective of my inquiry was to attempt to reconcile the B-theoretic, static conception of timeM of multi-dimensional spacetime with dynamic, flowing timeE that permeates, in one way or another, our representations of reality and their linguistic expressions (timeL ). It was my methodological choice to focus on beliefs and associated concepts in order to shed more light on the object they pertain to. First, I endorsed the idea of the indexing of the flow to the emergent ego perspective, calling it metaindexicality. Next, I added the suggestion that the timeE be analysed as a metarepresentation of timeM —a semi-propositional, and as such not fully understood, representation that at the same time is faithful to the theory of static universe. Having asked why and how, in virtue of what, timeE is dynamic, I moved to the question in what sense it is dynamic, asking about the properties of the concept itself: TIMEE . Using arguments from its externalization as timeL in natural languages, I further developed the view, originally proposed within the theory of Default Semantics, that temporal concepts are complex concepts that are built from conceptual building blocks standing for degrees of epistemic modality. I argued that when combined with the idea of metarepresentation, metaindexicality, and my two operators loosely modelled on Torrengo’s (2017a) phenomenal modifier, we arrive at a concept of the flow of timeE that explains our temporal beliefs.

DOES HUMAN TIME RE ALLY FLOW?

275

In short, there is a level of human conceptualization of reality on which timeE flows. This flow is explained through a combination of interrelated explanantia: (i) the emergent, metaindexical, ego-perspective on timeM through timeE ; (ii) the metarepresentational nature of the belief about time— timeE , but also, through metarepresentation, timeM ; as well as (iii) the fact that TIMEE is itself a complex concept. This concept arises together with the concept of TIMEE FLOW only on the level of language- and culture-specific combinations of primitive concepts that pertain to degrees of epistemic modality, and as such are static, like timeM . The proposal in (iii) is ‘neo-Whorfian’ in the sense in which Levinson (2003) tentatively adopted this term while analysing the conceptualization of space across languages and cultures: on the level of the basic concepts, there is universalism. The jury is still out, but there is growing support for the view that basic concepts from which thoughts are built do not differ from language to language or from culture to culture. However, on the level of their combination into complex concepts, linguistic and conceptual relativity arise. I understand this outcome as meaning that static timeE , the basic-level human time qua modality, is compatible with the timeM of spacetime in the physicists’ description of reality—there are no mismatches to explain. Mismatches arise on the level of complex concepts. My analysis offers a hypothesis, corroborated with some linguistic evidence as well as an argument I constructed here, of what exactly this human, dynamic timeE is, a sketch of how it can be represented in conceptual semantics, and how it dispels the myth that flow is so elusive. Finally, arguing ‘backwards’ to my assumptions, the argument for the underlyingly static timeE can be taken as partial argument for static universe in the ongoing debates between the A- and B-theorists in the metaphysics of time. But, arguably, these days this is no longer needed!

11 Underspecified temporal semantics in Pirahã Compositional transparency and semiotic inference Daniel L. Everett

‘Perhaps it might be said rightly that there are three times: a time present of things past; a time present of things present; and a time present of things future. For these three co-exist somehow in the soul, for otherwise I could not see them. The time present of things past is memory; the time present of things present is direct experience; the time present of things future is expectation.’ St Augustine ([1], Book 11, Chapter 20, Heading 26)

11.1 Introduction 11.1.1 Summary of major points In what follows I review, revise, and move beyond the analysis of temporal semantics and grammar in Pirahã first developed in Everett (1993).¹, ², ³ This revised analysis reaches several conclusions. First, as argued by Everett (1993), the theory of tense developed in Reichenbach, if adopted at all, must allow for languages which lack what Reichenbach (see also Hornstein (1990)) called the ¹ I would like to thank Kasia Jaszczolt for her invitation for me to participate in the workshop Understanding Human Time and to all the participants for their individual research and for the many who commented on what follows. The Zoom format had the advantages of keeping the workshop green and the discussions up close and personal. ² This chapter borrows and paraphrases frequently from Everett 1993. Abbreviations used in this chapter are: 1, 2, 3 indicate ‘first person,’ ‘second person,’ and ‘third person’ in the glosses of Pirahã examples; CONT. ‘continuative’, COMIT. ‘comitative’, COMP.CERT. ‘complete certainty’, DES. ‘desiderative’, FRUST.INI. ‘frustrated initiation’, IMP. PRO. ‘imperative pronoun’, IT. ‘iterative’, LOC. ‘locative’, REL. CERT. ‘relative certainty aspect’, REM. ‘remote: out of temporal control of speaker’. ³ Pirahã is a language of the Brazilian Amazon, the only surviving member of the Mura language family (Loukotka 1968).

Daniel L. Everett, Underspecified temporal semantics in Pirahã. In: Understanding Human Time. Edited by Kasia M. Jaszczolt, Oxford University Press. © Daniel L. Everett (2023). DOI: 10.1093/oso/9780192896445.003.0011

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

277

‘R(eference) point’. Because Pirahã lacks perfect tenses and perfect interpretations, I argue that the R-point is unnecessary in Pirahã. Second, I argue that Pirahã’s temporal structures require a theory of compositionality that, reminiscent of Partee (1995), Jaszczolt (2005, 2009), among others, fills in gaps of structures and meanings that are missing from the syntax, not merely via the pragmatics or implicatures, but by the general process of semiotic inference (see below). Although there are many theories of compositionality that address similar issues, I will argue that it is more parsimonious and accurate to treat compositionality as a subtype of inferential processes ranging at least over culture, grammar, and real-world knowledge. This better fits a particular approach to meaning, begun by Charles Sanders Peirce more than 150 years ago, influential across the natural sciences. I also argue that by treating compositionality as a subspecies of inference, we are better able to understand a number of otherwise puzzling interpretations, including the recursive interpretations in Pirahã in the absence of recursive syntax, Pirahã tense, and interpretations in other languages. Finally, I argue that the inferential determination of Pirahã temporal semantics, even though underdetermined by the syntax in the language, allows better understanding of Everett’s (2017) and Barham and Everett’s (2020) conclusions that there are three types of language based on the tightness of fit between syntax and semantics. If correct, this conclusion illustrates a possibility mentioned by Barbara Partee regarding syntactic underdetermination of quantification (based on a tripartite semantic structure for quantification found in several approaches): If we were to regard such a structure as a common semantic interpretation of the very different syntactic structures in question, we might find ourselves asserting that some languages are more compositional than others. Partee (1995: 541)

Though Partee rejects this possibility, I argue here that it is necessary and natural.

11.1.2 Organization of the chapter Section 11.2 of this chapter provides a summary of Pirahã time words and temporal interpretations as discussed in Everett (1993), as well as motives for rejecting the universality of Reichenbach’s full temporal theory (even as modified by Hornstein (1990)). Section 11.3 discusses getting from Pirahã temporal syntax to Pirahã temporal semantics. Section 11.4 discusses the implications

278

DANIEL L. EVERETT

of our findings for compositionality. Section 11.4 also discusses the possible fit, alluded to by Partee (1995: 541) above, between grammars and inference/compositionality, supporting Everett (2017) and Barham and Everett (2020). Section 11.5 summarizes the findings and discusses some additional implications.

11.2 Pirahã temporal signs and interpretants Although there is a rich system of verbal aspectual suffixation in Pirahã, there is no purely temporal morphology (Everett 1983, 1986, 1993). There are also a very limited number of time words. Thus Pirahã temporal semantics is underdetermined by lexico-syntactic sources. And yet Pirahãs are aware of past, present, and future, even though they do not mark them overtly in the language. This mismatch in Pirahã of the absence of overt marking of traditional tenses with the presence of temporal interpretations corresponding in part to such tenses leads ultimately to a better understanding of the role of inference, culture, and real-world knowledge in linguistic interpretation.⁴ Of course, the role of culture and other information in temporal interpretations is not unique to Pirahã (e.g. Jaszczolt 2009, Everett 2012a, b, 2016, inter alia). The facts of Pirahã temporal meanings remind us that the study of littleknown languages can pay back the effort expended by discoveries of principles or better examples of principles than on better-known languages might reveal.⁵, ⁶ From the perspective of Everett (2012b, 2016) all temporal systems begin with real-world physical and cultural experiences (also cognitive, but this term, according to Everett (2016) is but a hyponym of culture—cf. Sapir 1921, 1931, inter alia). One need not adopt a particular philosophical position on time and space to recognize that bodies require space and thoughts require time (cf. Prosser 2016). Of course, we do not know this when we are born. We just live. Our interpretations of our proprioceptions and perceptions are constrained by the ⁴ Note that my claim is not that I am the first to notice the long-known fact that real-world knowledge influences semantic and pragmatic interpretations. Rather I use this observation to urge a different conception of the nature of this influence, as inference rather than as purely a linguistic principle. ⁵ Pirahã is a ‘tenseless language’ in the sense of Comrie (1985: 50ff ): languages ‘where time reference per se is not grammaticalized’, i.e. where there is time reference but it is not reflected morphosyntactically. ⁶ The reader is urged to consult more cross-linguistic work on temporal semantics of less-studied languages, e.g. Dr Lydia Rodriguez on temporal representation in the Mayan language Chol (Rodriguez 2019), Simeon Floyd (2016) on Nheengatu (Brazilan creole language), and my own work (Everett and Kern 1995) on Wari’.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

279

cultures from which we emerge, which shape our perspectives on the meaning and structure of time and space and the way in which these perspectives in turn affect our social expectations, knowledge structures, and values (Everett 2016). As we say in Portuguese when we want someone to arrive on time: ‘Hora britˆanica, ouviu?’ (‘British time, you hear?’) According to this Brazilian Portuguese saying, the British valuation and meaning of specific time interpretations guide one’s behaviour differently from those in Brazilian Portuguese. (Though the expression is ultimately intended to be humorous, it is frequently used to clarify beginning times of meetings.) My proprioceptive awareness of my body’s extension tells me something about the space I occupy. My eyes show differential distances. My awareness of time is in part formed by external factors, such as the sun’s rays’ intensity changing based on the time of day and the time of year. Our hunger tells us it is time to eat; our thirst that it is time to drink. At a party the host ‘realizes’ it’s time for guests to leave, though guests might not share this realization— vantage points are crucial. Events and our lives feel sequential, especially our thoughts as Peirce pointed out 153 years ago: To say, therefore, that thought cannot happen in an instant, but requires a time, is but another way of saying that every thought must be interpreted in another, or that all thought is in signs. Peirce (1868: 207)

This sequentiality becomes a basis for time and the time-line metaphor (left to right = past to future) found in many languages, just as it can serve as a foundation for other time systems, such as the temporal semantics in Pirahã, wherein culture arguably plays a stronger role than the morphosyntax in time reference. Culture permeates our lives, as I have argued in many other places (e.g. Everett 2005, 2012, 2016), in ways that are often unspoken and even ineffable. One of the first to write about this with clarity and elegance was Edward T. Hall (1973 (1959), 1976, 1990). In 1959, Hall published The Silent Language: It wasn’t just that people ‘talk’ to each other without the use of words, but that there is an entire universe of behavior that is unexplored, unexamined, and very much taken for granted. It functions outside conscious awareness … What is most difficult to accept is the fact that our own cultural patterns are literally unique and therefore not universal. Hall (1973 (1959): vii)

280

DANIEL L. EVERETT

Hall is not merely offering here the banal observation that there is an unconscious. He is talking about cultural tacit knowledge—a more articulated notion. The concept of tacit knowledge (Polanyi 1974, 2009 [1966]) is one that I have expanded into the wider idea of ‘dark matter of the mind’, which I define as follows: Dark matter of the mind is any knowledge-how or any knowledge-that that is unspoken in normal circumstances, usually unarticulated even to ourselves. It may be, but is not necessarily, ineffable. It emerges from acting, ‘languaging,’ and ‘culturing’ as we learn conventions and knowledge organization, and adopt value properties and orderings. It is shared and it is personal. It comes via emicization, apperceptions, and memory, and thereby produces our sense of ‘self ’. Everett (2016: 11)

The concept of dark matter includes our understanding of time, individually and culturally, and, as the definition claims, is acquired gradually over time as we grow up in a particular culture. For example, some cultures offer a distinct application of their concepts of time for their gods. From some gods’ perspective, the past, present, and future present themselves to the deity simultaneously. For such a being there would be no past, present, or future—He or She created all times at once and sits outside of time (e.g. Psalms 90:2 ‘Before the mountains were born or you brought forth the whole world, from everlasting to everlasting you are God’; or, from Ephesians 1:4, ‘For he chose us in him, before the foundation of the world, to be holy and blameless in love before him’). Similarly Plato, a ‘deity’ of the Western secular canon, according to Reichenbach, viewed time as a ‘moving image of eternity’, where ‘eternity’ is a ‘reality not controlled by time flow’ (Reichenbach 1956: 5). One topic that has long intrigued anthropological linguists is the relationship between humanity’s various senses of time and how these cultural manifestations are represented in temporally based philosophies, sciences, theologies, and grammars. How do human cultures effect the linguistic expression of time? In Peircean terms, the process of (i) sensing, (ii) tracking, and eventually (iii) representing the movement of the phenomenological/phaneroscopic perception of time from a vague feeling (e.g. the warmth of sunlight on our face that we do not focus on consciously) to a recognition of temporal stages of the daytime (say, morning and noon), to words, morphology, and reasoning about time is the result of his three phaneroscopic (phenomenological) characteristics. Peirce labels these three components of experience ‘firstness’,

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

281

‘secondness’, and ‘thirdness’. The thirdness in particular is culturally and linguistically constrained and is the focus of this chapter.

11.3 A review of Pirahã time 11.3.1 Absolute tenses In Everett (1983, 1986, 1993, 2005) it is claimed that Pirahã lacks any dedicated morphosyntactic markers for past, future, or present. However, as presented in Everett (1986: 288ff ), Pirahã has a rich set of aspectual suffixes to mark such notions as duration of action, realization of action, internal division of action, continuation of action, beginning of action or state, actions within and outside of the control of the speaker (which overlaps with the notion of proximate vs remote tense), iteration of action, and resultative aspect. There are also two separate positional classes with three suffixes, each to indicate the relative certainty of the speaker with regard to the action being asserted (complete certainty, relative certainty, uncertainty) and the speaker’s source of evidence, e.g. deduction, induction/observation/abduction, or hearsay, all of which systems can have temporal implications. These represent distinct semiotic subsystems and the signs they are composed of work in tandem with discourse, culture, and real-world knowledge to source temporal interpretations for Pirahã utterances. 11.3.1.1 Time words There are no time words in Pirahã which unambiguously mark future or past times. In English, such time words and expressions include yesterday (moment of speech preceded by event) tomorrow (event follows moment of speech), next year, etc. The final interpretation of such time words and expressions as there are in Pirahã is crucially dependent on context. A partial list of Pirahã time words and expressions is given below:⁷ PAST/PRESENT (1) a. so ʔóá ‘already’ time -used/already experienced (This is a diachronic formation, no longer seen as productive in the language) ⁷ More data is available from the MIT corpus of Pirahã texts: https://osf.io/kt2e8/

282

DANIEL L. EVERETT

b. tı´i soʔóá kaháp 1 already go ‘I’m going now.’

-i -PROXIMATE

-ı´ -ACTION

c. tı´i soʔóá kaháp -á há 1 already go -REMOTE -COMP.CERT. ‘I already left’ PAST (2) a. so ʔógió time big (Also, like the examples to follow, a diachronic formation no longer productive in the language) ‘long time ago/from now’ b. ʔaoói soʔógió koa - á -i foreigner long ago die -REMOTE -ACTION ‘The foreigner died a long time ago’ c. ʔaoói soʔógió ʔowá -boı´ -haı´ foreigner long ago purchase -MOVEMENT -REL.CERT. baósai cloth ‘The foreigner will buy/bought cloth some time from now’ FUTURE/PAST (3) a. pii ai -so water thin -TELIC ‘The water becomes thin’ (Freer –‘Summer/dry season is approaching’) b. ʔoogiáihi piiáiso ʔab -o -óp Dan 3 dry season remain -STEM VOWEL-go a -i baisı´ hi REMOTE -REL.CERT. Maici:river 3

ai be

‘ʔoogiái (Dan) will return in the summer (=low water time)’ OR ‘Dan returned in the summer.’ (4) a. pii bı´gaı´ -so water thicken -TELIC ‘when water is thickened’ (winter/rainy season) b. ʔoogiái Dan

hi piibigaiso ʔabo -óp 3 rainy season remain -go

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

-a -i. baisı´, hi ai REMOTE -REL.CERT. Maici:river 3 be ‘Dan will return in the rainy season’ ‘Dan returned in the rainy season’ (Separate clauses, with separate intonational breaks) (5)

a. tı´ihı´ káobı´ -so Brazil nut fall -TELIC ‘Brazil nuts fallen’ (used for latter part of rainy season) b. ʔao aáı´bá -koı´ tı´ihı´káobı´so foreigner many -intense latter period of rainy season ‘There are many foreigners (here) in Brazil nut season’

(6)

a. ʔa ho ái at fire be ‘night’ b. ʔoı´ kab áo ʔab i ´ı biosphere not move REMAIN STEM VOWEL PROXIMATE -haı´ ʔahoái -REL.CERT. night ‘Don’t go out at night’

(7)

a. hoa hoı´ hi -o fire couple at -it ‘more than one day b. Hi hoahoı´hio ʔab iig á 3 couple of days remain -CONT. -REM. ‘He is staying several days.’

Other examples of time words include words like: (8)

ʔahoa koho -ai -hio night eat -do -in ‘early morning’

(9)

hi bigı´ 3 ground/sky ‘sunrise/sunset’

(10)

bagá ʔais -o on top of do into

hiso ógi ái sun big do/be ‘noon’ (lit: ‘big sun time’)

283

284

(11)

DANIEL L. EVERETT

pi ʔái water -do/be ‘now’ (lit: ‘this water’)

11.3.1.2 Sequential markers Sequential markers or ordering words are formed metaphorically in Pirahã (as in many other languages), including such expressions as: (12)

a. ʔapaı´ ‘first’ (literally: ‘head’) b. Gói ʔapai ʔopı´ -ta -á -ti imp.pro head go -IT. -REMOTE -DIRECTIVE ‘You go first.’ OR ‘You go in front.’

(13).

a. ti ohi -ó butt side -LOC. ‘next’ (lit: ‘at the butt’) b. Kóhoi ʔapaı´ kao -p -á Poioı´ hi Kóhoi head emerge -VERTICALLY -REMOTE Poioı´ 3 tioh -ó kao -p -á -há butt -at emerge -VERTICALLY -REMOTE -COMP.CERT. ‘Kóhoi was born first, then Poioi’

(14)

a. gaaba ‘then/next’ b. Kóhoi ʔapaı´ kaopá. Poioı´ hi tiohó kaopáhá ‘Kohoi was born first, then Poioi’ c. ti koho -ái -kab -á -ob -áo. 1 eat -do -end -REMOTE -TURN -TELIC. gaaba ti gı´ kobai -sog -abagaı´ Next 1 2 see -des. -FRUST.INI. ‘When I finish eating, then I (almost) want to see you’

Each of the temporal expressions above can be made precise only via specific contextual inferences. Note, though, that in context each of these expressions is ultimately interpreted by simply interpreting the moment of speech as prior to the future event, following the event, or simultaneous present with the event.⁸ ⁸ That is, as per Everett (1993), in a Reichenbachian system these can all be generated by orderings of the coordinates E ‘event time’ and S ‘moment of speech’.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

285

11.3.1.3 Other temporal representations Temporal clauses In Everett (1983/1986), I analysed a particular suffix, -ao/-so (vowel-initial following consonants, s-initial otherwise) as indicating telic aspect. (15)

kab -á -o -b -áo finish -REMOTE -away -down -TELIC ‘I just finished.’ OR ‘I am about to finish.’

soxóá already

The -áo is not a temporal connective/subordinator but rather it indicates ‘boundary of an action’. Moreover, it can be omitted from what one might consider to be similar to syntactically subordinate clauses in other languages, as in (16) and (17):⁹ (16)

a. kohóai -xiigá. Tı´i gı´ ʔahoái -soog -abagaı´ eat -CONT. 1 2 talk -want -FRUST.INI. ‘(you) are eating. I want to talk to you’ (Free translation: ‘When you finish eating, I want to talk to you’) b. kohóai eat

-kabá -o -finish -STEM VOWEL

-b -áo -DOWN -TELIC.

ti gı´ ʔahoái -soog -abagaı´ 1 2 talk -want -FRUST.INI. ‘(You) are about finished eating/you just finished eating. I want to talk to you’ (Free translation = ‘When you are done eating I want to talk to you’¹⁰ (17)

a. pii boı´ -so. ti kahápiı´ -hiaba rain move -TELIC. 1p go -NEGATIVE ‘Rain complete. I won’t go’ (This is because when the ground is wet, the snakes come out of their holes.) b. pii -boı´ -iigá. rain -move down CONT. ‘It (is) raining. I won’t go’

ti kahápiı´ -hiaba 1 go -NEGATIVE

⁹ As we see, a sentence may be semantically hypotactic to another without being syntactically subordinate. ¹⁰ See also Caudal, this volume, on frustrative markers.

286

DANIEL L. EVERETT

If the suffix –sai (‘old information’) is used in these examples instead of –so/-ao the result is different: (18)

a. pii -boı´ -sai. Ti kahápiı´ -hiaba rain -move down -OLD INFORMATION. 1 go -NEGATIVE ‘It rains (topic of discussion). I won’t go.’

This can also be interpreted as a conditional (e.g. ‘If it rains, I won’t go’), but need not be. This arises just in case the conversation is about raining and one adds information about their plans. But there is no evidence for subordination apart from the English translation. Additional examples Past readings: -iig -á (19) ti kahi ob -áo -b 1 arrow see -TELIC -DOWN/PERFECTIVE -CONT. -REMOTE ‘I was looking at the arrow’ (20)

xi hiab -ab -óxóı´ animal NEGATIVE. -REMAIN -INTERROGATIVE ‘ Did the meat run out?’

(21)

ti xi koho -áo -p -iig -á 1 animal eat -TELIC -UP-CONT. -REMOTE ‘I was eating meat’

Future readings (22) hi koab -ai 3 die -ATELIC ‘ ’ He will die (23)

-p -IMPERFECTIVE

-hix -INTERROGATIVE

-a -REMOTE

ʔipóihiı´ ʔab -óp -ai -so. ʔaaxái woman return -go -be -TELIC PROPER NAME ʔipóihiı´. tı´i ʔahá -p -ı´ -t -aó woman 1p go -VERT/IMP. -PROX. -IT. -TELIC. ‘he woman returned I am gone again’ Free translation: ‘When the woman returns, the wife of ʔaaxái, I will leave again.’

Present reading (24) kahaibó bogı´aga -hoag -á arrow point warp come/turn -REMOTE -há -taı´o -COMP.CERT. RESULTATIVE ‘Therefore, the arrow head does not warp’

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

(25)

kó Kohoibiı´ihiai, ti gı´ ʔahoa -ai VOCATIVE PROPER NAME 1 2 speak -be abagaı´ FRUST:INIT ‘Hey, Kohoibiı´ihiai, I want to talk to you’

287

-sóog -DES. –

Now that we have a basic concept of absolute tense expressions in Pirahã let‘s consider more complex cases, beginning with English complex sentences and proceeding to ask why perfect tenses are not found in Pirahã, beginning with a brief introduction to Reichenbach’s theory.

11.3.2 Reichenbach’s temporal theory¹¹ There are clearly logical and real-world constraints on tense, so that, for example, we cannot have someone in the past being affected by events in the future: (26)

∗

John ate when he will get his meal.

Tense is constrained by logic and the real world. But if the properties of tense grammar and interpretation in natural languages were all logically necessary, then there would be no language-specific variation in the tense system and thus little need for crosslinguistic studies in this area. For purposes of discussion, Reichenbach’s (1947) system will be used, though I propose a nontrivial modification of that system (cf. Everett 1993). Reichenbach (1947: 288ff ) argues that tense relies on three distinct temporal points: the event time, E, the speech time, S, and the reference time, R. These are all illustrated in the English perfect tenses: (27)

John will have finished when Bill arrives. S1 ___E1 ,___R1 S2 ___E2 ,R2 (where R1 __E2 )

In (27), the event of John‘s finishing precedes that of Bill‘s arriving, and both are subsequent to the moment of speech. So, in both the main clause and the adjunct clause of (27) the Moment of Speech (S) precedes the Events (E): S___E. The perfect tense reading comes from the R points. In the matrix clause ¹¹ This section borrows heavily from Everett (1993), repeating some of what was given there and rewriting those portions where my opinion/analysis has changed. Mood or modality refers to the relation of an utterance to reality. Indicative mood is associated with reality. Desiderative mood is associated with desire or a wish that a currently a non-real situation come about. ‘Remote’ in my usage indicates a situation in which one or more interlocutors lack control of all or aspects of what is happening. This is the case for either the future or the past. ‘Proximate’ in my terminology indicates a time or situation in which the interlocutor has control. I have used these terms since the early 1980s and they should not be confused with ‘remote’ and ‘proximate’ as simple tense divisions.

288

DANIEL L. EVERETT

the event of finishing precedes the event of Bill arriving. To achieve this result, we say that both R points follow the moment of speech, but that the R point of the matrix clause precedes the E point of Bill’s arrival, as illustrated. Future tense in Reichenbach’s system is also simply expressed: the moment of speech, S, precedes the event, E and R. Future tense using Reichenbach’s system is: S___E, R. The representation above indicates that S(peech time) precedes E(vent time), i.e. that the event is future with regard to my speaking. For future perfect tenses, the event of the main clause follows the moment of speech, but it will precede the reference point, R, the time of Bill‘s arrival in (27), which is the E of the adjunct clause (Bill’s arrival). R is also seen in past perfects: (28)

John had finished when Bill arrived. E1 ___R1 ___S1 E2 , R2 ___S2 (R1 = E2 )

The matrix tense structure of (28) is: E _ R _ S; the adjunct tense structure is E, R _ S. That is, in the adjunct clause the Event (Bill’s arrival) and the Reference point are simultaneous, both preceding the Moment of Speech. The main clause Event, John‘s finishing, takes place prior to the Reference point (the Event of the adjunct clause) which is itself prior to the Moment of Speech. Thus, the event of the matrix clause, finished, is not situated directly in relation to S but in relation to a point of reference fixed by the adjunct clause. The crucial point is that the adjunct clause’s temporal structure, in particular its E point, serves as the anchor for the R point of the matrix clause. Obviously, then, Reichenbach’s reference point is important for talk about tense in English. Reichenbach goes so far as to claim that all tenses have R points and he suggests that basic tenses may all be accounted for straightforwardly based on relations between E, S, and R as in (29): (29)

a. b. c. d. e. f.

S, R, E E, R ___ S S___ R, E E ___ S, R E ___ R ___ S S ___ E ___ R

present tense past tense future tense present perfect past perfect future perfect

Note crucially, however, that R is not a logical necessity, even within Reichenbach’s theory. For example, there are some tenses that we can talk about without it, as shown with Pirahã below. So, while it is clearly important to the interpretation of perfect tenses, it is not necessary for the interpretation of the simple past, present, or future tenses. In these tenses, R is always simultaneous with E, the event time and is therefore redundant (at least for simple clauses). On the other hand, E and S do seem necessary if we are to interpret

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

289

the world around us temporally. Without them an event cannot be situated along any temporal dimension. Indeed, it is difficult to conceive of talk about time without a concept of event (E) or a way to indicate whether that event precedes or follows a deictic anchor, S, at least pragmatically or contextually. If this is correct, then E and S are epistemologically prior to R. They make talk about tense possible in the first place, whereas R ‘merely’ enables us to draw finer distinctions and to mark relations between these. A further argument for conceding epistemic priority to the Event point over the Reference point comes from work by Davidson (1967) on action sentences, as further developed by Higginbotham (1985) and Leder (1991). Both Leder and Higginbotham argue at length that E(vent) is a crucial component of all verbal (and even nominal) structures, independent of tense. Leder further argues that this independently needed notion of Event is indeed the same notion of Event that Reichenbach-influenced tense specialists refer to. Assuming something along the lines of the Leder and Higginbotham approaches, then this provides further support for the contention that E and S are the primary building blocks of temporal interpretation, while R is secondary. Of course, many proposals on tense have arisen since Reichenbach (1947), both within philosophical logic (e.g. Montague 1973) and linguistics (for an important survey of facts about tense in English and discourse-based interpretations, see Declerck (1991)). Within Generativism, Hornstein (1990) was one of the earlier proposals. Hornstein accepts most of Reichenbach’s original ideas but, further develops Reichenbach’s model by extending the model to account for a wider range of intrasential relations between tense in matrix and embedded clauses. I will not provide a complete introduction to Hornstein’s model here, referring the reader to Hornstein (1990). I will however discuss those of his elaborations which are necessary in order to understand some relevant implications of Pirahã. Consider how Reichenbach’s theory might handle the famous Double Access Reading cases such as (31) and (32) below. To analyse these examples in Reichenbach’s tense theory, Hornstein (1990) argued that Sequence of Tense (SOT) structures, such as Double Access Readings, require an additional syntactic component in Reichenbach’s model, the Rule for Temporal Connectives (RTC). The latter is stated in (30) (from Hornstein 1990): (30)

Rule for Temporal Connectives (RTC): Write the Basic Tense Structure (BTS) (the linear arrangement of E, S, and R) of the adjunct temporal clause (TNS2 ) under the BTS of the matrix clause, TNS1 ). Associate the S points. Associate the R points by moving R2 to R1, placing E accordingly.

290

DANIEL L. EVERETT

(31)

a. John heard that Mary is pregnant. b. John heard that Mary was pregnant.

(32)

a. John said that Harry will leave. b. John said that Harry would leave.

These are referred to as Double Access Readings (DAR) because the tense of the adjunct clause is checked against both the time of the main clause and the moment of the entire utterance (cf. Giorgi 2010). Under a particular reading, the a and b examples in (31) and (32) are temporally synonymous. They differ in that in the (a) examples, both matrix and subordinate S points are anchored to the actual moment of utterance. In the (b) examples, however, the S point of the subordinate clause is anchored to the E point of the (immediately) dominating clause, as in (33), representing example (31) above:8 (33)

[John heard that [Mary was pregnant.]] TNS1 (E1 ,R1 ___S1 ) TNS2 (E2, R2 , ___ S2 )

(34)

SOT rule:

TNS1 (E1 ,R1 ___S1 ) | TNS2 (E2, R2 , ___ S2 )¹²

Another example of the utility of Reichenbach‘s system includes its ability to account for ungrammatical temporal modification. In (35) for example, the problem is that the Basic Tense Structure of the example is simple past, or E, R _ S. But the modifying expression, at this very moment, requires that R and S be simultaneous, since it requires a present tense meaning. This produces a grammatical/logical violation and the sentence is ruled out: (35)

∗

John left at this very moment.¹³

Example (12) illustrates another problem: (36)

∗

John left tomorrow.

The tense structure of John left is a simple past: E, R S. However, tomorrow requires that R follow S. This produces another violation and (36) sentence is therefore also ruled out.

¹² The SOT rule is provided here for completeness’ sake in presenting Hornstein’s 1990 adaptation of Reichenbach’s system and is not crucial to the argumentation. ¹³ For some speakers this example is fine. In that case ‘at this very moment’ must be interpreted as non-literal.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

291

Now consider what happens when a matrix structure is a simple future and the adjunct clause tense is present (subscripts indicate relevant clause): (37)

John will arrive when you eat pickles. TNS1 (S1 ___R1 , E1 ) TNS2, (S2 , R2 , E2 ). The tenses of each clause must be aligned via the RTC: RTC –> S___R1 , E1 |

S2 , R 2 ,

E2

This eating pickles is linked to John’s arrival, producing the grammatical (simultaneous future events) reading. As noted, however, the R-point is not always needed, in even English. For example, one could say that the R point is unnecessary in past tense. Mary’s pregnancy is unrelated to John’s hearing when both matrix and subordinate tenses are past. Their only relationship is arguably that each occurs prior to the moment of speech. As we see below, for Pirahã the R point never seems necessary. To conclude this summary, the tense points E, S, and, occasionally, R, as well as syntactic conditions on the ways in which these maybe ordered and related intrasententially are crucial to an understanding of English tense. Reichenbach’s system would similarly operate in Pirahã, in sentences like (38): (38)

ti

gáı´ -sai. asi ti soʔóá 1 Say -OLD INFORMATION thus 1 already -áp -á -up REMOTE kapiiga -kakaı´ -sai ʔoogiái hi paper -mark -OLD INFORMATION Dan 3 I arrived. I want to study with/teach Dan.

ʔáab -óp turn -go

ʔigı´ with

o -loc¹⁴

In English this might be best expressed as ‘When I arrived, I wanted to study with Dan’, which would have the Reichenbach structure of R_S_E and R, S_E, where the Rs both refer to ‘my arrival’ and S and E have their standard values. My modified Reichenbach structure for the Pirahã, however, would be: E___S. S___E. No reference (R) point is needed for the Pirahã examples, though one is for the English. These sentences can be understood as time or ¹⁴ ‘Old information’ means that sentence references material already mentioned in the discourse or something that is otherwise known to all Pirahãs.

292

DANIEL L. EVERETT

causally related clauses or, depending on speaker intentions, unrelated events as stated. Nothing connects them morphosyntactically. There are thus two reasons why a simple application of either Reichenbach’s basic system or Hornstein’s (1990) modifications do not work here. First, Pirahã sentences are never embedded (Everett 2005, 2009a, 2012a, 2012b, 2017) and so the temporal interpretation of one clause is bound by the discourse and cultural context rather than morphosyntactic connections to another clause. Normally, the rightmost sentence in (38), ‘I want to study with Dan’, would be interpreted as following the speaker’s arrival noted in the first sentence. But it might also be the inverse, e.g. someone wanting to study with Dan (me) to earn money because they just arrived, the sentence having no implied temporal relationship to when Dan is studying (could have been four days ago, might be tomorrow). Thus due to the lack of sentential recursion in Pirahã, SOT constraints do not exist in Pirahã.¹⁵ We have seen that Reichenbach suggests that tense can be understood in terms of the three points, E, S, and R, but that E, event time, and S, speech time are epistemologically prior to R in talk about tense. As Pirahã and many other languages illustrate, however, temporal systems vary significantly. The discussed variation predicts that tense systems crosslinguistically could vary with regard to R (or its equivalent) but likely never in regard to E or S. The crucial role for R arises not in the basic tenses (where Hornstein (1990: 112) admits that R has ‘… no interpretive reflex’), but in complex tense relations, some forms of temporal modification, and SOT structures. In Section 11.4, we see that this grammatical, ‘naive compositionality’ (e.g. in Montagovian semantics) solution falls short and that generally to interpret tense structures ¹⁵ Paul Postal comments on this section (p.c.) that in Arc Pair Grammar it is possible to treat discourses as sentences, via ‘overlapping arcs’. He says: Now, suppose we apply that view of anaphora to Pirahã’s purported lack of embedding with one further assumption. Assume that so-called discourses are not different from sentences, just are one specific type of sentence. Think of coordination as involving neighboring arcs 1….n, each with the label ‘Coordination’, while discourses involve neighboring arcs 1…n with the label ‘Discourse’. And then assume that Pirahã temporal sequences are just such discourse sentences… The point then is that under the assumptions above, the representation of the Pirahã examples do involve syntactic embedding. This is not an ad hoc assumption to achieve some result about the example, but is simply a consequence of (i) the treatment of anaphora as in Johnson and Postal (1980) plus (ii) the assumption the discourses are/can be single sentences. But of course the problem with doing away with discourse as Postal suggests in this case, is that discourse is a separate domain of which sentences are constituents and is marked by non-sentential particles, coherence, topic-marking, and many other of the facts that have emerged in the vast discourse literature over the years. Moreover, the view that discourse is just a large sentence (not uncommon unfortunately) cannot account for the Pirahã facts or the myriad other aspects of discourse grammar, as many of the works referred to this paper are at pains to demonstrate.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

293

in languages, we must use additional information from culture and the real world. The stronger point is that while Pirahã manifests interpretations for past, present, and future tenses, though it lacks direct syntactic markers for them, it nevertheless lacks interpretations corresponding to perfect tenses, SOT structures, Double Access Readings, complex tense structures, or multiple temporal modification structures, due to the combination of its missing R-point and the absence of sentential recursion.

11.3.3 Conclusion to Section 11.3 To summarize, Pirahã lacks any affixes for tense and otherwise has a very restrictive set of temporal nouns, adverbs, and adjectives. However, as presented in Everett (1986: 288ff ), Pirahã also has a rich aspectual system including suffixes to mark such notions as duration of action, realization of action, internal division of action, continuation of action, beginning of action or state, actions within and outside of the control of the speaker (which overlaps with the notion of proximate vs remote tense), iteration of action, and resultative aspect. There are also two separate positional classes with three suffixes each to indicate the relative certainty of the speaker with regard to the action being asserted (complete certainty, relative certainty, uncertainty) and the speaker’s source of evidence (deduction, observation, hearsay).¹⁶ ¹⁶ However, superficially there may appear to be counterexamples to the claim that Pirahã lacks perfect tenses, represented in (i). Such examples might superficially look like perfect tense: (i)

kaoáíbógí ʔab -o -óp -ai -ta -ha jungle entity remain -STEM -go -ATELIC -REPEAT -COMP.CERT. gíxai soxóá koho -ái -p -há you already eat ATELIC -VERT/IMP -REMOTE -COMP. CERT. ‘The jungle entity completely returned. You already ate’ ∗ ‘When you arrived I had eaten’

-ó -POSITION

Example (i) does not express a perfect tense. Nor is a perfect reading found in either the literal translation or in the most common usage of the Pirahãs. Very similar structures are found in other languages with similar aspectual markings and do not entail perfect tense readings (e.g. Chol, Lydia Rodriguez, p.c.). In other words, not only does Pirahã lack a marker for the grammatical category ‘perfect’ but it also lacks any obvious translation for such a category. It is possible that the Pirahãs might have this conceptually (e.g. as in images/icons of different types of temporal relations), but it is not found in their language. And indeed in Everett (1993) I deny that this structure type has a perfect reading. On the other hand, I have not yet conducted experiments on such readings with native speakers. Therefore I cannot rule out the possibility that (i) might be used as the closest equivalent to a past perfect example for some speakers. Similar examples could exist for future and present perfects, though they are, like (i), translations of independent clauses without overt tense marking. I doubt that this is the case, but, again, experimental work is needed. Of course, if these did receive perfect tense translations, this would

294

DANIEL L. EVERETT

11.4 Peircean Semeiotics and Compositionality 11.4.1 Compositionality The modern idea of Compositionality traces back, as many ideas about language do, to Frege (1984 [1892]) and Peirce (a body of work on language dating back to before 1865). Although Peirce predates Frege in his writings about the formal nature of human language, most current researchers rightly see Frege (1984 [1892]) as another foundational work. The so-called ‘Frege’s Principle’ states that, as Dowty (2007: 3) puts it: (39)

Frege’s Principle: ‘The meaning of a sentence is a function of the meanings of the words in it and they way that they are combined syntactically.’¹⁷

Although Frege’s principle is not without opposition (see Gayral et al. 2005, among many others), one thing is clear to most linguists—the grammatical structure and lexical items in a specific linguistic unit are not irrelevant to its meaning. Just how relevant they are will depend on the degree of ‘compositional transparency’ (Dowty 2007: 8) involved. Languages and constructions, I argue (Everett 2017, in progress) can be more or less transparent and this variation falls into one of at least three types of languages, which I label G1 , G2 , and G3 languages; see Section 11.4.3 below. Consider two English examples: (40)

Ol’ John kicked the bucket.

(41)

Sally kissed Mary.

Example (41) is compositionally transparent, while example (40) is not— idioms are idioms because their literal meaning is not what they actually mean or, as Peirce would put it, they have grown as symbols. All languages likely allow grammatical utterances to vary according to whether they are more or less compositionally transparent. And as we have seen in the examples above, Pirahã temporal interpretations are not compositionally transparent. Yet some linguists and philosophers believe that compositional transparency is universal, with occasional exceptions, all of which can be fixed present a problem for my 1993 analysis that Pirahã lacks the R reference point for temporal interpretations since this point is required for a perfect tense in the Reichenbachian system. But they do not seem to have perfect tense interpretations. ¹⁷ Once again, like most researchers of his era and the modern era, unlike Peirce, Frege’s view of language was sentence-bound. In Everett (in progress), I suggest that this is because of the centrality of the proposition, often expressed by a single sentence in natural language, to much philosophy of language and mind. But as Stjernfelt (2014) and others make clear, propositions can be expressed perfectly clearly external to language.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

295

by plugging the syntax into the pragmatics and downloading conversational implicatures and the like. I refer to this position as ‘naive compositionality’ (Montague (1973) being perhaps the best example). The problem with naive compositionality, however, as currently practised, is that it makes two erroneous assumptions—that all we need for literal semantics is in the syntax and that compositionality takes place at the sentence level, leading to the unfortunate fact that most syntactic theories are largely sentence-bound. There are many non-naive views of compositionality (i.e. where compositionality is not driven by the structure of the sentence), of course. Jaszczolt (2005) and (2009) are two such examples. But although Szabó (2000), Jacobson (2014), Dowty (2007), Pelletier (1994), Kamp (2019), Fodor and Lepore (2002), Lepore and Stone (2015) and Peirce (1909), inter alia, offer more nuanced views of compositionality; they are by and large still sentence-bound and omit any theoretical role for culture.¹⁸ Thus study of compositional transparency variation across constructions and across languages is vital for improving our theories of human language. In this regard, Pirahã joins all other languages of the world in forcing us to go beyond naive compositionality and pragmatics in order to model understanding of temporal intensions and extensions. As we see below (and see also Everett 2005, 2012, 2016, 2017), Pirahã allows for less compositional transparency between the syntax and the semantics than some other languages. I am going to argue that what unites all languages is not identical compositional transparency but what I will call ‘semiotic inference’ (using signs to infer other signs, linguistic or otherwise; see Section 11.5.2). As we see in Section 11.4.3. compositionality can vary in its direct connection to the syntax. Therefore we must be careful to avoid proposing a single model of transparency designed to fit all languages. See Jaszczolt (2012) as well for additional discussion.

11.4.2 Compositionality challenges A modern statement of Frege‘s Principal is found in Szabo (2000): (42)

‘Principle of Compositionality: the meaning of a complex expression is determined by meanings of its constituents and by its structure.’ (Szabó 2000, 3)

¹⁸ It might sound strange to say that Kamp’s Discourse Representation Theory is sentence-bound. In several ways it clearly is not sentence-bound. But it lacks any account of the standard features of discourse that, say, Longacre (1983, 1976), Grimes (1976), inter alia). Thus it extends sentence principles across sentence boundaries, but omits vital characteristics that distinguish discourse and justify not treating those boundaries as insignificant. Discourse is not a mere set of conjoined sentences.

296

DANIEL L. EVERETT

But this bare statement on its surface does not offer a solution for Pirahã temporal interpretation (and neither do ‘fixes’ such as empty categories or contextual enrichments. See Everett (in progress) for details.) The question remains as to how we achieve temporal interpretations for time words, affixes, and discourses in Pirahã without either precision time words or tense morphosyntax, if not by naive compositionality along the lines of (42). In light of (42) let’s reconsider examples like the following from Pirahã: (43)

a. kohóai -xiigá. tı´i gı´ ʔahoái -soog -abagaı´ eat -CONT. 1 2 talk -want -FRUST.INI. ‘(You) are eating. I want to talk to you’ (Free translation: ‘When you finish eating, I want to talk to you.’ )

The interpretation of this example depends on seeing someone eating or quoting someone. The interpretation also depends on the cultural understanding of -abagaı´ which can also be an illocutionary force marker, i.e. this is an indirect speech act (literally it is ‘I almost begin to want to talk to you’, making it not quite a direct statement). So we only know the temporal meaning of this example, like the English examples in (56)–(66) below by a combination of real-world knowledge and experience, with cultural values on how best to express ideas. Thus again we see that in spite of the fact that lack of tense morphemes and precise temporal lexical items, the speakers of Pirahã are nevertheless able to come up with precise temporal understandings, modulo culture (see below).¹⁹ To accomplish this, they must infer from the context, the discourse, the words, expressions, gestures and so on what time frame is implicated in the meaning of individual utterances. Peirce argued that all reasoning takes place via semiotics—using one sign to interpret another. For Peirce, however, signs, including entire propositions, need not be ‘expressed’ linguistically. An interpretation can take words, gestures, immediate context, cultural values, and so on into account, not being exclusively bound by the linguistically expressed elements (see also Jaszczolt (2005, 2009, 2016a), inter alia, for similar findings).²⁰ We have seen that Piraha

¹⁹ By the expression ‘modulo culture’ I mean that cultures have different rankings of time values and their relevance (see Everett 2017). Thus their understandings of what a ‘long’ time or a ‘distant’ time will vary, among other aspects of the mapping between temporal ontology (Steedman 1997) and linguistic meaning. ²⁰ Consider in this regard the wind blowing the weather vane to the west when it is blowing in from the east. The configuration/movement of the weather vane expresses the proposition ‘The wind blows westward’ nonlinguistically but fully semiotically (showing that linguistics is but a subspecies of semiotics).

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

297

uses inferences from a variety of sources to come up with temporal interpretations, for example. This use of linguistic and non-linguistic signs to infer meanings for utterances is what I refer to as ‘semiotic inference’ (which is one way to classify the work of Floyd (2016) and Rodrı´guez (2014, 2019), for example). Semiotics is crucial here because linguistic forms and divisions (phonemics, semantics, morphosyntax, pragmatics, semantics) are all subdivisions of semiotics. Yet another excellent example of the failure of naive compositionality in temporal interpretations is found in Wari’ Intentional State Constructions (ISCs; Everett 2009b). These are used to express a variety of concepts, including temporality. Consider, for example, the use of these constructions to express quotatives and future tense (Everett and Kern 1997 analyse these examples in the context of the wider Wari’ grammar). The embedded sentence carries stress only on the last syllable of the sentence, i.e. the embedded sentence/predicate is stressed like a word; the embedded sentence is otherwise a non-idiomatic, fully productive sentence: QUOTATIVE (44) ma’ co mao na -ini Guajarái that:prox:hearer m/f:rp/p go:sg 3s:rp/p -3n Guajará (Brazilian city) (45)

naj -namk ‘oro narimak ’ taramaxiconj . 3s:rp/p -3f collective woman chief ‘Who went to Guajará?’ (said) the chief to the women’

In what follows, I use the node predicator as a neutral term for lumping together verbs and the semantic nucleus of ISCs (these nodes are labelled NUCLEUS in Role and Reference Grammar, RRG; VanValin 2001). Also, the grammatical relations (subject, object, indirect object) in the tree diagrams below are informal labels. Everett (2009b) restates them in RRG terminology (VIC stands for verbal inflectional clitic the normally second-position clitic, which agrees with the subject and the indirect object; see Everett and Kern (1997: 5ff ) and Everett (2009b) for more details on the language and this construction in particular).

298

DANIEL L. EVERETT

(46)

S

PREDICATOR

VIC

INDIRECT OBJECT

SUBJECT

S

najnamk

‘oro narimak

taramaxiconj

PREDICATOR Ma’

co

mao

VIC

INDIRECT OBJECT

nalini

Guajarái

In (46) we see an (asymmetrically) embedded sentence, ma‘co mao ‘who go‘, in the predicator position of the larger clause nanam ‘oro narimataramaxicon which is followed by the agreement-tense clitic complex, nain, where na agrees with an understood masculine subject (not part of the structure) and in agrees with Guajará. The literal meaning of this most embedded clause is ‘Who went to Guajará’, where Guajará is the indirect object of the verb mao. This is in turn embedded in a larger structure (lacking a verb), ‘he-to-them (fem) Chief women’. It means literally ‘Who went to Guajará (said) Chief to women’. Though a verb of saying is necessary to the English translation, it is not necessary in the Wari’ clause (more on this below). These intentional state constructions are also used to express future time reference (even though Wari’ also has future tense morphemes that such constructions are gradually replacing; Everett and Kern 1997). Example (47) shows a temporal ISC construction, used to communicate future time reference (see the next section) though it has the form of a quotative (Everett and Kern 1997: 55ff; Everett 2009b). (47)

cao‘ xi’ carawa nana hwijima’ eat 1pincl:rf animal 3:rp/p children ‘The children will eat food.’ (lit: ‘We will eat food, the children (say).’ )

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

(48)

299

S PREDICATOR S

VIC

SUBJECT

nanai

hwijama'i

PREDICATOR

VIC

OBJECT

cao'

xi'

carawa

In such examples, the meaning is not compositionally transparent, but is computed based on the speakers’ knowledge of Intentional State Constructions. One could of course force the interpretation of (47) to be compositionally transparent by adding a stipulation to Wari’ grammar, e.g. ‘A sentence under the NUC/Predicator position is interpreted as a predicate and no other predicate is allowed.’ This stipulation allows the speaker to compose the meaning from the syntax in Wari’. But although this works mechanically, it begs the questions of (i) how this language-specific stipulation is learned and (ii) how such stipulations fit into other aspects of temporal interpretation, as in English and Pirahã, etc. That is, in a theory of linguistics in which inference, rather than a series of language-specific components is implicated, such stipulations are unnecessary. Continued exposure to such examples leads to ‘habits of inference’ (Hartshorne and Weiss 1932) in which nothing further need be stated. The unifying answer is inference—speakers learn and interpret ISCs by induction (going from parts of utterances to the entire utterance) and deduction (having learned a grammar they apply deductive principles to determine that ISCs have a stipulation attached that tells them how to parse the utterance topdown). And the child learning the language begins its learning via inferential principle of Peircean abduction (Section 11.5), and tests its guesses by induction (to build a theory of the language) and deduction (to use that theory to interpret individual utterances).²¹ Now let us turn back to Pirahã. Just as temporal interpretations are not computationally transparent, neither are recursive interpretations. This is ²¹ This does not rule out a role for innate knowledge. This chapter does not take on that larger issue. But innate knowledge is used inferentially like any other knowledge (for Peirce innate knowledge is phylogenetic learning while non-innate is ontogenetic learning, though he was relatively agnostic on the amount, if any, of the former in human cognition). On the other hand, I have argued elsewhere, e.g. Everett (2012a, 2016) there is very little, if any, convincing evidence for nativist, rationalist beliefs on the acquisition of knowledge.

300

DANIEL L. EVERETT

interesting because it reinforces the point that sentence interpretations need not be exclusively computed from the linguistic structure of single sentences. In other words, Pirahãs use inference to learn the unique principles of their grammar and inference to parse and interpret the meanings of their discourse, sentence, word, and other signs. If inference is indeed crucial to the learning of individual languages, beyond the lexicon—e.g. constructions, stipulations, idioms, etc., then all speakers of all languages must use inference to construct their grammars, modulo innate parts. But, again, compositionality is nothing more than the use of linguistic/semiotic knowledge in the inference of meanings, appropriate contexts of usage, grammaticality, and so on, consciously or unconsciously (see below and Harman 1973; Brandom 1998; Peregrin 2014). Especially in light of Peirce‘s (1868) arguments against intuition and introspection as ‘capacities of man’ (see footnote 26 below). Inference is also important in understanding the relationship between structures relating syntax and semantics. Take the inference of meaning from combined sentences in discourse (for example Peirce 1909; Kamp 2019; Grimes (1976; Givon 2020; Longacre 1976, 1996, and so on). Independent sentences can be interpreted semantically by a formally non-sentence-based compositionality (see especially Peirce 1909) for the original theory that is isomorphic to modern-day Discourse Representation Theory). To take one example, Sauerland (2018) argues that Pirahã sentences manifest recursion, based on the fact that recursive interpretations are possible for some sequences of linguistic units and his apparent assumption of naive compositionality. His conclusion is that the linguistic units must form a single recursive unit in order to account for the recursive semantics. But this does not follow. Taking his examples (very poorly transcribed, so my own corrections are included) we can see why the point does not follow, owing to lack of compositional transparency. (49)

Spoken by speaker 1 (Toe):²² ce kahápe ogéhiai igeuo I go star up (This should be: ti kahápiı´Ɂogihı´ai Ɂigı´—o)²³ (I go star ALONGSIDE-LOC.) ‘I have been to the stars’

²² These transcriptions lack much grammatical information, unfortunately. The phonetics is also only partially represented. ²³ The reason that I constantly make the effort to correct the mere phonetics of data collected by people without any speaking knowledge of Pirahã, as I have been asked on multiple occasions, is not to make these researchers look sloppy (though that is indeed an effect), but because as any field researcher knows, without careful phonetics, one is shooting blind in what he or she considers relevant or even

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

(50)

301

Spoken by speaker 2: Toi he gáisai ce kahápe ogéhiaiigeuo Toe say 1p-SINGULAR have-been stars (This should be: Tooı´ hi gái-sai ti kahápiı´Ɂogihı´ai Ɂigı´ – o) (proper name he spoke. I go star ALONGSIDE-LOC.) a. co-ordinate interpretation: ‘Toe talked, and I have been to the stars’ b. subordinate interpretation: ‘Toe said “I have been to the stars”’

Sauerland hypothesizes that there are two interpretations of the sequence of words in (48) and (49): the co-ordinate interpretation in (48) and the subordinate interpretation in (49b). Sauerland further hypothesized that the subordinate interpretation requires syntactic recursion in order to be interpretable as such, whereas the co-ordinate interpretation does not require syntactic recursion.²⁴ He constructed ten items like (49) and (50), and a further ten control items like (51) and (52) where speaker 2 misreports what speaker 1 says: (51)

Spoken by speaker 1 (Toe): ce kahápe kahe’ai igeuo I go moon ALONGSIDE (This should be: ti kahápiı´ kahaiɁaı´iı´Ɂigı´ – o) ‘I have been to the moon.’

(52)

Spoken by speaker 2: Toi hi gái-sai ce kahápehai heesé igeuo ‘Toi said ‘I have been to the sun’ (This should be: Tooı´ hi gái-sai Ti kahápihaı´ hisı´Ɂigı´ – o) (Tooı´ he spoke. I will go sun ALONGSIDE-LOC.) a. co-ordinate interpretation: ‘Toe talked, and I have been to the sun’ b. subordinate interpretation: ‘Toe said “I have been to the sun”’

Critically, both interpretations of (52) are false. Sauerland then had sixteen Pirahã speakers take part in his survey. In this survey, participants were asked perceptible in the syntax. All field workers understand that every statement on the grammar of a given language ultimately rests on a solid phonetic foundation or it is of little worth. ²⁴ It is worth mentioning that it is a mistake to confuse apparent embedding with recursion. Recursion is the ability for a rule to apply to its own output, leading to constituents of arbitrary length. If the constituents are limited such that, as in Pirahã (Everett 2012b) there is a longest sentence, then something other than recursion seems to be occurring. In the case of Pirahã a quotative is formed of two juxtaposed clauses. Even if, for example, it could be shown that a language has two clauses or sentences functioning for some purposes as a single unit, this does not entail that the language has recursion. The two clauses could be united only by their interpretations, rather than syntactically, just as in Peirce’s (1909) existential graphs.

302

DANIEL L. EVERETT

to decide whether each of the twenty items were correctly understood by Speaker 2 (specifically, they were asked ‘Did Speaker B hear well?’). Participants were trained on both versions of one item: they were told that they should say ‘no’ to the control item and they should say ‘yes’ to the target item. They were then tested on the remaining eighteen items (nine target, nine control). Sauerland reported above chance behaviour on the target items, and concluded that Pirahã contains true syntactic embedding. There are several problems with this analysis. (The reader is referred to Everett and Gibson (2019) for more details, which also criticize the design and interpretations of Sauerland’s experiments.) Most importantly, once again, this analysis equates a potential embedded interpretation with a need for syntactic embedding to obtain that interpretation. In particular, there is no reason to assume that interpreting (55) as ‘Toe said “I have been to the stars”’ requires any syntactic recursion. As many others have noted in the discussion of recursion, sets of non-embedded syntactic materials can easily give rise to an embedded semantic interpretation, especially if such an interpretation is contextually supported. For example, Hollebrandse (2018) makes exactly this point about English examples like (52), as I do as well in Everett (2010): (53)

Malcolm is guilty. The jury knows that. The judge knows that. (Example is from Hollebrandse p. 37)

An available interpretation of (53) is that the judge knows that the jury knows that Malcolm is guilty, in spite of the fact that there is no syntactic embedding in this example. And in English, given a context in which someone has just said ‘I have been to the stars’, if a second speaker says ‘someone said something. I have been to the stars’, most listeners will agree that the meaning of this in the context is that Speaker 1 said that he has been to the stars, even though there was no syntactic embedding in the original statement. Indeed, this alternative possibility to Sauerland’s assumed reading is testable and so Everett and Gibson (2019) report on Gibson’s testing of the alternative. Gibson ran the relevant control experiment in English, with 20 participants from Amazon’s Mechanical Turk, using the written versions of all ten of Sauerland’s items (as presented in the appendix in his paper), and using the very instructions that Sauerland provided to the Pirahãs (‘Did Speaker B hear well?’). Example target and control items are given in (54) and (55). (54)

Example target item: John: I have been to the stars. Bill: John said something. I have been to the stars.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

(55)

303

Example control item: John: I have been to the moon. Bill: John said something. I have been to the sun.

Note that there is no syntactic embedding in the written form of what Bill says in each discourse: there is no quotation or embedded sentence. The embedded meaning would have to be inferred because it is not present in the presented syntax.²⁵ Sauerland’s experiments are based on the now-falsified view of naive compositionality. So where do these interpretations come from? Where do any meaning interpretations come from? Jaszczolt (2005, 2009, 2016a) and Everett (2016) argue that interpretations are pieced together using whatever information is available to the speaker. In fact, in Jaszczolt‘s theory (see references cited immediately above) the structure of the sentence need not even provide the backbone for compositional semantics. Notice that if we were to step outside the confines of naive compositionality, we could refer to any piecing together of information (from solving crimes to baking cakes from recipes) as what they are—inference (at least as I discuss it below and as found in the references. Inference can be very fast and unconscious or slow and conscious). In fact, we might say that the main function of the cerebral cortex (see Rolls 2016; Kortylewski et al. 2021; Baroni 2020; Bienenstock et al. 1996; inter alia) is to ‘compose meaning’ via inference. As visual compositionality pieces together the visual components of the environment along with learned visual recognition patterns, so linguistic compositionality is that form of inference which uses the clues of the linguistic elements to piece together partial meanings but is simply a subtype of Peircean inference (deduction, retro-/abduction, and induction).²⁶ As Peirce says: The elements of every concept enter into logical thought at the gate of perception [via signs, DLE] and make their exit the gate of purposive action [interpretation of signs]; and whatever cannot show its passport these gates is to be arrested as unauthorized by reason. (Peirce CP 5.130)

Peirce says that all knowledge, including all linguistic knowledge, is the result of inference. He attacks the idea that there is any special power used for making linguistic judgements, other than inference, rejecting Cartesian intuition ²⁵ Our English participants agreed with the target sentence on ninety-nine per cent of the trials, demonstrating that they obtained the embedded interpretation in spite of the lack of embedded syntax. Furthermore, they disagreed with the control (as desired) on ninety-eight per cent of the trials. All materials are available at osf.io/z86k2/. ²⁶ Clearly, as the works cited make clear, our inferential abilities are underwritten by the structures of our brains.

304

DANIEL L. EVERETT

and introspection, and was the first American social scientist to demand that research be quantitative and not merely qualitative, emphasizing replicability and mathematical precision.²⁷ It is no surprise, therefore, that in recent years several works, e.g. Russell (2012), have argued that statistical inference is crucially implicated in speakers’ interpretations of utterances (cf. Lepore and tone 2015). To better illustrate the role of inference relative to naive compositionality, consider the following examples from English:²⁸ In these examples the grammaticality (or felicity, depending on one’s theory) of each example varies depending on real-world and cultural knowledge. (56)

a. John reported that Mary has COVID-19/is happy. b. Yesterday, John reported that Mary has COVID-19/?is happy. ∗ c. Almost a hundred years ago, scientists concluded that the ivory-billed woodpecker is pregnant/happy/flying. (modification of example from Barbara Partee, p.c.) d. Almost a hundred years ago, scientists concluded that the ivory-billed woodpecker is extinct. (Example from Barbara Partee, p.c.)

(57)

a. John claimed that Mary is pregnant. b. Yesterday, John claimed that Mary is pregnant.

(58)

?a. Twelve months ago, John claimed/reported that Mary has COVID-19. ∗b.Twelve months ago, John claimed/reported that Mary is pregnant.

(59)

a. Twelve months ago, John reported that the elephant is pregnant. ∗b. Thirty six months ago, John reported that the elephant is pregnant.

²⁷ Peirce (1868) argues at length that there is no such thing as intuition, a central part of the cognitive research of Descartes and many linguists (e.g. Chomsky). Defining intuition as a cognition unlinked to a previous cognition, Peirce concludes that no such cognitions exist and that all cognition is part of a chain of inference with other cognitions. He claims that the only evidence for intuition is that we think we have it. To this dubious claim he replies: ‘A child has, as far as we know, all the perceptive powers of a man. Yet question him a little as to ∗how∗ he knows what he does. In many cases, he will tell you that he never learned his mother tongue; he always knew it, or he knew it as soon as he came to have sense. It appears, then, that he does not possess the faculty of distinguishing, by simple contemplation, between an intuition and a cognition determined by others [which would be an inference, DLE].’ ²⁸ Different authors mark some of these examples as pragmatic infelicities or grammaticality violations (see Giorgi 2010 for example). Such distinctions are irrelevant if we recognize all interpretations as forms of inference.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

305

(60) ∗a. Two years ago, John reported that his neighbor Tricia is happy. ?b. Two years ago, John reported that the Virgin Mary is happy. ?c. Today, John reported that his neighbor Mary is happy. d. Two thousand years ago, John reported that the Virgin Mary is happy. e. Zookeepers who examined Ellie the elephant in the Cincinnati zoo five months ago announced/published that she is pregnant. (Example from Barbara Partee, p.c.) (61) ∗a. One thousand years ago, John reported that Bill is his friend. b. One thousand years ago, John reported that Muhammed is God’s prophet. c. Two thousand years ago, John reported that Jesus is alive. ∗d. Two thousand years ago, John reported that Bill is alive. e. The ancient Egyptians believed that the earth is flat. (Example from Barbara Partee, p.c.) People interpret and evaluate the grammaticality of the sentences above inferentially, via cultural and real-world knowledge, in conjunction with their knowledge of the words and structures of their languages. All of the judgements in (56)–(61) (whether we call the judgements pragmatics, semantics, syntax, or whatever is irrelevant) rely on inferential reasoning. Thus in (61c), Jesus, as an eternal being to some religions, can be alive after a thousand years, while Bill cannot be in a physical sense (though for some he is eternal in heaven). Each of the contrasts in (56)–(61) depend on inference in which the linguistic information is just one part of the final felicity/grammaticality of the example. One might of course argue that compositionality plays its role and then submits the result to pragmatics to determine if the constructed meaning fits the context. But there are a couple of problems with this suggestion. First, we must ask why this division is desirable or whether it only arises in order to artificially distinguish compositionality from the general inferential abilities independently known to be possessed by humans (indeed, all animals). Again, cooking by a recipe or solving crimes seem to be abilities that require a power and process of inference identical in operation to the construction of sentence meanings. The difference is that in a recipe part of our inferential procedure draws on our knowledge of cooking whereas the inferential procedure in natural language draws on our knowledge of story structure, conversational structure, our culture more broadly, etc. Both recipes and natural language inference draw upon our knowledge of language as well.

306

DANIEL L. EVERETT

The idea that there are only inferences in understanding our native tongues, not some other special capacity (intuition) of the mind, obviously means that no speaker is able to make intuitive judgments about what is grammatical or not, because intuition doesn’t exist. We only judge whether something is grammatical or not just as we only judge what something means in the first place—via inference, using one or more of the three ‘-ductions’—induction, abduction, or deduction. If I ask you if the following sentences are OK, what is the process by which you answer me? (62)

John is three years old and is CEO of a major company.

(63)

John are the nicest guy I know.

(64)

Talking about Mary, he is a smart woman.

(65)

Who do you wonder whether saw John?

(66)

Who do you wonder whether John saw?

According to Peirce there is only one answer for any form of reasoning, inference. You (child or adult) know the answer to these questions because you infer that it is ungrammatical or grammatical and why based on its comparison to other sentences, using known signs to infer properties of unknown signs (as we have seen, Peirce demonstrates the vacuity of notions like ‘intuition’ and ‘introspection’, replacing them with inference). But for Peirce the forms of inference are dependent on his theory of signs, his semiotic. Simply put, Peirce’s semiotic system differs from all others in its strict triadicity. A sign must have three components (not merely the Saussurian dyadic form+meaning). These are the Object, Interpretant, and Representamen (the form of the sign). So ‘apple’ has the phonemic form ‘apple’ that varies by dialect and it has as object, say, the red, sweet fruit that we make cider with. But the form and the object can only come together as a sign of some type if they have an interpretant—if they can be interpreted by other signs in the language. In ‘apple of my eye’ the interpretant of apple will be different than in ‘apple in my eye’ and so on. Semiotic inference uses knowledge of linguistic signs and their arrangement (the arrangement is also a sign) and other forms of knowledge in an inferential process built on cultural learning and semiotic principles, such as ‘closeness in function closeness in syntax’ (also known as iconicity).

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

307

11.4.3 Compositional transparency and linguistic typology This leads us to a proposal in Everett (2017) and Barham and Everett (2020), in which it is argued that there is more than one bauplan behind the organization of the individual languages of the world. These language types (which are not exhaustive) are classified according to the relative degree of compositional transparency, i.e. how tight the fit is between the syntax and the semantics. I propose three broad organizational structures and argue that we find examples of all three in the world‘s currently spoken languages. Only one of these organizational plans might support the compositionality that Szabó (2000) and others assume, however. For the other two types, compositionality is less directly connected to the syntax. The broad language types that I defend here are: a linear order grammar, G1 , (subject–verb–object in (69)) that conveys meaning (represented in (70) and (71)). G2 languages, which have hierarchical structures but no recursion, and G3 languages, which have recursion, as well as all of what G1 and G2 languages have (73) (Everett 2017: Chapter 9). In this hierarchy of grammars, there is no need for a protolanguage in language evolution; a G1 language is sufficient to convey nuanced, abstract meaning. G1 languages may evolved first, with recursion a late and unnecessary expectation for early languages (Karlsson 2009; Everett 2017; 2012b). G1–G3 coexist today with G1 and G2 languages found in some societies without graphic traditions (Everett 2005; Gil 2008; Pullum 2020).²⁹ The empirical differences in these three grammars are illustrated diagrammatically using sentences (66)–(69) and (70)–(73): (67) (68) (69)

John came in the room. John sat. John slept. (Interpreted as ‘John came into the room sat and then slept’.) John entered the room by the garden. John slept. (Again interpreted recursively.) John came in the room, sat, and slept.

²⁹ Perhaps a more workable definition of compositionality in light of the above discussion is that of Pietarinen’s Pragmatic Principle of Compositionality (PPC): (i) ‘The meaning of a sentence is the meaning of all sentences that follow from that sentence either by inductive or deductive principles and permissions under all authorized circumstances (i.e., those arising out of mutual consent by [the interlocutors, DLE]’. Pietarinen (2005: 525ff ) But this would still be a subtype of inference. And it also fails to account for the greater amount of information found entirely outside of sentences, as per Jaszczolt (2009).

308

DANIEL L. EVERETT

The illustrations in (70)–(71) conform to a G1 grammar. In these diagrams, there are no category labels, e.g. ‘noun’ or ‘verb’, and no phrase labels, such as ‘verb phrase’. The simplest grammatical structure would be a linear arrangement of words as a proposition/sentence. There are modern languages represented by G1 grammars, for example, Pirahã (see also Futrell et al. 2016; Everett 2005, 2009a; Everett and Gibson 2019) but also Warlpiri, Wargamay, Hixkaryána, Kayardild, Gavião, and Amele among others (Pullum 2020). A G2 grammar would allow the structure in (72) which shows hierarchical nesting of sub-phrases. A G3 grammar would allow structures such as that shown in (73). Two sentences are contained in or ‘dominated by’ the highest sentence making this a grammar without constraints on recursion. These different grammars can be understood as a set of ‘templates’ in the Role and Reference Grammar sense (Van Valin 2001), i.e. there are similar proposals in the literature.

(70)

Sentence

John

(71)

came

into the room

Sentence

John

sat

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

(72)

309

Sentence

V(erb) P(hrase)

N(oun) P(hrase)

V(erb)

N(oun)

NP Det(erminer)

entered

John

N

the

room

(73) is a diagram of the embedded structure of a G3 language with recursion.

(73)

S(entence) NP

VP

N

V

John

came

PP P

S

NP

NP

VP

S

(John) V into Det

N

the room

sat

NP

VP

N

V

(John)| slept These grammar types are hypothesized to reveal the differential degrees of compositional transparency found in languages of the world. Thus temporality in Pirahã and other languages has much more to teach us about human cognition and the interpretation of language than merely a novel tense system. If there were space, we might further explore how this inferencebased account of interpretation can account for a variety of other interesting facts about Pirahã such as the contrast between interpretation of numerical

310

DANIEL L. EVERETT

concepts vs colour concepts, the distinction in the language between the absence of quantifiers (of certain types) and generic terms, and so on (cf. Everett, in progress).

11.5 Conclusion 11.5.1 Dark matter and culture The Pirahãs, the Wari’s, English speakers, and others know what times are relevant and important in their languages because of their knowledge of the external world, cultural values, the ways in which these values have been encoded into their languages, and the ways that they are learned and applied— all examples of inference. This correspondence between the world, culture, grammar, and meaning also supports the idea that we ‘get’ meanings because of our species’ advanced capacity for the general animal ability to infer from one sign (or datum) to another. This raises several additional questions. Setting those aside for now, if this widely accepted process is not wrong, compositionality is but a special form of inference that is applicable once inference has built up a sufficient knowledge of the language to enable the further, more specialized inference from form, culture, real-world knowledge, discourse context, sentence structure, lexical choice, etc. to meaning. Compositionality, a subdomain of inference, is conducted by the same means just in a distinct semiotic domain, i.e. linguistic form (for another semiotic domain, vision, cf. Kortylewski et al. 2021; Baroni 2020; Bienenstock et al. 1996). Using culture and inference further answers the question of why any languages have tense marking in the first place and, if so what kinds. After all, if all languages could construct interpretations without tense as Pirahã does, why have tense morphemes or grammatical operations? Because cultures value (Everett 2016) more or less precision in temporal interpretations, precision that is underwritten by a wider array of terms, and constructions. Cultures create tense markings for the same reasons that they create colour terms, numerals, and so on—what terms a language has and how it builds them into discourses, the lexicon, sentences, and so on is a matter of cultural values. No special linguistic ability or endowment will in general be necessary. Some cultures value one set of words, others value another.³⁰ From these terms inferential relations and interpretations are born. ³⁰ For example, my subculture does not value golf. My male relatives considered golf a ‘sissy’ sport and any man who played it was not quite a real man. So, growing up in such a culture, I have no terms (and I never left those cultural confines in this respect) for golf. I do, however, know a lot of words about cattle and the raising of them, since this was a respectable male activity in my upbringing.

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

311

On the other hand, the question of why people appear to have semantic interpretations for absolute tenses in every language known even without tense-marking doesn’t seem to be a cultural problem but an issue of universal facts about animal cognition and the relationship of life to the world. Yet even here, we still must learn these tenses inferentially. Another crucial cultural component of time in Pirahã is what Everett (2005) calls the ‘Immediacy of Experience Principle’. Stated informally with regard to time/tense/temporal interpretation, this simply means that as far as temporal semantics goes (see Everett 2005 for more details), Pirahãs do not talk about the distant past or the distant future (‘distant’ being loosely defined as beyond two generations). They may talk about what their children will be when they grow up, but this is uncommon. They may talk about a time when they had not seen outboard motors, but this is not a distant past and is easily inferrable from the fact that the motors are known to have postdated the arrival of the first missionaries. This is not to say that the Pirahãs have no idea that there will be a distant future or that there is/was a distant past. But they do not as a rule talk about these. It is similar in this respect to taboos in many cultures (but not in Pirahã) against naming the dead. Or like raising your voice to children in Western culture. These are, as cultural values usually are, violable constraints (in the sense of Everett 2016), but strong (highly ranked) constraints nonetheless. This principle may very well have affected the development of the tense system in Pirahã or, vice-versa, may have come into the culture due to the language— as I point out in Everett (2016) the direction of causality in language-culture relations is not always transparent, largely because these two domains form a symbiosis.

11.5.2 The role of inference Inference is in a sense the subject to which Peirce devoted his entire research career, across physics, chemistry, mathematics, logic, philosophy (especially pragmatism), and so on. In this sense Peirce is similar to Herbert Simon who dedicated most of his research life (Nobel-Prize-winning economic theory, management theory, cognitive science, computer science, and so on) to problem-solving (an application of inference). Peirce is careful to emphasize throughout his body of work that there is only one way that humans know anything—inference. Inference for Peirce has three forms (prior to Peirce, only deduction and induction were understood/known):

312

DANIEL L. EVERETT

(70)

Deduction: 1. Rule (major premise) 2. Case (minor premise) 3. Result (conclusion)

(71)

Example of Deduction: All phrases are formed by Merge (major premise/rule) ‘John’s brother’s friend’ is a phrase (minor premise/case) ‘John’s brother’s friend’ is formed by Merge (conclusion/result)

(72)

Induction: 1. Result 2. Case 3. Rule

(73)

Example of Induction: ‘John’s brother’s friend’ is formed by Merge (result) ‘John’s brother’s friend’ is a phrase (case) All phrases are formed by Merge (rule)

(74)

Abduction: 1. Result 2. Rule 3. Case

(75)

Example of Abduction: ‘John’s brother’s friend’ is formed by Merge (result) All phrases are formed by Merge (rule/guess) ‘John’s brother’s friend’ is a phrase.

We negotiate our paths through life by inference. According to Peircean inference, we either build generalizations (induction), we decompose general ideas into particulars (deduction), or we guess (abduction). New knowledge, as Peirce argued at length, arises primarily via abduction, which he also referred to as ‘ampliative inference’. We guess. Then we make inferences about that guessing and a combination of induction and deduction will tell us whether our abduction is also on the right track. It might surprise some to think of abduction as the formalization of guessing. This is why it is important that Peircean abduction (and Peirce is the originator of the term) should not be confused with more recent bastardization of the term to mean ‘inference to the best explanation’, which should have a separate

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

313

term, because it has only the most tenuous of connections to Peirce’s superior work (superior because it follows from a theory and is not simply a useful conceptual category). It is of course crucial to remember that induction, deduction, and abduction—the trinity of inference—are not necessarily conscious processes. We inference our way through most activities. If I grab my coat leaving the house, that is based on an inference. If I use an honorific in Japanese, that is an inference. If I use a ‘frustrative’ verbal affix, that is the result of inference. A child learns their language through inference (even if one believes that the grammar is given a priori, the words are learned inferentially for the most part. And of course I believe all of language is used inferentially, whether based in large or small part on innate dispositions). Just as Peirce believed that all philosophy and reasoning should be firmly based in mathematical principles, Russell (2012) offers a quantificational theory of inference in certain domains of linguistic knowledge. The avoidance of such quantificational inference in modelling human interpretations of sentences, actions, events, and so on, is done at science’s peril (cf. Lepore and Stone 2015). Compositionality, then, deduces, induces, or abduces meaning from our linguistic world. It is not, in this view, a property of utterances. Rather it is a special domain (language forms, meaning, context) that uses any of Peirce’s three forms of inference to attach meaning to utterances, discourses, etc. Moreover, in compositional meaning, prior knowledge of linguistic units may be necessary, though it is never sufficient. Prior knowledge is not always necessary because we can use inference to figure out utterances that have foreign words we do not know in them or even entire foreign sentences we have no previous knowledge of (especially in the appropriate contexts). Of course, in computing the meaning of a sentence, priority will be given to the information contained in the sentences—the words contained and their arrangement. But this particular inferential operation is always sensitive to context, which in its turn may include different types of information (leading occasionally to ‘coercion’ (Pustejovsky 1995) of word meanings), e.g. the introduction of foreign words into the sentence, implicatures (Grice 1991), and so on. As stated earlier, by way of simple illustration, this is similar to how we piece together the disparate forms of knowledge and the conclusions required to follow a recipe, another inferential process. A recipe for biscuits may be followed carefully with the final biscuit the ‘compositional interpretation’ of the recipe. But recipes, like linguistic meanings, can be overridden. A given recipe can be altered to fit dietary restrictions, different altitude, extra spice, lack of

314

DANIEL L. EVERETT

ingredients, and so on. This is the same operation of inference as interpreting a sentence. The purpose of dwelling on this is to remind us that it is not necessarily the case that linguistic meaning is as special as we might have otherwise thought. The above discussion is further support for the old claim by Bar-Hillel (1964: 174) that ‘the idea of fully automatic high-quality translation (FAHQT) is just a dream which will not come true in the foreseeable future’. This is not because we lack understanding of syntax, semantics, pragmatics, or culture, but because our understanding of how to link these types of knowledge (along with psychological states, dialectal variants, diachronic change in progress, i.e. variation, and the like) is presently not widely recognized as the problem in linguistics and the cognitive sciences more generally. This is Peirce’s problem of synechism—understanding the links between all things, that all knowledge is part of a continuum in a non-trivial sense—and it is our job to figure out how the flow of knowledge works within the continuum—imposing secondness as needed to isolate individual portions of that continuum (see Everett in progress; among many others). Serious and interesting attempts to address this issue in a formal way do exist, however (beyond efforts like Russell 2012). For example, the work by Gibbon and Griffiths (2017), Zaslavsky, Hu, and Levy (2020), Frank and Goodman (2012), all provide promising analyses and testings of the inference problem going beyond the standard considerations of compositionality found in mainstream linguistics. Everett (2016) discusses more such problems of compositionality and inference, as well as their philosophical, psychological, anthropological, computational, linguistic, and historical roots. As we approach a quarter-century into the Second Millennium CE it is occasionally disheartening to see how little progress linguistics, for example, has made in solving the inference problem—the foundational problem of the field. But one finds optimism in the knowledge that attempts are becoming more numerous.

11.5.3 Back in time After this discursus on inference we come back to time in the world’s languages and, in particular, in Pirahã. The case of time in Pirahã is, as a review of the chapters in this volume will quickly show, not all that unusual. In every temporal system there are well-specified constraints on temporal interpretation and semantics provided by the morphosyntax. In English we have John

UNDERSPECIFIED TEMPORAL SEMANTICS IN PIRAHÃ

315

is running vs John ran where the mapping is close to one-to-one between temporal morphosyntax and semantics. But we also have expressions such as John used to run, in which the habitual aspect is indicated periphrastically/idiomatically. And then we have things like Sun shinin’ I’m travelin’, in which no tense is marked, though one can interpret this as a statement of habit or future intent, depending on the context. Note that in this case no tense needs to be marked anywhere in the discourse. This works fine as an isolated utterance (in my dialect at least). Constructions like this illustrate the use of inference in English temporal semantics. The point of this chapter has been to focus on these latter types of constructions, inferentially interpreted constructions, and show that they are the most common type of temporal interpretation in languages like Pirahã. My Peircean perspective of semantic interpretations as inferences (induction, deduction, abduction) takes compositionality to be a special case of inference, in which the syntax bears a heavier burden in the determination of the meaning of the linguistic unit. To see the contrast between syntactic vs non-syntactic inference, consider two men at a bar. In the syntactic case, the two men see a drunk at the bar with his wallet hanging from his back pocket. One man says to his friend ‘You want to take that wallet?’ The other one responds ‘No’. In the non-syntactic case, two men see a wallet hanging out of the back pocket of an inebriate. One looks at the other then points to the wallet and raises his eyebrows. The other looks and shakes his head from side to side. They have communicated the same thing, the former case with words and syntax and the other without words and syntax. But both using inference. Linguists would say that the former case relies on compositionality, but in fact the cases are more similar than that. Both use inference (in Peirce’s sense or any other). Inference is also crucial for temporal systems like Piraha’s, which hold special interest for theories of understanding time in human language. As we have discussed, this is because in such languages general inference (cultural inference in particular) is the unmarked case, as opposed to the role of the special form of inference, compositionality, that seems to be the unmarked case in many, perhaps most, other languages. Again, what I mean by this is that because Pirahã chooses to severely restrict the amount of temporal information in the morphosyntax to a few time words and a limited number of aspectual morphemes, it relies more on piecing together interpretations from context than a language with more temporal specifiers in the language and the Pirahã strategy places more of the interpretational burden on inference from the non-linguistic context.

APPENDI X

Pirahã Text Xigábaı´ almost bitten by snake Told by Xahoápati to Dan Everett ca. 2000³¹

xaı´ ti ig -a -ı´ Thus. 1 COMIT. -sound -PROXIMATE ‘So I spoke at first.’

apaı´. first (literally ‘head’)

Xigábaı´ gı´ bası´ -ig -a -b Xigábaı´ 2 bed -COMIT. do. -down

-o -direct

-p -a -p -ı´ -up -vertical -up -PROXIMATE ‘Xigábaı´ was bringing your bed back up.’ Xı´s ib -á animal hit -VERTICAL

-i -ta -a -b -PROXIMATE -REPETIVE -do -down

-o -i -haı´ tigaiti. direct -transition -REL:CERT snake ‘He arrowed the snake a couple of times.’ Xı´s a -xáa -há. animal bite -ATELIC -COMP:CERT ‘The snake bit (at him)’ xaı´ Toı´toı´ hi xigı´ a -xáa Thus Toı´toı´ 3 COMIT. bit -ATELIC ‘Thus Toı´toı´ with (him) it bit at (also)’

-i -há. -PROXIMATE COMP:CERT.

hi agı´ai tigaiti gı´i xibiga á. 3. thus snake real DIRECT OBSERVATION be ‘Thus it was a real snake (a bushmaster) they saw’

³¹ This appendix is included so that the interested reader can trace temporal interpretations through a text in Pirahã to better exemplify temporal readings in context.

PIRAHÃ TE XT

tigaiti -gı´i snake -real

xaı´ xis thus animal

ib -á hit -VERTICAL

-i -PROXIMATE

-ta -b -og -aáti. repetitive -down -DES. -UNCERT/IMP. ‘It is a bushmaster. Shoot it’ Hi agı´a xogi 3 thus big xis aihi animal that:one -b -down -direct -up

-áaga -ó -be -DIRECTIONAL og -a -á want -find -INTENSE -ó -p -a -REMOTE -up -COMP.CERT

-a -VERTICAL -p -á. -up -COMP.CERT.

xaı´ tigaiti. thus snake ‘Thus the bulk of the people wanted to search carefully down on the ground. Thus (for that) snake’ Ti agı´a xis 1 thus animal

igı´ -o COMIT. -DIRECTIONAL

k -á -ab -ó -p -á OBJECT -INTENSELY -bite -DIRECT -up -VERTICAL -p -á -xaı´. -up -VERTICAL -do ‘(When I was) with the animal, (it) struck up (at me)’ Ti ig áı´ -sai. 1 COMIT. Do -OLD:INFO xi ab -áo -p -ı´ -saxáa 3 bite -TELIC -up -PROXIMATE -FRUST.:EVENT -g -abaga -á. -DES. -FRUST.INI. -COMP.CERT. ‘I spoke (as I have been doing). It almost bit him, almost wanted to’

317

318

DANIEL L. EVERETT

Xiabáı´si k human:being OBJECT

-á -INTENSELY

-ab -o -bite -MOVEMENT

-ó -DIRECTIONAL

-p -a -i -hiabi -VERTICAL -REMOTE -TRANSITION VOWEL -NEGATIVE -só -(a)i. xáihá. -telic -do thus. ‘Thus, it (the bushmaster) didn‘t bite the human being (upwards and viciously)’ xi 3

go -ó -i xa -ó focus -LOC. -PROXIMATE bite -DIRECTION

-o -DIRECT

-p -á -á -ha -xaı´. -up -INTENSE -VERTICAL -COMP:CERT -do ‘He was there (almost) bitten (It was there that he was almost bitten)’ Xi go -ó -i 3 FOCUS -LOC. -PROXIMATE -p -á. -up -INTENSE ‘He was there (almost) bitten’

xa -ó bite -DIRECTION

-o -DIRECT

xoı´ hi aı´ -si -xı´ga. environment 3 is -OLD:INFO -EMPHATIC ‘It was there in the jungle’ Xigábaı´ hi aa -b -áo Xigábaı´ 3 bite -down -TELIC

-b -PERFECTIVE

-ı´i -sa -xá -abaı´. -INTENTIVE -OLD:INFO -INTENSIVE -FRUST.COMPLETION ‘(The snake) almost bit Xigábaı´, as we have been saying’ Xigábaı´ hi áa -b -o Xigábaı´ 3 bite -down -DIRECT -ái -hiab -do -NEGATIVE

-ó -LOC.

-sói -xáihi. -DOUBT -DECLARATIVE

quase medo almost afraid (Portuguese) ‘(The snake) barely didn’t bite Xigábaı´’

-p -up

References Abusch, Dorit (2012). Circumstantial and temporal dependence in counterfactual modals. Natural Language Semantics 20: 273–297. Aiken, Conrad (1920). The House of Dust. Boston: The Four Seas Company. Aikhenvald, Alexandra Y. (2012). The Languages of the Amazon. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. (2018). Me, us, and others: Expressing the self in Arawak languages of South America, with a focus on Tariana. In Minyao Huang and Kasia M. Jaszczolt (eds), Expressing the Self: Cultural Diversity and Cognitive Universals. Oxford: Oxford University Press, 13–39. Albert, David. Z. (2000). Time and Chance. Massachusetts: Harvard University Press. Alexandrova, Anna (2016). Avertive constructions in Europe and North Asia: An areal typology. Paper presented at the Chronos 12 (12th International Conference on Actionality, Tense, Aspect, Modality/Evidentiality), Université Caen Normandie. Alpher, Barry, Nicholas Evans, and Mark Harvey (2003). Proto-Gunwinyguan verb suffixes. In Nicholas Evans (ed.), The Non-Pama-Nyungan Languages of Northern Australia: Comparative Studies of the Continent’s Most Linguistically Complex Region. Canberra: Pacific Linguistics, 305–352. Altshuler, Daniel (2014). A typology of partitive aspectual operators. Natural Language & Linguistic Theory 32(3): 735–775. Alverson, Hoyt (1994). Semantics and Experience: Universal Metaphors of Time in English, Mandarin, Hindi, and Sesotho. Baltimore: Johns Hopkins University Press. Amaral, Luiz, Marcus Maia, Andrew Nevins, and Tom Roeper (eds) (2018). Recursion Across Domains. Cambridge: Cambridge University Press. Ander Bois, Scott, and Maksymilian Dąbkowski (2021). A’ingae =sa’ne “APPR” and the semantic typology of apprehensional adjuncts. Semantics and Linguistic Theory 30: 43–62. Andersen, Holly (2014). The development of the specious present and James’ views on temporal experience. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press, 25–42. Anscombe, Gertrude E. M. (1959). An Introduction to Wittgenstein’s Tractatus. London: Hutchinson. Arkadiev, Peter (2019). The Lithuanian “buvo + be-present active participle” construction revisited. Baltic Linguistics 10: 65–108. Arkadiev, Peter (2020). Grammaticalization by semantic enrichment: from progressive to proximative to avertive in Lithuanian. Paper presented at the SLE 2020, Online conference. Arregui, Ana (2007). When aspect matters: the case of would-conditionals. Natural Language Semantics 15(3): 221–264. Arregui, Ana (2009). On similarity in counterfactuals. Linguistics and Philosophy 32(3): 245–278.

320

REFERENCES

Arregui, Ana, Marı´a Luisa Rivero, and Andrés Salanova (2014). Cross-linguistic variation in imperfectivity. Natural Language & Linguistic Theory 32(2): 307–362. Arstila, Valtteri (2015a). Defense of the brain time view. Frontiers in Psychology 6: 1350. DOI: 10.3389/fpsyg.2015.01350. Arstila Valtteri (2015b). Theories of apparent motion. Phenomenology and the Cognitive Sciences 15(3): 337–358. Arstila, Valtteri (2018). Temporal experiences without the specious present. Australasian Journal of Philosophy 96(2): 287–302. Arstila, Valtteri, and Dan Lloyd (eds) (2014a). Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press. Arstila, Valtteri, and Dan Lloyd (2014b). Subjective time from past to future. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press, 309–322. Arunachalam, Sudha, and Anubha Kothari (2010). Telicity and event culmination in Hindi perfectives. In Pier Marco Bertinetto, Anna Korhonen, Alessandro Lenci, Alissa Melinger, Sabine Schulte im Walde, and Aline Villavicencio (eds), Proceedings of Verb 2010: Interdisciplinary Workshop on Verbs. The Identification and Representation of Verb Features. Pisa: Scuola Normale Superiore, 16–19. Arunachalam, Sudha, and Anubha Kothari (2011). An experimental study of Hindi and English perfective interpretation. Journal of South Asian Linguistics 4(1): 27–42. Asher, Nicholas (2011). Lexical Meaning in Context: A Web of Words. Cambridge: Cambridge University Press. Atlas, Jay D. (1977). Negation, ambiguity, and presupposition. Linguistics and Philosophy 1: 321–336. Atlas, Jay D. (1979). How linguistics matters to philosophy: Presupposition, truth, and meaning. In Choon-Kyu Oh and David A. Dinneen (eds), Syntax and Semantics, vol. 11. New York: Academic Press, 265–281. Atlas, Jay D. (1989). Philosophy Without Ambiguity: A Logico-Linguistic Essay. Oxford: Clarendon Press. Austin, John L. (1962). How to Do Things with Words. Oxford: Oxford University Press. Auwera, Johan van der, and Vladimir Plungian (1998). Modality’s semantic map. Linguistic Typology 2(1): 79–124. Bachelard, Gaston (1993 [1936]). La Dialectique de la Durée. Paris: Presses Universitaires de France. Bachelard, Gaston (1994 [1932]). L’Intuition de l’Instant. Paris: Le Livre de Poche. Baranzini, Laura, and Louis de Saussure (2017). Le futur épistémique en italien: de la modalité à l’évidentialité. In Laura Baranzini (ed.), Le Futur dans les Langues Romanes. Berne: Peter Lang, 299–316. Bar-el, Leora (2005). Aspectual Distinctions in Skkwxxwú7mesh. PhD dissertation, University of British Columbia. Bar-El, Leora, Henry Davis, and Lisa Matthewson (2006). On Non-Culminating Accomplishments. In Leah Bateman and Cherlon Ussery (eds), Proceedings of the Thirty-Fifth Annual Meeting of the North East Linguistic Society (NELS 35), vol. 1. Amherst, Mass.: GLSA (Graduate Linguistic Student Association), Department of Linguistics, South College, University of Massachusetts, 87–102. Barham, Larry, and Daniel Everett (2020). Semiotics and the origin of language in the Lower Paleolithic. Journal of Archaeological Method and Theory. DOI: 10.1007/s10816020-09480-9

REFERENCES

321

Bar-Hillel, Yehoshua (1964). A demonstration of the nonfeasibility of fully automatic high quality translation. In Language and Information: Selected Essays on their Theory and Application, 2nd edn. Reading, Mass.: Addison-Wesley, 174–179. Barnes, Elizabeth, and Ross Cameron (2009). The open future: bivalence, determinism and ontology. Philosophical Studies 146: 291–309. Barnes, Elizabeth, and Ross Cameron (2011). Back to the open future. Philosophical Perspectives 25: 1–26. Baron, Sam (2017). Feel the flow. Synthese 194: 609–630. Baron, Sam (2019). The curious case of spacetime emergence. Philosophical Studies 177: 2207–2226. Baroni, Marco (2020). Linguistic generalization and compositionality in modern artificial neural networks. Philosophical Transactions of the Royal Society B: Biological Sciences, 375(1791): 20190307. DOI: 10.1098/rstb.2019.0307 Bednall, James (2020). Temporal, aspectual and modal expression in Anindilyakwa, the language of the Groote Eylandt archipelago, Australia. PhD dissertation, ANU and Université de Paris-Diderot. Belnap, Nuel, Michael Perloff, and Ming Xu (2001). Facing the Future: Agents and Choices in our Indeterminist World. Oxford: Oxford University Press. Bender, Andrea, and Sieghard Beller (2014). Mapping spatial frames of reference onto time: A review of theoretical accounts and empirical findings. Cognition 132: 342–382. Bergen, Benjamin (2005). Mental simulation in literal and figurative language understanding. In Seana Coulson and Barbara Lewandowska-Tomaszczyk (eds), The Literal and Non-literal in Language and Thought. Berlin: Peter Lang, 255–280. Bergen, Benjamin (2012). Louder than Words: The New Science of How the Mind Makes Meaning. New York: Basic Books. Bergson, Henri (1960 [1889]). Time and Free Will: An Essay on the Immediate Data of Consciousness. Translated by Frank L. Pogson. New York: Harper and Row. Bernard, Timothée, and Lucas Champollion (2018). Negative events in compositional semantics. Semantics and Linguistic Theory 28: 512–532. Bhatt, Rajesh (1999). Covert Modality in Non-finite Contexts. PhD dissertation, University of Pennsylvania. Bienenstock, Elie, Stuart Geman, and David Potter (1996). Compositionality, MDL priors, and object recognition. Proceedings of the Conference on Neural Information Processing: 838–844. Bittner, Maria (2005). Future discourse in a tenseless language. Journal of Semantics 22: 339–387. Blackburn, Patrick, and Klaus Frovin Jørgensen (2016) Reichenbach, Prior and hybrid tense logic, Synthese 193: 3677–3689. Blake, Barry J. (1979a). A Kalkatungu Grammar (Pacific linguistics: Series B, Monographs), vol. 57. Canberra: Pacific Linguistics, Australian National University. Blake, Barry J. (1979b). Pitta-Pitta. In Robert M. W. Dixon and Barry J. Blake (eds), Handbook of Australian Languages, vol. 1. Canberra: Australian National University Press, 183–242. Block, Richard A. (1982). Temporal judgments and contextual change. Journal of Experimental Psychology: Learning, Memory, and Cognition 8: 530–544. Block, Richard A. (1990). Models of psychological time. In Richard A. Block (ed.), Cognitive Models of Psychological Time. Hillsdale, NJ: Lawrence Erlbaum, 1–36. Boccardi, Emiliano, and Federico Perelda (2018). The delusive illusion of passage. Analysis 78: 387–396.

322

REFERENCES

Boghossian, Paul A., and David J. Velleman (1989). Color as secondary quality. Mind 98(389): 81–103. Bohnemeyer, Ju¨rgen (2002). The Grammar of Time Reference in Yukatek Maya. Mu¨nchen: Lincom Europa. Bordini, Davide, and Giuliano Torrengo (2022). Frightening times. European Journal of Philosophy 30(1): 293–306. Bourne, Craig (2002). When Am I? A tense time for some tense theorists? Australiasian Journal of Philosophy, 80: 359–371. Bourne, Craig (2006). A Future for Presentism. Oxford: Oxford University Press. Bourova, Viara (2005). A la recherche du ‘conditionnel latin’: Les constructions ‘Infinitif + forme habere’ examinées à partir d’un corpus électronique. In Claus D. Pusch, Johannes Kabatek, and Wolfgang Raible (eds), Romanistische Korpuslinguistik II. Tu¨bingen: Naar, 303–316. Bourova, Viara, and Liliane Tasmowski (2007). La préhistoire des futurs romans: Ordre des constituants et sémantique. In Louis de Saussure, Jacques Moeschler and Genoveva Puskas (eds), Études sémantiques et pragmatiques sur le temps, l’aspect et la modalité (Cahiers Chronos 19), 25–41. Amsterdam/Atlanta: Rodopi. Braddon-Mitchell, David (2004). How do we know it is now now? Analysis 64: 199–203. Brandom, Robert B. (1998). Making it Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, Mass.: Harvard University Press. Briggs, Rachael, and Graeme A. Forbes (2012). The real truth about the unreal future. In Karen Bennett and Dean W. Zimmerman (eds), Oxford Studies in Metaphysics, vol. 7. Oxford: Oxford University Press, 257–304. Briggs, Rachael, and Graeme A. Forbes (2019). The future, and what might have been. Philosophical Studies 176: 505–532. Broad, C.D. (1959). A reply to my critics. In Paul Arthur Schilpp (ed.), The Philosophy of C. D. Broad. New York: Tudor Publishing Company. 741–764. Brown, Jessica (2000). Against temporal externalism. Analysis 60(2): 178–188. Bybee, Joan, Revere Perkins, and William Pagliuca (1994). The Evolution of Grammar: Tense, Aspect and Modality in the Language of the World. Chicago: University of Chicago Press. Callender, Craig (2017). What Makes Time Special? Oxford: Oxford University Press. Cameron, Ross (2015). The Moving Spotlight: An Essay on Time and Ontology. Oxford: Oxford University Press. Campbell, John (1994). Past, Space and Self. Cambridge, Mass./London: MIT Press/A Bradford Book. Campbell, Lyle (2012). Typological characteristics of South American indigenous languages. In Lyle Campbell and Verónica Grondona (eds), The Indigenous Languages of South America. Berlin/Boston: Mouton De Gruyter, 259–330. Cappelen, Herman, and Josh Dever (2013). The Inessential Indexical: On the Philosophical Insignificance of Perspective and the First Person. Oxford: Oxford University Press. Carol, Javier, and Andrés Pablo Salanova (2017). Frustratives and aspect. Paper presented at the Symposium on Amazonian Languages II (SAL 2), UC Berkeley. Casasanto, Daniel (2016). Temporal language and temporal thinking may not go hand in hand. In Barbara Lewandowska-Tomaszczyk (ed.), Conceptualizations of Time. Amsterdam: John Benjamins Publishing Company, 169–186. Casasanto, Daniel, and Kyle Jasmin (2012). The Hands of Time: Temporal Gesture in English Speakers. Cognitive Linguistics 23(4): 643–674.

REFERENCES

323

Caudal, Patrick (2005). Stage structure and stage salience for event semantics. In Paula Kempchinsky and Roumyana Slabakova (eds), Aspectual Inquiries. Dordrecht: Springer, 239–264. Caudal, Patrick (2011). Contribution aspectuelle des temps et de la phrase: des affinités électives? In Sylvie Hancil (ed.), Perspectives théoriques et empiriques sur l’aspect en anglais. Rouen: Publications des Universités de Rouen et du Havre, 11–52. Caudal, Patrick (2012). Pragmatics. In Robert Binnick (ed.), The Oxford Handbook of Tense and Aspect. Oxford/New York: Oxford University Press, 269–305. Caudal, Patrick (2018a). De la théorie du sens, à celle des appariements formes/sens: synthèse de quinze ans de recherche sur le TAM(E). Université de Paris 7, Mémoire d’habilitation à diriger les recherches. Caudal, Patrick (2018b). Demodality: profiling a novel category at the tense/aspect – modality divide. Presented at the 13th International Conference on Tense, Aspect, Modality and Evidentiality (Chronos 13), Université de Neuchˆatel. Caudal, Patrick (2020a). From ‘actuality entailments’ to avertivity: on some postmodal meanings in French. Paper presented at the ALS 2020 (Conference of the Australian Linguistic Society), Sydney. https://als.asn.au/Conference/Conference2020 Caudal, Patrick (2020b). Coercion for the ages? A thousand years of parallel inchoative histories for the French passé simple and passé composé. In Martı´n Fuchs and Joshua Philipps (eds), Proceedings of Linguistic Society of America, vol. 5. Washington, DC: Linguistic Society of America, 51–66. Caudal, Patrick, and Eva Schultze-Berndt (2016). Unusual possibilities? Exploring the Jaminjung modal system. Paper presented at the New Ideas in Semantics and Modeling 2016 (NSIM 2016), EHESS, Paris. Caudal, Patrick, and Robert Mailhammer (2021). Tensed periphrastic vs. synthetic modal inflections in Iwaidja – A novel insight into grammaticalization cycles for modality in Northern Australia? Paper presented at Evidentiality and Modality: At the crossroads of grammar and lexicon, Montpellier, 10-11 June. Caudal, Patrick & Robert Mailhammer. (2022). Linear lengthening in Iwaidja: An eventquantifying intonation at the phonology to semantics/pragmatics interface. Languages. Multidisciplinary Digital Publishing Institute 7(3). 209–231. Caudal, Patrick, Robert Mailhammer, and James Bednall (2019). A comparative account of the Iwaidja and Anindilyakwa modal systems. Paper presented at the ALW2019 (Australian Languages Workshop), Camp Marysville, VIC. Chadha, Monima (2019). Reconstructing memories, deconstructing the self. Mind and Language 34: 121–138. Chalmers, David J. (2006). The foundations of two-dimensional semantics. In Manuel Garcia-Carpintero and J. Macia (eds), Two-Dimensional Semantics: Foundations and Applications. Oxford: Oxford University Press, 55–140. Chang, Hasok (2012). Is Water H2O? Evidence, Realism and Pluralism. New York: Springer. Chávez, Margarethe (2003). I Want to But I Can’t: the Frustrative in Amahuaca. Dallas/Arlington,TX: SIL. Chomsky, Noam (2009). Cartesian Linguistics, 3rd edn. Cambridge: Cambridge University Press. Chomsky, Noam (2016). What Kind of Creatures are We? New York: Columbia University Press. Chumakina, Marina (2013). Periphrasis in Archi. In Marina Chumakina and Greville G. Corbett (eds), Periphrasis: The Role of Syntax and Morphology in Paradigms. Oxford/New York, NY: British Academy/Oxford University Press, 27–52.

324

REFERENCES

Cinque, Guglielmo (1999). Adverbs and Functional Heads: A Crosslinguistic Perspective. Oxford: Oxford University Press. Clark, Herbert H. (1973). Space, Time, Semantics, and the Child. In Timothy. E. Moore (ed.), Cognitive Development and the Acquisition of Language. New York: Academic Press, 28–63. Clendon, Mark (2014). Worrorra: A Language of the North-West Kimberley Coast. Adelaide: University of Adelaide Press. Collins, J.M. (2006). Temporal Externalism, Natural Kind Terms and Scientifically Ignorant Communities. Philosophical Papers 35(1): 55–68. Comrie, Bernard (1976). Aspect. Cambridge: Cambridge University Press Comrie, Bernard (1981). On Reichenbach’s approach to tense, Communication and Linguistics Studies 17: 24–30. Comrie, Bernard (1985). Tense. Cambridge: Cambridge University Press Condoravdi, Cleo (2002). Temporal interpretation of modals - Modals for the present and for the past. In David. I. Beaver, Luis D. Casillas Martı´nez, Brady. Z. Clark, and Stefan Kaufmann (eds), The Construction of Meaning. Stanford, CA.: CSLI, 59–88. Condoravdi, Cleo, and Sven Lauer (2012). Imperatives: Meaning and illocutionary force. In Christopher Piñón (ed.), Empirical Issues in Syntax and Semantics 9, 37–58. http:// www.cssp.cnrs.fr/eiss9/eiss9.pdf Condoravdi, Cleo, and Sven Lauer (2016). Anankastic conditionals are just conditionals. Semantics and Pragmatics 9(8): 1–69. Copley, Bridget (2005). Ordering and reasoning. In Jon Gajewski, Valentine Hacquard, Bernard Nickel, and Seth Yalcin (eds), New Work on Modality, vol. 51. Cambridge, Mass.: MITWPL, 120–151. Copley, Bridget (2007). When the actual world isn’t inertial: Tohono O’odham ‘cem’. In Michael Becker and Andrew McKenzie (eds), Proceedings of SULA 3: The Semantics of Under-Represented Languages in the Americas. Amherst, Mass.: University of Massachusetts, 1–18. Copley, Bridget, and Heidi Harley (2014). Eliminating causative entailments with the force-theoretic framework: The case of the Tohono O’odham frustrative cem. In Bridget Copley and Fabienne Martin (eds), Causation in Grammatical Structures. Oxford: Oxford University Press, 120–151. Coseriu, Eugenio (1976). Das romanische Verbalsystem. Tu¨bingen: TBL-Verlag Narr. Cossins, Daniel (2019). The time delusion. New Scientist 3237: 32–36. Creissels, Denis, Gerrit J. Dimmendaal, Zygmunt Frajzyngier, and Christa Ko¨nig (2007). Africa as a morphosyntactic area. In Bernd Heine and Derek Nurse (eds), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, 86–150. Csikszentmihalyi, Mihaly (1990). Flow: The Psychology of Optimal Experience. New York: Harper Collins. Dainton, Barry (2000). Stream of Consciousness: Unity and Continuity in Conscious Experience. London: Routledge. Dainton, Barry (2008). Sensing Change. Philosophical Issues 18: 362–384 Danto, Arthur Coleman (1985). Narration and Knowledge. New York, NY: Columbia University Press. Davidson, Donald (1967). The logical form of action sentences. In Nicholas Rescher (ed.), The Logic of Decision and Action. Pittsburgh, PA: University of Pittsburgh Press, 81–95. Davis, Henry, and Lisa Matthewson (2016). Against all expectations: The meaning of St’át’imcets séna7. In Marianne Huijsmans, Thomas J. Heins, Oksana Tkachman, and Natalie Weber (eds), Papers for ICSNL 51: The Fifty-First International Conference on

REFERENCES

325

Salish and Neighbouring Languages (UBCWPL 42). Vancouver: University of British Columbia, 37–67. Deasy, Daniel (2017). What is presentism? Noûs 51: 378–397. Declerck, Renaat (1991). Tense in English. London: Routledge. Delancey, Scott (2012). Still mirative after all these years. Linguistic Typology 16(3): 529–564. Dendale, Patrick (1994). Devoir épistémique, marqueur modal ou évidentiel? Langue française 102: 24–40. Dendale, Patrick (1999). Devoir au conditionnel: valeur évidentio-modale et origine du conditionnel. In Svetlana Vogeleer, Andrée Borillo, Marcel Vuillaume and Carl Vetters (eds), La modalité sous tous ses aspects (Cahiers Chronos 4), 7–28. Amsterdam/Atlanta: Rodopi. Deng, Natalja (2013a). On explaining why time seems to pass. The Southern Journal of Philosophy 51(3): 367–382. Deng, Natalja (2013b). Our experience of passage on the B-theory. Erkenntnis 78(4): 713–726. Deng, Natalja (2019). One thing after another: Why the passage of time is not an illusion. In Valtteri Arstila, Adrian Bardon, Sean Power, and Argiro Vatakis (eds), The Illusions of Time: Philosophical and Psychological Essays on Timing and Time Perception. Basingstoke: Palgrave Macmillan, 3–16. Dennett, Daniel (1993). Consciousness Explained. London: Penguin. Dennett, Daniel, and Marcel Kinsbourne (1992). Time and the observer: The where and when of consciousness in the brain. The Behavioral and Brain Sciences 15(2): 183–201. Detges, Ulrich, and Richard Waltereit (2002). Grammaticalization vs. reanalysis: A semantic-pragmatic account of functional change in grammar. Zeitschrift fu¨r Sprachwissenschaft 21(2): 151–195. Dickinson, Emily (1998). Crumbling is not an Instant’s Act. In The Poems of Emily Dickinson: Variorum Edition. Cambridge, Mass.: The Belknap Press of Harvard University Press. Dobson, Henry Austin (1923). The Paradox of Time. In The Complete Poetical Works of Henry Austin Dobson. Oxford: Oxford University Press. Donnellan, Keith (1983). Kripke and Putnam on natural kind terms. In Carl Ginet and Sydney Shoemaker (eds), Knowledge and Mind: Philosophical Essays. Oxford: Oxford University Press, 84–104. Dowe, Phil (2000). Physical Causation. Cambridge: Cambridge University Press. Dowty, David (1979). Word Meaning and Montague Grammar: The Semantics of Verbs and Times in Generative Semantics and Montague’s PTQ. Dordrecht: Reidel. Dowty, David (1986). The effects of aspectual class on the temporal structure of discourse: Semantics or Pragmatics? Linguistics and Philosophy 9(1): 37–61. Dowty, David (2007). Compositionality as an Empirical Problem. In Chris Barker and Pauline I. Jacobson (eds), Direct Compositionality. Oxford: Oxford University Press, 14–23. Droit-Volet, Sylvie (2013). Time perception, emotions and mood disorders. Journal of Physiology, Paris 107(4): 255–264. Droit-Volet, Sylvie (2014). What emotions tell us about time. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press, 477–506. Droit-Volet, Sylvie, and John H. Wearden (2016). Passage of time judgments are not duration judgments: Evidence from a study using experience sampling methodology. Frontiers in Psychology 7: 176. DOI: 10.3389/fpsyg.2016.00176

326

REFERENCES

Droit-Volet, Sylvie, and Sandrine Gil (2009). The time-emotion paradox. Philosophical Transactions of The Royal Society B: Biological Sciences 364: 1943–1953. Dummett, Michael (1969). The reality of the past. Proceedings of the Aristotelian Society 69: 239–258. Dyke, Heather (2002). Tokens, dates and tenseless truth conditions. Synthese, 131: 329–351. Dyke, Heather (2008). Metaphysics and the Representational Fallacy. New York: Routledge. Earman, John (1969). The anisotropy of time. Australasian Journal of Philosophy 47(3): 273–295. Earman, John (1970). Space-time, or how to solve philosophical problems and dissolve philosophical muddles without really trying. The Journal of Philosophy 67(9): 259–277. Earman, John (1974). An attempt to add a little direction to ‘the problem of the direction of time’. Philosophy of Science 41(1): 15–47. Eather, Bronwyn (2011). A Grammar of Nakkara (Central Arnhem Land Coast). Mu¨nchen: Lincom Europa. Einstein, Albert (1916) Relativity: The Special and General Theory. London: Methuen & Co. Ltd. El Refaie, Elisabeth (2019). Visual Metaphor and Embodiment in Graphic Illness Narratives. New York: Oxford University Press. Elder, Chi-Hé and Michael Haugh (2018). The interactional achievement of speaker meaning: Toward a formal account of conversational inference. Intercultural Pragmatics 15: 593–625. Elias, Norbert (1992). Time: An Essay. Translated by Edmund Jephcott. Oxford: Blackwell. Eliot, Thomas Stearns (2004a). Four Quartets. In The Complete Poems and Plays of T. S. Eliot. London: Faber & Faber, 117–138. Eliot, Thomas Stearns (2004b). The Waste Land. In The Complete Poems and Plays of T. S. Eliot. London: Faber & Faber, 37–50. Elliott, Jennifer R. (2000). Realis and irrealis: Forms and concepts of the grammaticalisation of reality. Linguistic Typology 4(1): 55–90. Erelt, Mati, and Helle Metslang (2009). Some notes on proximative and avertive in Estonian. Linguistica Uralica 45(3): 178–192. Evans, Nicholas. (1985). Kayardild - The language of the Bentinck Islanders of North West Queensland. Canberra: Australian National University PhD Thesis. Evans, Nicholas (1995). A Grammar of Kayardild with Historical-Comparative Notes on Tangkic. Berlin: Mouton de Gruyter. Evans, Nicholas (2003). Bininj Gun-wok: A Pan-dialectal Grammar of Mayali, Kunwinjku and Kune. Canberra: Australian National University. Evans, Nicholas, and Honoré Watanabe (eds) (2016). Insubordination. Amsterdam: John Benjamins Publishing Company. Evans, Nicholas, and Stephen C. Levinson (2009). The myth of language universals: Language diversity and its importance for cognitive science. Behavioral and Brain Sciences 32: 429–492. Evans, Vyvyan (2003). The Structure of Time: Language, Meaning and Temporal Cognition. Amsterdam: John Benjamins Publishing Company. Everett, Daniel L. (1983). A Lingua Pirahã e a Teoria da Sintaxe. Sc.D. dissertation, Universidade Estadual de Campinas. Everett, Daniel L. (1986). Pirahã. In Desmond Derbyshire and Geoffrey Pullum (eds), Handbook of Amazonian Languages, vol. 1. Berlin: Mouton de Gruyter, 200–325. Everett, Daniel L. (1993). Sapir, Reichenbach, and the Syntax of Tense in Pirahã. Pragmatics & Cognition 1(1): 89–124.

REFERENCES

327

Everett, Daniel L. (2005). Cultural constraints on grammar and cognition in Pirahã: Another look at the design features of human language. Current Anthropology 76(4): 621–664. Everett, Daniel L. (2008). Don’t Sleep, There Are Snakes: Life and Language in the Amazonian Jungle. London: Profile. Everett, Daniel L. (2009a). Pirahã culture and grammar: A reply to some criticisms. Language, 85(2): 405–442. Everett, Daniel L. (2009b). Wari’ intentional state construction predicates. In Robert Van Valin (ed.), Investigations of the Syntax-Semantics-Pragmatics Interface. Amsterdam: John Benjamins Publishing Company: 381–409. Everett, Daniel L. (2010). You drink. You drive. You go to jail. Where’s recursion? Lingbuzz. https://ling.auf.net/lingbuzz/001141 Everett, Daniel L. (2012a). Language: The Cultural Tool. New York: Pantheon Books. Everett, Daniel L. (2012b). What does Pirahã have to teach us about human language and the mind? Wiley Interdisciplinary Reviews. Cognitive Science 3(6): 555–563. Everett, Daniel L. (2016). Dark Matter of the Mind: The Culturally Articulated Unconscious. Chicago: University of Chicago Press. Everett, Daniel L. (2017). How Language Began: The Story of Humanity’s Greatest Invention. London: Profile. Everett, Daniel L. (2019). The American Aristotle. Aeon Magazine. https://aeon.co/essays/ charles-sanders-peirce-was-americas-greatest-thinker Everett, Daniel L. (2020). O Papel da Cultura na Lı´ngua e na Cognição. In Miguel Oliveira Jr. (ed.), 50th Anniversary Edition of the Proceedings of ABRALIN (Brazilian Linguistics Association). São Paulo: ABRALIN Press, 1–37. Everett, Daniel L. (in progress). Peircean Linguistics: A Chapter in the History of Empiricist Thought. Oxford: Oxford University Press. Everett, Daniel L., and Barbara Kern (1997). Wari’: The Pacaas Novos Language of Western Brazil. London: Routledge. Everett, Daniel L., and Edward Gibson (2019). Review of Recursion Across Domains by Luis Amaral et al. (2018). Language 95(4): 777–790. Everett, Daniel L., and Edward Gibson (2020). Semantics requires knowledge of meaning: A reply to Rodrigues. Lingbuzz. https://ling.auf.net/lingbuzz/005060 Farr, Matt (2012a). On A- and B-theoretic elements of branching spacetimes. Synthese 188: 85–116. Farr, Matt (2012b). Towards a C Theory of Time: An Appraisal of the Physics and Metaphysics of Time Direction. Ph. D. thesis, University of Bristol. Farr, Matt (2020a). C-theories of time: On the adirectionality of time. Philosophy Compass 15(12): 1–17. Farr, Matt (2020b). Causation and time reversal. The British Journal for the Philosophy of Science 71(1): 177–204. Farr, Matt (2020c). Explaining temporal qualia. European Journal for Philosophy of Science 10(8). DOI: 10.1007/s13194-019-0264-6 (online) Farr, Matt (2022a). Conventionalism about time direction. Synthese 200. DOI: 10.1007/s11229-022-03540-1 (online) Farr, Matt (2022b). The C-theory of time. Unpublished ms. Fauconnier, Gilles, and Mark Turner (2008). Rethinking Metaphor. In Raymond W. Gibbs (ed.), The Cambridge Handbook of Metaphor and Thought. Cambridge: Cambridge University Press, 57–66.

328

REFERENCES

Ferreira, Marcelo (2014). Displaced aspect in counterfactuals: Towards a more unified theory of imperfectivity. In Luka Crnicˇ and Uli Sauerland (eds), The Art and Craft of Semantics, vol. 1. Cambridge, Mass.: MITWPL, 147–164. Filip, Hana (2017). The semantics of perfectivity. The Italian Journal of Linguistics 29: 167–200. Fine, Kit (2006). The reality of tense. Synthese 150: 399–414. Fischer, Luke (2015). The Poet as Phenomenologist: Rilke and the New Poems. New York: Bloomsbury. Fischer, Raymond (1984). An aged man remembers April. Poetry (Chicago), 144(1): 39. Fischoff, Baruch (1975). Hindsight ≠ foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1: 288–299. Flaherty, Michael (1999). A Watched Pot: How We Experience Time. New York: New York University Press. Flaherty, Michael G. (2011). The Textures of Time: Agency and Temporal Experience. Philadelphia: Temple University Press. Flaherty, Michael G. (2018). An S-shaped pattern in the perceived passage of time: How social interaction governs temporal experience. Language and Cognition 10: 1–25. Floyd, Simeon (2016). Modally hybrid grammar? Celestial pointing for time-of-day reference in Nheengatu, Language, 92(1): 31–64. Fodor, Jerry A., and Ernest Lepore (2002). The Compositionality Papers. Oxford: Oxford University Press. Foley, William A. (1991). The Yimas Language of Papua New Guinea. Stanford: Stanford University Press. Forbes, Graeme A. (2016) The Growing Block’s past problems. Philosophical Studies 173: 699–709. Forbes, Graeme A., and Nathan Wildman (2022). Enduring senses. Synthese 200(291):1–21. Ford, Lysbeth and D. McCormack (2007). Murrinh tetemanthay ngarra murrinh law kardu bamam thanguna: Murrinhpatha – English Legal Glossary. http://www.bowdenmccormack.com.au Forrest, Peter (2004). The real but dead past: A reply to Braddon-Mitchell. Analysis: 64(4): 358–362. Fraisse, Paul (1963). The Psychology of Time. New York: Harper & Row. Frank, Michael C., and Noah D. Goodman (2012). Predicting pragmatic reasoning in language games. Science 336: 998. Frankish, Keith (2016). Illusionism as a theory of consciousness. Journal of Consciousness Studies 23(11-12): 11–39. Frege, Gottlob (1956). The thought: A logical inquiry. Mind 65(259): 289–311. Frege, Gottlob (1984). Collected Papers on Mathematics, Logic, and Philosophy. Edited and translated by Brian McGuiness. Oxford: Blackwell. Frege, Gottlob (1984 [1892]). On concept and object. In Translations from the Philosophical Writings of Gottlob Frege. Translated by Peter Geach. Oxford: Blackwell: 182–194. Futrell, Richard, Laura Stearns, Steven T. Piantadosi, Daniel Everett, and Edward Gibson (2016). A corpus investigation of syntactic embedding in Pirahã’, PLOS One, March 2. Galton, Antony (2011). Time flies but space does not: Limits to the spatialisation of time. Journal of Pragmatics 43: 695–703. Garcı´a Salido, Gabriela (2014). Clause Linkage in Southeastern Tepehuan: A Uto-Aztecan Language of Northern Mexico. PhD dissertation, University of Texas.

REFERENCES

329

Garrett, Brian (1988). ‘Thank goodness that’s over’ Revisited. The Philosophical Quarterly, 38: 201–205. Gayral, Françoise, Daniel Kayser and François Lévy (2005). Challenging the principle of compositionality in interpreting natural language texts. In Edouard Machery, Markus Werning, and Gerhard Schurz (eds), The Compositionality of Meaning and Content, vol 2. Frankfurt: Ontos Verlag, 83–105. Gepshtein, Sergei, and Kubovy Michael (2007). The lawful perception of apparent motion. Journal of Vision, 7(8–9): 1–15. Gibbon, Dafydd, and Sascha Griffiths (2017). Multilinear grammar: Ranks and interpretations. Open Linguistics, 3: 265–307. Gibbs, Raymond W., and Ana Cristina Pelosi Silva de Macedo (2010). Metaphor and embodied cognition. DELTA Documentação de Estudos em Lingu¨´ıstica Teórica e Aplicada 26: 679–700. Gibbs, Raymond W. (2006). Metaphor as embodied simulation. Mind and Language 21(3): 434–458. Gibbs, Raymond W. (2017). Metaphor Wars: Conceptual Metaphors in Human Life. Cambridge: Cambridge University Press. Gil, David. (2008). How much grammar does it take to sail a boat? (or, what can material artefacts tell us about the evolution of language?). In Andrew D. M. Smith, Kenny Smith and Ramon Ferrer i. Cancho (eds), The Evolution of Language. New Jersey: World Scientific, 123–130. Gil, Sandrine, and Sylvie Droit-Volet (2009). Time perception, depression and sadness. Behavioural Processes 80: 169–176. Giorgi, Alessandra (2010). About the Speaker: Towards a Syntax of Indexicality. Oxford: Oxford University Press. Givon, Tom (2020). Coherence. Amsterdam/Philadelphia,PA: John Benjamins Publishing Company. Goddard, Cliff (1983). A Semantically-Oriented Grammar of the Yankunytjatjara Dialect of the Western Desert Language. PhD dissertation, Australian National University. Goddard, Cliff (1985). A grammar of Yankunytjatjara. Alice Springs, NT: Institute for Aboriginal Development. Gold, Thomas (1966). Cosmic processes and the nature of time. In Robert Garland Colodny (ed.), Mind and Cosmos. Pittsburgh, PA: University of Pittsburgh Press, 311–329. Gosselin, Laurent (2010). Les modalités en français: la validation des représentations. Amsterdam/New York, NY: Rodopi. Gow, Laura (2016). The limitations of perceptual transparency. The Philosophical Quarterly, 66(265): 723–744. Gow, Laura (2017a). Everything is clear: All perceptual experiences are transparent. European Journal of Philosophy 27(2): 412–425. Gow, Laura (2017b). Why externalist representationalism is a form of disjunctivism. Ratio 31: 35–50. Green, Rebecca (1995). A Grammar of Gurr-goni (North Central Arnhem Land). PhD dissertation, Australian National University. Green, Rebecca (2003). Proto Maningrida within Proto Arnhem: evidence from verbal inflectional suffixes. In Nicholas Evans (ed.), The Non-Pama-Nyungan Languages of Northern Australia: Comparative Studies of the Continent’s Most Linguistically Complex Region. Canberra: Pacific Linguistics, 369–421. Grice, Paul (1991). Studies in the Way of Words. Cambridge, Mass.: Harvard University Press.

330

REFERENCES

Grice, Paul (2001). Aspects of Reason, ed. Richard Warner. Oxford: Clarendon Press. Grimes, Joseph E. (1976). The Thread of Discourse. The Hague: Mouton de Gruyter. Grimes, Joseph (1976). The Thread of Discourse (Janua Linguarum. Series Minor). Berlin: DeGruyter. Grondin, Simon (2010). Timing and time perception: A review of recent behavioral and neuroscience findings and theoretical directions. Attention, Perception, & Psychophysics, 72: 561–582. Grondin, Simon (2020). The Perception of Time: Your Questions Answered. London: Routledge. Grush, Rick (2007). Time and experience. In Thomas Mu¨ller (ed.), Philosophie der Zeit. Frankfurt: Klosterman, 1–18. Guillaume, Gustave, and Annette Vassant (1992). Leçons de linguistique de Gustave Guillaume 1938-1939, vol. 12. Lille: Presses Universitaires du Septentrion. Gumperz, John J., and Stephen C. Levinson (eds) (1996). Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Hacking, Ian. (1967). A slightly more realistic personal probability. Philosophy of Science 34: 311–325. Hacking, Ian (2001). An Introduction to Probability and Inductive Logic. Cambridge: Cambridge University Press. Hacquard, Valentine (2009). On the interaction of aspect and modal auxiliaries. Linguistics and Philosophy 32(3): 279–315. Haiman, John (1980). Hua. A Papuan language of Eastern Highlands of New Guinea. Amsterdam: John Benjamins Publishing Company. Haiman, John (1988). Inconsequential clauses in Hua and the typology of clauses. In John Haiman and Sandra A. Thompson (eds), Clause Combining in Grammar and Discourse, 49–70. Amsterdam/Philadelphia, PA: John Benjamins Publishing Company. Hale, Kenneth (1969). Papago /cˇim/. International Journal of American Linguistics 35(2): 203–212. Hale, Kenneth (1976). The adjoined relative clause in Australia. In Robert Dixon (ed.), Grammatical categories in Australian languages. Canberra: Australian Institute of Aboriginal Studies Press, 78–105. Hall, Edward T. (1973 [1959]). The Silent Language. New York: Anchor Books. Hall, Edward T. (1976). Beyond Culture. New York: Anchor Books. Hall, Edward T. (1990). The Hidden Dimension. New York: Anchor Books. Hardy, Heather K., and Lynn Gordon (1980). Types of adverbial and modal constructions in Tolkapaya. International Journal of American Linguistics 46(3): 183–196. Harman, Gilbert (1973). Thought. Princeton, NJ: Princeton University Press. Hartle, James. B (2005). The physics of now. American Journal of Physics 73 (2):101–109. Hartshorn, Charles, and Paul Weiss (eds) (1932). Collected Papers of Charles Sanders Peirce. Cambridge, Mass.: Harvard University Press. Harvey, Mark (2002). A Grammar of Gaagudju. Berlin/New York: De Gruyter Mouton. Haspelmath, Martin (2001). The European linguistic area: Standard average European. In Martin Haspelmath, Ekkehard Ko¨nig, Wulf Oesterreicher and Wolfgang Dressler (eds), Language Typology and Language Universals/Sprachtypologie und sprachliche Universalien/La typologie des langues et les universaux linguistiques, vol. 2. Berlin: Walter de Gruyter, 1492–1510. Hawking, Stephen W. (1988). A Brief History of Time: From the Big Bang to Black Holes. London: Bantam Press. Hawley, Katherine (2001). How Things Persist. Oxford: Oxford University Press.

REFERENCES

331

Heath, Jeffrey (1984). Functional Grammar of Nunggubuyu. Canberra/Atlantic Highlands, NJ: Humanities Press/Australian Institute of Aboriginal Studies. Hedden, Brian (2019). Hindsight Bias is not a Bias. Analysis 79: 43–52. Heidegger, Martin (1996 [1927]). Being and Time. Translated by Joan Stambaugh. Albany, NY: State University of New York Press. Heine, Bernd (1994). On the genesis of aspect in African languages: The proximative. Annual Meeting of the Berkeley Linguistics Society 20(2): 35–46. Heine, Bernd (2002). On the role of context in grammaticalization. In Ilse Wischer and Gabriele Diewald (eds), New Reflections on Grammaticalization. Amsterdam/Philadelphia: John Benjamins Publishing Company, 83–101. Heine, Bernd, and Tania Kuteva (2002). World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Heinemann, Eric G., Endel Tulving, and Jacob Nachmias (1959). The effect of oculomotor adjustments on apparent size. America Journal of Psychology, 72: 32–45. Hercus, Louise (1994). A Grammar of the Arabana-Wangkangurru Language, Lake Eyre Basin, South Australia. Canberra: Department of Linguistics, Research School of Pacific and Asian Studies, Australian National University. Higginbotham, James (1985). On semantics. Linguistic Inquiry 16(4): 547–594. Higginbotham, James (2003). Remembering, imagining, and the first person. In Alex Barber (ed.), Epistemology of Language. Oxford: Oxford University Press, 496–533. Hintz, Daniel J. (2011). Crossing Aspectual Frontiers: Emergence, Evolution, and Interwoven Semantic Domains in South Conchucos Quechua Discourse. Berkeley/Los Angeles: University of California Press. Hoerl, Christoph (2009). Time and tense in perceptual experience. Philosophers’ Imprint 9: 1–18. Hoerl, Christoph (2013a). Tense and the Psychology of Relief. Topoi 34(1): 217–231. Hoerl, Christoph (2013b). A succession of feelings, in and of itself, is not a feeling of succession. Mind 122(486): 373–417. Hoerl, Christoph (2013c). Husserl, the absolute flow, and temporal experience. Philosophy and Phenomenological Research 86: 376–441. Hoerl, Christoph (2014). Do we (seem to) perceive passage? Philosophical Explorations 17(2): 188–202. Hoerl, Christoph (2018). Experience and time: Transparency and present. Ergo, 5(5): 127–151. Hoerl, Christoph, and Teresa McCormack (2019). Thinking in and about time: A dual systems perspective on temporal cognition. Behavioral and Brain Sciences 42: 1–69. Hollebrandse, Bart (2018). Indirect recursion: The importance of second-order embedding and its implications for cross-linguistic research. In Luis Amaral, Marcus Maia, Andrew Nevins, and Tom Roeper (eds), Recursion across Domains. Cambridge: Cambridge University Press, 35–47. Homer, Vincent (2011). French modals and perfective: A case of aspectual coercion. In Mary Byram-Washburn, Katherine McKinney-Bock, Erika Varis, Ann Sawyer and Barbara Tomaszewicz (eds), Proceedings of the 28th West Coast Conference on Formal Linguistics (WCCFL 28). Somerville, Mass.: Cascadilla, 106–114. Hopper, Paul J. (1991). On some principles of grammaticization. In Elizabeth Closs Traugott and Berndt Heine (eds), Approaches to Grammaticalization, vol. 1. Amsterdam/ Philadelphia,PA: John Benjamins Publishing Company, 17–36. Hornstein, Norbert (1990). As Time Goes by: Tense and Universal Grammar. Cambridge, Mass.: The MIT Press.

332

REFERENCES

Howson, Colin (2000). Hume’s Problem: Induction and the Justification of Belief. Oxford: Oxford University Press. Huggett, Nick (2014). Skeptical notes on a physics of passage. Annals of the New York Academy of Sciences, 1326(1): 9–17. Hume, David (1978 [1739]). A Treatise of Human Nature. New York: Oxford University Press. Husserl, Edmund (1964 [2019]). The Phenomenology of Internal Time-Consciousness, ed. Martin Heidegger. Translated by James S. Churchill. Bloomington: Indiana University Press. Husserl, Edmund (1991 [1928]). On the Phenomenology of the Consciousness of Internal Time (1893–1917), part A. Translated by John B. Brough. Dordrecht: Kluwer. Ippolito, Michela (2003). Presuppositions and implicatures in counterfactuals. Natural Language Semantics 11(2): 145–186. Ippolito, Michela (2006). Semantic composition and presupposition projection in Subjunctive Conditionals. Linguistics and Philosophy 29(6): 631–672. Ismael, Jenann T. (2006). The Situated Self. Oxford: Oxford University Press. Ismael, Jenann T. (2016). How Physics Makes Us Free. Oxford: Oxford University Press. Ismael, Jenann T. (2017). Passage, flow, and the logic of temporal perspectives. In Christophe Bouton and Philippe Huneman (eds), Time of Nature and the Nature of Time: Philosophical Perspectives of Time in Natural Sciences. Cham, Switzerland: Springer. 23–38. Jackman, Henry (2005). Temporal externalism, deference, and our ordinary linguistic practice. Pacific Philosophical Quarterly 86: 365–380. Jackson, Frank (1998). From Metaphysics to Ethics. Oxford: Oxford University Press Jacobson, Pauline (2014). Compositional Semantics: An Introduction to the Syntax/Semantics Interface. Oxford: Oxford University Press. James, William (1890). The Principles of Psychology, 2 volumes. London, Macmillan; New York: Dover (1950). Janssen, Theo M.V. (1997). Compositionality. In J. Van Benthem and Alice Ter Meulen (eds), Handbook of Logic and Language. Cambridge, Mass.: The MIT Press, 417–473. Jaszczolt, Kasia M. (2005). Default Semantics: Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. Jaszczolt, Kasia M. (2009). Representing Time: An Essay on Temporality as Modality. Oxford: Oxford University Press. Jaszczolt, Kasia M. (2010). Default Semantics. In: Bernd Heine and Heiko Narrog (eds), The Oxford Handbook of Linguistic Analysis. Oxford: Oxford University Press, 193–221. Jaszczolt, Kasia M. (2012). Cross-linguistic differences in expressing time and universal principles of utterance interpretation. In Luna Filipovic´ and Kasia M. Jaszczolt (eds), Space and Time in Languages and Cultures: Linguistic Diversity. Amsterdam: John. Benjamins Publishing Company, 95–121. Jaszczolt, Kasia M. (2013). Temporality and epistemic commitment: An unresolved question. In Kasia M. Jaszczolt and Louis de Saussure (eds), Time: Language, Cognition, and Reality. Oxford: Oxford University Press, 193–209. Jaszczolt, Kasia M. (2016a). Meaning in Linguistic Interaction: Semantics, Metasemantics, Philosophy of Language. Oxford: Oxford University Press. Jaszczolt, Kasia M. (2016b). Temporal reference without the concept of time? In Barbara Lewandowska-Tomaszczyk (ed.). Conceptualizations of Time. Amsterdam: John Benjamins Publishing Company, 3–24.

REFERENCES

333

Jaszczolt, Kasia M. (2017). Temporal reference. In: Anne Barron, Yueguo Gu, and Gerard Steen (eds), The Routledge Handbook of Pragmatics. London: Routledge, 227–240. Jaszczolt, Kasia M. (2018a). Time, perspective and semantic representation. Language and Cognition 10: 26–55. Jaszczolt, Kasia M. (2018b). Pragmatic indexicals. In Minyao Huang and Kasia M. Jaszczolt (eds), Expressing the Self: Cultural Diversity and Cognitive Universals. Oxford: Oxford University Press, 260–286. Jaszczolt, Kasia M. (2020). Human imprints of real time: From semantics to metaphysics. Philosophia 48: 1855–1879. Jaszczolt, Kasia M. (2021). Functional proposition: A new concept for representing discourse meaning? Journal of Pragmatics 171: 200–214. Jaszczolt, Kasia M. (2023). Semantics, Pragmatics, Philosophy: A Journey through Meaning. Cambridge: Cambridge University Press. Jaszczolt, Kasia M. and Lidia Berthon. Forthcoming. Negotiation and joint construction of meaning (or why health providers need philosophy of communication). In Maria Grazia Rossi and Sarah Bigi (eds), A Pragmatic Agenda for Healthcare: Fostering Inclusion and Active Participation through Shared Understanding. Amsterdam: J. Benjamins. Jaszczolt, Kasia M., and Louis de Saussure (eds) (2013). Time: Language, Cognition, and Reality. Oxford: Oxford University Press. Jaszczolt, Kasia M., and Maciej Witek (2018). Expressing the self: From types of de se to speech-act types. In Minyao Huang and Kasia M. Jaszczolt (eds), Expressing the Self: Cultural Diversity and Cognitive Universals. Oxford: Oxford University Press, 187–221. Johnston, Alan, and Shin’ya Nishida (2001). Time perception: brain time or event time? Current Biology 11(11): R427–R430. Kahneman, Daniel (2011). Thinking, Fast and Slow. London: Allen Lane. Kamp, Hans (1979). Events, instants, and temporal reference. In Rainer Bauerle, Urs Egli, and Arnim von Stechow (eds), Semantics from Different Points of View. Berlin: Springer, 367–417. Kamp, Hans (2013 [1999]). Deixis in discourse: Reichenbach on temporal reference. In Klaus von Heusinger and Alice ter Meulen (eds), Meaning and the Dynamics of Interpretation: Selected Papers of Hans Kamp. Leiden: Brill, 105–159. Kamp, Hans (2019). Tense and aspect in Discourse Representation Theory. In Robert Truswell (ed.), The Oxford Handbook of Event Structure. Oxford: Oxford University Press, 523–583. Kamp, Hans, and Uwe Reyle (1993). From Discourse to Logic. Dordrecht: Kluwer Academic Publishers. Kant, Immanuel (1838). Critique of Pure Reason. Translated by Francis Haywood. London: William Pickering. Kant, Immanuel (1987 [1790]). Critique of Judgment. Indianapolis: Hackett Publishing Co. Kant, Immanuel (1999 [1781/1787]). Critique of Pure Reason. Cambridge: Cambridge University Press. Kapitonov, Ivan (2019). A Grammar of Kunbarlang. Melbourne: The University of Melbourne. Kapitonov, Ivan, and Caroline Gentens (2018). Yimarne in Kunbarlang: from similative to quotative. Paper presented at the 8th Syntax of the World’s Languages conference (SWL 8), INALCO, Paris. Kaplan, David (1989). Demonstratives: An Essay on the Semantics, Logic, Metaphysics, and Epistemology of Demonstratives and other Indexicals. In J. Almog, J. Perry and H. Wettstein (eds), Themes from Kaplan. Oxford: Oxford University Press, 481–563.

334

REFERENCES

Karlsson, F. (2009). Recursion and iteration. In Harry van der Hulst (ed.), Recursion and Human Language. Berlin: Mouton de Gruyter, 43–68. Keats, John (1977). Hyperion. In John Keats: The Complete Poems. London: Penguin Classics, 283. Kehayov, Petar, and Florian Siegl (2006). The evidential past participle in Estonian reconsidered. Études finno-ougriennes 38: 75–117. Kempson, Ruth M. (1975). Presupposition and the Delimitation of Semantics. Cambridge: Cambridge University Press. Kempson, Ruth M. (1979). Presupposition, opacity, and ambiguity. In: Choon-Kyu Oh and David A. Dinneen (eds), Syntax and Semantics, vol. 11. New York: Academic Press, 283– 297. Kempson, Ruth M. (1986). Ambiguity and the semantics-pragmatics distinction. In Charles Travis (ed.), Meaning and Interpretation. Oxford: Blackwell, 77–103. Kind, Amy (2008). How to believe in qualia. In Edmond Leo Wright (ed.), The Case for Qualia. Cambridge, Mass.: The MIT Press, 285–298. Kiverstein Julian and Valtteri Arstila (2013). Time in mind. In Heather Dyke and Adrian Bardon (eds), A Companion to the Philosophy of Time. Oxford: Wiley-Blackwell, 444–469. Klein, Wolfgang (1994). Time in Language. London/New York: Routledge. Koenig, Jean-Pierre, and Lian-Cheng Chief (2008). Scalarity and state-changes in Mandarin (and other languages). In Olivier Bonami and Patricia Cabredo Hofherr (eds), Empirical Issues in Syntax and Semantics, vol. 7. Paris: CSSP, 241–262. Koenig, Jean-Pierre, and Nuttannart Muansuwan (2000). How to end without ever finishing: Thai semi–perfectivity. Journal of Semantics 17(2): 147–182. Korn, Agnes, and Irina Nevskaya (eds) (2017). Prospective and Proximative in Turkic, Iranian and Beyond. Wiesbaden: Reichert Verlag. Kortylewski, Adam, Ju He, Qing Liu, Christian Cosgrove, Chenglin Yang, and Alan L. Yuille (2021). Compositional generative networks and robustness to perceptible image changes. 55th Annual Conference on Information Sciences and Systems (CISS): 1–8. Kothari, Anubha. 2008. Event Culmination as Implicature in Hindi Perfectives. Stanford University, CA., ms. Kozlov, Alexey (2019). Iterative and avertive polysemy in Moksha Mordvin. Sprachtypologie Und Universalienforschung (STUF) 72(1): 133–159. Kratzer, Angelika (1979). Conditional necessity and possibility. In Rainer Bauerle, Urs Egli, and Arnim von Stechow (eds), Semantics from Different Points of View. Berlin: Springer, 117–147. Kroeger, Paul (2017). Frustration, culmination, and inertia in Kimaragang grammar. Glossa, 2(1). 1–29. Kronning, Hans (1996). Modalité, cognition et polysémie: sémantique du verbe modal ‘devoir’. Stockholm, Sweden: Almqvist & Wiksell International. Kronning, Hans (2001a). Pour une tripartition des emplois du modal ‘devoir’. In Patrick Dendale and Johan van der Auwera (eds), Les verbes modaux (Cahiers Chronos 8), 67–84. Amsterdam/Atlanta: Rodopi. Kronning, Hans (2001b). Nécessité et hypothèse: devoir non déontique au conditionnel. In Patrick Dendale and Lilliane Tasmowski (eds), Le conditionnel en français, Metz: Université de Metz, 251–276. Ku¨nnapas, Teodor M. (1968). Distance perception as a function of available visual cues. Journal of Experimental Psychology 77: 523–529.

REFERENCES

335

Kutach, Douglas (2002). The entropy theory of counterfactuals. Philosophy of Science 69(1): 82–104. Kutach, Douglas (2013). Causation and its Basis in Fundamental Physics. Oxford: Oxford University Press. Kuteva, Tania (1998). On identifying an evasive gram: Action narrowly averted. Studies in Language 22(1): 113–160. Kuteva, Tania (2000a). TAM-auxiliation and the avertive category in Northeast Europe. In Jocelyne Fernandez-Vest (ed.), Grammaticalisation aréale et sémantique cognitive: les langues fenniques et sames. Louvain: Peeters, 27–41. Kuteva, Tania (2000b). Areal grammaticalization: The case of the Bantu-Nilotic borderland. Folia Linguistica 34(3-4): 267–284. Kuteva, Tania (2001). Auxiliation: An Enquiry into the Nature of Grammaticalization. Oxford/New York: Oxford University Press. Kuteva, Tania, Bas Aarts, Gergana Popova and Anvita Abbi (2019). The grammar of ‘nonrealization.’ Studies in Language 43(4): 850–895. Lakoff, George, and Mark Johnson (1980). Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, George, and Mark Johnson (1999). Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books. Lakoff, George, and Mark Turner (1989). More Than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press. Landman, Fred (1992). The progressive. Natural Language Semantics 1(1): 1–32. Larrivée, Pierre, and Adeline Patard (2019). Pathways of evolution, contiguity and bridging contexts. In Daniël Van Olmen, Tanja Mortelmans, and Frank Brisard (eds), Aspects of Linguistic Variation. Berlin/New York: De Gruyter Mouton, 185–206. Latham, Andrew, Kristie Miller, and James Norton (2020). An empirical investigation of purported passage phenomenology. The Journal of Philosophy 117(7): 353–386. Laughren, Mary (1999). Constraints on the pre-auxiliary position in Warlpiri and the nature of the auxiliary. In John Henderson (ed.), Proceedings of the 1999 Conference of the Australian Linguistic Society. UWA, Perth: Australian Linguistic Society, 1–18. Laughren, Mary (2002). Syntactic constraints in a ‘free word order’ language. In Mengistu Amberber and Peter Collins (eds), Language Universals and Variation. Westport, Conn.: Praeger Publishers, 83–130. Le Poidevin, Robin (1998). The past, present and future of the debate about tense. In Robin Le Poidevin (ed.), Questions of Time and Tense. Oxford: Oxford University Press, 13–42. Le Poidevin, Robin (2007). The Images of Time: An Essay on Temporal Representation. Oxford: Oxford University Press. Le Poidevin, Robin (2011). The temporal prison. Analysis 71: 456–465. Le Poidevin, Robin (2015). Perception and time. In Mohan Matthen (ed.), The Oxford Handbook of Philosophy of Perception. Oxford: Oxford University Press, 459–474. Leder, Harry (1991). Tense and Temporal Order. PhD dissertation, Massachusetts Institute of Technology. Lee, Geoffrey (2014). Temporal experience and the temporal structure of experience. Philosophers’ Imprint, 14(3): 1–21. Leeding, Velma J. (1989). Anindilyakwa phonology and morphology. University of Sydney PhD Thesis. https://ses.library.usyd.edu.au/handle/2123/1558. (accessed 20 December, 2022). Legate, Julie Anne (2003). The morphosemantics of Warlpiri counterfactual conditionals. Linguistic Inquiry 34(1): 155–162.

336

REFERENCES

Legate, Julie Anne (2009). Clausal recursion and embedding in Warlpiri. LingBuzz. https:// ling.auf.net/lingbuzz/000872 Leibowitz, H., and Donovan Moore (1966). Role of changes in accommodation and convergence in the perception of size. Journal of the Optical Society of America, 56(8): 1120–1123. Leininger, Lisa (2018). Objective becoming: In search of A-ness. Analysis 78(1): 108–117. Lepore, Ernie, and Matthew Stone (2015). Imagination and Convention: Distinguishing Grammar and Inference in Language. Oxford: Oxford University Press. Levinson, Stephen C. (2003). Space in Language and Cognition: Explorations in Cognitive Diversity. Cambridge: Cambridge University Press. Lewis, David (1986). On the Plurality of Worlds. Malden, Mass.: Blackwell Publishing. Lewis, Peter (2016). Quantum Ontology: A Guide to the Metaphysics of Quantum Mechanics. New York: Oxford University Press. Liao, Shen-yi (2012). What are centered worlds? The Philosophical Quarterly 62(247): 294–316. Lichtenberk, Frantisek (2008). A Grammar of Toqabaqita. Berlin: Walter de Gruyter. Lin, Jo-Wang (2005). Time in a language without tense: The case of Chinese. Journal of Semantics 23: 1–53. Lloyd, Dan, and Valtteri Arstila (2014). The disunity of time. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press, 657–664. Loewer, Barry (2007). Counterfactuals and the second law. In Huw Price and Richard Corry (eds), Causation, Physics and the Constitution of Reality: Russell’s Republic Revisited. Oxford: Oxford University Press, 293–326. Loewer, Barry (2012). Two accounts of laws and time. Philosophical Studies 160: 115–137. Longacre, Robert E. (1976). An Anatomy of Speech Notions. Lisse: The Peter de Ridder Press. Longacre, Robert E. (1996). The Grammar of Discourse. New York, NY: Springer. Loukotka, Čestmı´r (1968). Classification of South American Indian languages. Los Angeles: Latin American Center, UCLA. Ludlow, Peter (1999). Semantics, Tense, and Time: An Essay in the Metaphysics of Natural Language. Cambridge, Mass.: The MIT Press. Ludlow, Peter (2013). Tensism. In Kasia M. Jaszczolt and Louis de Saussure (eds), Time: Language, Cognition, and Reality. Oxford: Oxford University Press. 175–192. Lyons, John (1977). Semantics, vol. 2. Cambridge: Cambridge University Press. MacFarlane, John (2003). Future contingents and relative truth. The Philosophical Quarterly 53: 321–336. MacFarlane, John (2008). Truth in the garden of forking paths. In Manuel Garcı´aCarpintero and Max Ko¨lbel (eds), Relative Truth. Oxford: Oxford University Press: 81–102. Maclaurin, James, and Heather Dyke (2002). ‘Thank goodness that’s over’: The evolutionary story. Ratio, 15: 276–292. Magidor, Ofra (2015). The myth of the de se. Philosophical Perspectives, 29: 249–283. Malchukov, Andrej L. (2004). Towards a semantic typology of adversative and contrast marking. Journal of Semantics 21(2): 177–198. Mari, Alda (2013). Modalité et Temps: des modèles aux données. Berne: Peter Lang. Mari, Alda, and Fabienne Martin (2008). Perfective and Imperfective in French – Kinds of abilities and Actuality Entailment (And some notes on epistemic readings). Ms, Institut Jean Nicod, Paris. https://jeannicod.ccsd.cnrs.fr/ijn_00416168

REFERENCES

337

Martin, Fabienne, and Florian Scha¨fer (2012). The modality of ‘offer’ and other defeasible causative verbs. In Nathan Arnett and Ryan Bennett (eds), Proceedings of the 30th West Coast Conference on Formal Linguistics. Somerville, Mass.: Cascadilla Proceedings Project, 248–258. Martin, Fabienne, and Florian Scha¨fer (2017). Sublexical modality in defeasible causative verbs. In Ana Arregui, Marı´a Luisa Rivero, and Andrés Salanova (eds), Modality Across Syntactic Categories. Oxford: Oxford University Press, 87–108. Martin, Fabienne, and Hamida Demirdache (2020). Partitive accomplishments across languages. Linguistics 58(5): 1195–1232. Martin, Michael G. F. (2002). The transparency of experience. Mind and Language 17(4): 376–425. Matthewson, Lisa (2006). Temporal semantics in a superficially tenseless language. Linguistics and Philosophy 29: 673–713. Maudlin, Tim (2002). Remarks on the passing of time. Proceedings of the Aristotelian Society, 3: 259–274. Maudlin, Tim. (2007). The Metaphysics Within Physics. Oxford: Oxford University Press. Mayo, Deborah G. (1996). Error and the Growth of Experimental Knowledge. Chicago: University of Chicago Press. McArthur, Robert P. (1976). Tense Logic. Dordrecht/Boston: Reidel. McCall, Storrs (1976). Objective time flow. Philosophy of Science 43: 337–362. McCall, Storrs (1994). A Model of the Universe. Oxford: Oxford University Press. McConvell, Patrick, and Mary Laughren (2004). The Ngumpin-Yapa subgroup. In Claire Bowern and Harold James Koch (eds), Australian Languages: Classification and the Comparative Method. Amsterdam: John Benjamins Publishing Company, 151–177. McGregor, William (1990). A Functional Grammar of Gooniyandi. Amsterdam: John Benjamins Publishing Company. McKay, Graham Richard (1975). Rembarnga: A Language of Central Arnhem Land. PhD dissertation, Australian National University. McTaggart, John M. E. (1908). The unreality of time. Mind 17: 457–474. McTaggart, John M. E. (1927). The Nature of Existence, vol. 2. Cambridge: Cambridge University Press. Meakins, Felicity, and Rachel Nordlinger (2014). A Grammar of Bilinarra: An Australian Aboriginal Language of the Northern Territory. Berlin: Walter de Gruyter. Mellor, D. H. (1981). Real Time. Cambridge: Cambridge University Press. Mellor, D. H. (1998). Real Time II. London: Routledge. Mellor, D. H. (2009). The direction of time. In Rodin Le Poidevin, Peter Simons, Andrew McGonigal, and Ross Cameron (eds), The Routledge Companion to Metaphysics. London: Routledge, 449–458. Meredith, Owen (1900). The dead Pope. In Robert Lord Lytton (ed.), The Poetical Works of Owen Meredith: Meredith’s Poems. New York: Hurst & Co, 112. Merleau-Ponty, Maurice (1962). Phenomenology of Perception. Translated by Colin Smith. London/New York: Routledge. Miller, Izchak (1984). Husserl, Perception, and Temporal Awareness. Cambridge, Mass.: The MIT Press Miller, Kristie (2018). The new growing block theory vs presentism. Inquiry: An Interdisciplinary Journal of Philosophy 61(3): 223–251. Miller, Kristie, Alex Holcombe, and Andrew. J. Latham (2018). Temporal phenomenology: Phenomenological illusion versus cognitive error. Synthese 197(2): 751–771. Millikan, Ruth (1990). The myth of the essential indexical. Noûs, 24: 723–734.

338

REFERENCES

Millikan, Ruth. (1991). Perceptual content and Fregean myth. Mind 100(399): 439–459. Minkowski, Eugene (1970). Lived Time: Phenomenological and Psychopathological Studies. Chicago: Northwestern University Press. Moens, Marc, and Mark Steedman (1988). Temporal ontology and temporal reference. Computational Linguistics 14(2): 15–28. Moeschler, Jacques and Anne Reboul (2001). Conditionnel et assertion conditionnelle. In Patrick Dendale and Liliane Tasmowski (eds), Le conditionnel en français. Metz: Université Metz, 147–168. Montague, Richard (1973). The proper treatment of quantification in ordinary English. In Patrick Suppes, Julius Moravcsik and Jaakko Hintikka (eds), Approaches to Natural Language. Dordrecht: Reidel, 221–242. Moore, George E. (1903). The refutation of Idealism. Mind 12: 433–453. Moore, Kevin Ezra (2006). Space-to-time mappings and temporal concepts. Cognitive Linguistics 17(2): 199–244. Moore, Kevin Ezra (2011). Ego-perspective and field-based frames of reference: Temporal meanings of FRONT in Japanese, Wolof, and Aymara. Journal of Pragmatics 43: 759–776. Moore, Kevin Ezra (2014). The Spatial Language of Time: Metaphor, Metonymy, and Frames of Reference. Amsterdam: John Benjamins Publishing Company. Morpurgo Davies, Anna (1983). Mycenaean and Greek prepositions: o-pi, e-pi, etc. In Alfred Heubeck and Gu¨nter Neumann (eds), Res Mycenaeae: Akten des VII. Internationalen Mykenologischen Colloquiums in Nu¨rnmerg vont 6.-10. April 1981. Go¨ttingen: Vandenhoeck and Ruprecht, 287–310. Mozersky, Joshua M. (2001). Smith on times and tokens. Synthese 129: 405–411. Mozersky, Joshua M. (2013). The B-theory on the twentieth century. In Heather Dyke and Adrian Bardon (eds), A Companion to the Philosophy of Time. Oxford: Wiley and Sons, 167–182. Mozersky, Joshua M. (2015). Time, Language, and Ontology: The World from the B-Theoretic Perspective. Oxford: Oxford University Press. Mucha, Anne (2013). Temporal interpretation in Hausa. Linguistics and Philosophy 36: 371–415. Mu¨ller, Neele (2013). Tense, Aspect, Modality, and Evidentiality Marking in South American Indigenous Languages. PhD dissertation, Radboud Universiteit. Nagel, Thomas (1986). The View from Nowhere. New York: Oxford University Press. Nash, David George (1980). Topics in Warlpiri grammar. PhD dissertation, Massachusetts Institute of Technology. Nerlich, Graham (1998). Falling branches and the flow of time. Australasian Journal of Philosophy 76: 309–316. Nordlinger, Rachel (1998). A grammar of Wambaya, Northern Territory (Australia). Canberra: Pacific Linguistics. Nordlinger, Rachel and Patrick Caudal (2012). The tense, aspect and modality system in Murrinh-Patha. Australian Journal of Linguistics 32(1): 73–113. Norton, James (2021). Experimental philosophy on time. Philosophy Compass 16 (11): e12779. Núñez, Raphael, and Eve Sweetser (2006). With the future behind them: Converging evidence from Aymara language and gesture in the crosslinguistic comparison of spatial construals of time. Cognitive Science 30(3): 401–450. Núñez, Raphael, and Kevin Cooperrider (2013). The tangle of space and time in human cognition. Trends in Cognitive Sciences 17(5): 220–229.

REFERENCES

339

Nuyts, Jan (1992). Subjective and objective modality: What is the difference? In Michael Fortescue, Peter Harder, and Lars Kristoffersen (eds), Layered Structure and Reference in a Functional Perspective, Amsterdam, John Benjamins Publishing Company, 73–97. Nuyts, Jan (2001). Epistemic Modality, Language and Conceptualization. A Cognitive Pragmatic Perspective. Amsterdam: John Benjamins Publishing Company. O’Hagan, Zachary (2018). Lexical semantics and Caquinte frustrative constructions. Paper presented at the 2018 Annual Meeting of the Society for the Study of the Indigenous Languages of the Americas (SSILA 2018). O’Shaughnessy, Brian (2000). Consciousness and the World. Oxford: Oxford University Press. Oaklander, Nathan L., and V. Alan White (2007). B-time: A reply to Tallant. Analysis 67: 332–340. Olson, Eric T. (2009). The rate of time’s passage. Analysis 69: 3–9. Osgarby, David (2018). Reconstructing proto-Mirndi verbal morphology: from particles and clitics to prefixes. Australian Journal of Linguistics 38(2): 223–292. Overall, Simon E. (2017). A typology of frustrative marking in Amazonian languages. In Alexandra Y. Aikhenvald and R. M. W. Dixon (eds), The Cambridge Handbook of Linguistic Typology. Cambridge: Cambridge University Press, 477–512. Pahontu, Beatrice (forthcoming). Les périphrases progressives dans les langues romanes. PhD dissertation, Université de Paris/Universitatea din Bucures¸ti. Palmisano, Stephen, Robert S. Allison, Mark M. Schira, and Robert J. Barry (2015). Future challenges for vection research: definitions, functional significance, measures, and neural bases. Frontiers in psychology 6: 193. DOI: 10.3389/fpsyg.2015.00193 Papafragou, Anna (2006). Epistemic modality and truth conditions. Lingua 116(10): 1688– 1702. Papineau, David (2021). The Metaphysics of Sensory Experience. Oxford: Oxford University Press. Parsons, Josh (2002). A-theory for B-theorists. The Philosophical Quarterly 52: 1–20. Parsons, Josh (2003). A-theory for tense logicians. Analysis 63: 4–6. Partee, Barbara H. (1973). Some structural analogies between tenses and pronouns in English. Journal of Philosophy 70: 601–609. Partee, Barbara H. (1984). Nominal and temporal anaphora. Linguistics and Philosophy 7: 243–286. Partee, Barbara H. (1995). Quantificational structures and compositionality. In Emmon Bach, Eloise Jelinek, Angela Kratzer, and Barbara H. Partee (eds), Quantification in Natural Languages. Dordrecht: Kluwer Academic Publishers, 541–601. Paul, Ileana, Baholisoa Simone Ralalaoherivony and Henriëtte de Swart (2020). Culminating and non-culminating accomplishments in Malagasy. Linguistics 58(5): 1285–1322. Paul, Laurie. A. (2010). Temporal experience. The Journal of Philosophy 107(7): 333–359. Paul, Laurie. A. (2014). Experience and the arrow. In Alastair Wilson (ed.), Chance and Temporal Asymmetry. Oxford: Oxford University Press, 175–193. Peirce, Charles S. (1868). Questions concerning certain faculties claimed for man. Journal of Speculative Philosophy 2: 103–114. Peirce, Charles S. (1909). Manuscript 514. Transcribed by Michel Balat with commentary by J. F. Sowa. http://www.jfsowa.com/peirce/ms514.htm [accessed 15 November 2009] Peirce, Charles S. (1932 [1903]). Harvard Lectures on Pragmatism, Lecture 7. In Charles Hartshorne and Paul Weiss (eds), Writings of Charles S. Peirce. Vol. 1. Harvard: Harvard University Press, 112–127. Pelletier, Francis Jeffry (1994). The principle of semantic compositionality. Topoi 13: 11–24.

340

REFERENCES

Peregrin, Jaroslav (2014). Inferentialism: Why Rules Matter. London: Palgrave-Macmillan. Perovic´, Katarina (2019). Three varieties of Growing Block theory. Erkenntnis 86(3): 623–645. Perry, John (1977). Frege on demonstratives. The Philosophical Review 86: 474–497. Perry, John (1979). The problem of the essential indexical. Noûs 13: 3–21. Perry, John (1986). Thought without representation. Proceedings of the Aristotelian Society 60: 137–152. Phillips, Ian (2011). Indiscriminability and experience of change. The Philosophical Quarterly 61(245): 808–827. Phillips, Ian (2012). Attention to the passage of time. Philosophical Perspectives 26: 277–308. Phillips, Ian (2013a). Afterimages and sensations. Philosophy and Phenomenological Research 87(2): 417–453. Phillips, Ian (2013b). Perceiving the passing of time. Proceedings of the Aristotelian Society 113: 225–252. Phillips, Ian (2014a). Experience of and in time. Philosophy Compass, 9(2): 131–144. Phillips, Ian (2014b). The temporal structure of experience. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press: 139–158. Phillips, Joshua (forthcoming). Negation in Australian languages. In Claire Bowern (ed.), The Oxford Guide to Australian Languages. Oxford: Oxford University Press. Piata, Anna (2018). The Poetics of Time: Metaphors and Blends in Language and Literature. Amsterdam: John Benjamins Publishing Company. Piata, Anna (2019). When time passes quickly: A cognitive linguistic study on compressed time. Metaphor and Symbol 34(3): 167–184. Pietarinen, Ahti (2005). The compositionality of concepts and Peirce’s Pragmatic Logic. In Markus Werning, Eduoard Machery, and Gerhard Schurtz (eds), The Compositionality of Concepts and Meanings: Foundational Issues. Berlin/Boston: De Gruyter, 247–270. Pinson, Mathilde (2020). It looks like this complementizer used to be an adjective. Lexis 3(15): 1–19. Plungian, Vladimir (2001). Antirezul’tativ: do i posle rezul’tata. In Vladimir Plungian (ed.), Issledovanija po teorii grammatiki I: Glagol’nye kategorii. Moscow: Russkie slovari, 50–88. Plungian, Vladimir (2005). Irrealis and modality in Russian and in typological perspective. In Bjo¨rn Hansen and Petr Karlı´k (eds), Modality in Slavonic Languages: New Perspectives. Mu¨nchen: Verlag Otto Sagner, 135–146. Polanyi, Michael (1974). Personal Knowledge: Towards a Post-Critical Philosophy. Chicago: University of Chicago Press. Polanyi, Michael (2009 [1966]). The Tacit Dimension. Chicago: University of Chicago Press. Po¨ppel, Ernst (1978). Time perception. In R. Held, H. W. Leibowitz, and H. L. Teuber (eds), Handbook of Sensory Physiology, vol. 8. Berlin: Springer-Verlag, 713–729. Popper, Karl (1971). Conjectural knowledge: My solution to the problem of induction. Revue Internationale de Philosophie 25: 167–197. Portner, Paul (2018). Commitment to priorities. In Daniel Fogal, Daniel W. Harris, and Matt Moss (eds), New Work on Speech Acts. Oxford: Oxford University Press, 296–316. Potts, Christopher (2005). The Logic of Conventional Implicatures. Oxford/New York: Oxford University Press. Potts, Christopher (2007). The expressive dimension. Theoretical Linguistics 33(2): 165–198. Predelli, Stefano (1998). I am not here now. Analysis 58: 107–115.

REFERENCES

341

Price, Huw (1996). Time’s Arrow and Archimedes’ Point: New Directions for the Physics of Time. Oxford: Oxford University Press. Price, Huw (2002). Boltzmann’s time bomb. The British Journal for the Philosophy of Science 53(1): 83–119. Price, Huw (2007). Causal perspectivalism. In Huw Price and Richard Corry (eds), Causation, Physics, and the Constitution of Reality: Russell’s Republic Revisited. Oxford: Oxford University Press, 250–292. Price, Huw (2011). The flow of time. In Craig Callender (ed.), The Oxford Handbook of Philosophy of Time. Oxford: Oxford University Press, 276–311. Prior, Arthur N. (1959). Thank goodness that’s over. Philosophy 34: 12–17. Prior, Arthur N. (1967). Past, Present and Future. Oxford: Clarendon Press. Prior, Arthur N. (1968). Papers on Time and Tense. Oxford: Clarendon Press. Prior, Arthur N. (2003). Papers on Time and Tense, new edn. Oxford: Oxford University Press. Prosser, Simon (2012). Why does time seem to pass? Philosophy and Phenomenological Research 85(1): 92–116. Prosser, Simon (2013a). Passage and perception. Noûs 47: 69–84. Prosser, Simon (2013b). The passage of time. In Heather Dyke and Adrian Bardon (eds), A Companion to the Philosophy of Time. New Jersey: Wiley-Blackwell, 315–327. Prosser, Simon (2013c). Experience, thought, and the metaphysics of time. In Kasia Jaszczolt and Louis de Saussure (eds), Time: Language, Cognition, and Reality. Oxford: Oxford University Press, 157–174. Prosser, Simon (2015). Why are indexicals essential? Proceedings of the Aristotelian Society, 115: 211–233. Prosser, Simon (2016). Experiencing Time. Oxford: Oxford University Press. Prosser, Simon (2019). Locating the contradiction in our understanding of time. Behavioral and Brain Sciences 42: 45–46. Prosser, Simon, and François Recanati (eds) (2012). Immunity to Error through Misidentification. Cambridge: Cambridge University Press. Pullum, Geoffrey K. (2020). Theorizing about the syntax of human language: A radical alternative to generative formalisms. Cadernos de Lingu¨´ıstica 1: 1–33. Pustejovsky, James (1995). The Generative Lexicon. Cambridge, Mass.: The MIT Press. Putnam, Hilary (1975). The meaning of meaning. In Keith Gunderson (ed.), Language, Mind, and Knowledge. Minneapolis: Minnesota University Press, 215–271. Pym, Noreen, and Bonnie Larrimore (1979). Papers on Iwaidja Phonology and Grammar (Work Papers SIL-AAB, Series A). Darwin: Summer Institute of Linguistics. Quine, W. V. (1953 [1950]). Identity, ostention, and hypostasis. In From a Logical Point Of View. London: Harvard University Press, 65–79. R (on the application of Miller) (Appellant) v The Prime Minister (Respondent) (2019) UK Supreme Court. https://www.supremecourt.uk/watch/prorogation/judgment.html [accessed 28 July 2022] Radden, Gu¨nter (2003). The metaphor TIME AS SPACE across languages. Zeitschrift fu¨r Interkulturellen Fremdsprachenunterricht 8(2): 226–239. Rasmussen, Joshua (2012). Presentists may say goodbye to A-properties. Analysis 72: 270–276. Recanati, François (2007). Perspectival Thought: A Plea for (Moderate) Relativism. Oxford: Oxford University Press. Recanati, François (2012). Immunity to error through misidentification: What it is and where it comes from. In Simon Prosser and François Recanati (eds), Immunity to Error through Misidentification. Cambridge: Cambridge University Press, 180–201.

342

REFERENCES

Reichenbach, Hans (1938). Experience and Prediction. Chicago: University of Chicago Press. Reichenbach, Hans (1947). Elements of Symbolic Logic. New York: Macmillan. Reichenbach, Hans (1956). The Direction of Time. Berkeley: University of California Press. Richard, Mark (1981). Temporalism and eternalism. Philosophical Studies 39(1): 1–13. Richardson, Louise (2014). Space, time and Molyneux’s question. Ratio, 27(4): 483–505. Richardson, Louise (2018). Commentary on Christoph Hoerl’s ‘Experience and time: Transparency and presence’. https://philosophyofbrains.com/2018/09/24/symposiumon-christoph-hoerls-experience-and-time-transparency-and-presence.aspx [accessed 22 September 2019] Rodrı´guez, Lydia (2014). Time in Language, Gesture, and Thought: A Case Study in Chol Mayan. PhD dissertation, University of Virginia. Rodrı´guez, Lydia (2019). Time is not a line. Temporal gestures in Chol Mayan. Journal of Pragmatics 151: 1–17. Rolls, Edmund T. (2016). Cerebral Cortex: Principles of Operation. Oxford University Press, Oxford. Rose, David (2001). The Western Desert Code: An Australian Cryptogrammar. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The Australian National University. Rovelli, Carlo (2018). The Order of Time. Translated by Erica Segre and Simon Carnell. London: Penguin Books. Rullmann, Hotze, and Lisa Matthewson (2018). Towards a theory of modal-temporal interaction. Language 94(2): 281–331. Russell, Benjamin (2012). Probabilistic Reasoning and the Computation of Scalar Implicatures. PhD dissertation, Brown University. Russell, Bertrand (1936). On order in time. Journal of Symbolic Logic 1(2): 72–73. Salmon, Wesley (1997). Causality and Explanation. Oxford: Oxford University Press. Sansò, Andrea (2020). Routes towards the Irrealis. Transactions of the Philological Society 118(3): 401–446. Sapir, Edward (1921). Language. New York: Harcourt, Brace, and Co. Sapir, Edward (1931). Conceptual categories in primitive languages. Science, 74: 578. Sattig, Thomas (2019). The sense of temporal flow: a higher-order account. Philosophical Studies 176: 3041–3059. Sauerland, Uli (2018). False speech reports in Pirahã: A comprehension experiment. In Luiz Amaral, Marcus Maia, Andrew Nevins, and Tom Roeper (eds), Recursion across Domains. Cambridge: Cambridge University Press: 21–34. Saussure, Louis de (2013a). Background relevance. Journal of Pragmatics 59: 178–189. Saussure, Louis de (2013b). Perspectival interpretations of tenses. In Kasia Jaszczolt and Louis de Saussure (eds), Time: Language, Cognition and Reality. Oxford: Oxford University Press, 46–69. Saussure, Louis de, and Tim Wharton (2020). Relevance, effects and affect. International Review of Pragmatics, 12: 183–205. Savage, Heidi, Melissa A. Ebbers, and Robert M. Martin (2019). The Meaning of Language, 2nd edn. London: The MIT Press. Schultze-Berndt, Eva (2000). Simple and Complex Verbs in Jaminjung: A Study of Event Categorisation in an Australian Language. PhD Dissertation, Universiteit Nijmegen. Schwellenbach, Sarah (2019). Avertiv und Proximativ: Eine korpusbasierte synchrone und diachrone Untersuchung der romanischen Sprachen. Avertiv und Proximativ. Berlin: De Gruyter.

REFERENCES

343

Seno, Takeharu, Yuki Yamada, and Stephen Palmisano (2012). Directionless vection: a new illusory self-motion perception. i-Perception 3(10): 775–777. Shardlow, Jack, Ruth Lee, Christoph Hoerl, Teresa McCormack, Patrick Burns, and Alison S. Fernandes (2021). Exploring people’s beliefs about the experience of time. Synthese 198(11): 10709–10731. Sharp, Janet Catherine (2004). Nyangumarta – A Language of the Pilbara Region of Western Australia. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The National Australian University. Shoemaker, Sydney (1968). Self-reference and self-awareness. Journal of Philosophy 65: 555–567. Sider, Theodor (1996). All the world’s a stage. Australasian Journal of Philosophy, 74: 433–453. Sider, Theodor (2001). Four-Dimensionalism: An Ontology of Persistence and Time. Oxford: Clarendon Press. Sider, Theodore (2011). Writing the Book of the World. Oxford: Clarendon Press. Simard, Candide (2013). Prosody and function of ‘iconic lengthening’ in Jaminjung. SOAS Working Papers in Linguistics 16: 65–77. Simpson, J. (2012). Warlpiri Morpho-Syntax: A Lexicalist Approach. Dordrecht: Springer. Singer, Ruth, Nita Garidjalalug, Heather Hewett, Peggy Mirwuma and Phillip Ambidjambidj (2015). Mawng Dictionary v1.0. Warruwi: Mawng Ngaralk. Sinner, Carsten, and Jana Dowah (2020). The avertive hube de + infinitive in Spanish. Estudos de Lingu¨´ıstica Galega 12: 91–114. Skow, Bradford (2012). Why does time pass? Noûs 46(2): 223–242. Skow, Bradford (2015). Objective Becoming. Oxford: Oxford University Press. Slobin, Dan I. (1996). From ‘thought and language’ to ‘thinking for speaking’. In John J. Gumperz and Stephen C. Levinson (eds), Rethinking Linguistic Relativity. Cambridge: Cambridge University Press, 70–96. Smart, J. J. C. (1949). The river of time. Mind 58, 483–494. Smart, J. J. C. (1980). Time and becoming. In Peter Van Inwagen (ed.), Time and Cause: Essays Presented to Richard Taylor. Dordrecht: Reidel: 3–15. Smith, Carlota S. (1991). The Parameter of Aspect. Dordrecht: Kluwer. Smith, Carlota S. (1997). The Parameter of Aspect, 2nd edn. Dordrecht: Kluwer. Smith, Quentin (1993). Language and Time. Oxford: Oxford University Press. Smith, Quentin (2002). Time and degrees of existence: A theory of ‘degree presentism’. In Craig Callender (ed.), Time, Reality and Experience. Cambridge: Cambridge University Press, 119–136. Smith-Dennis, Ellen (2021). Don’t feel obligated, lest it be undesirable: the relationship between prohibitives and apprehensives in Papapana and beyond. Linguistic Typology 25(3): 413–459. Sokolowski, Robert (2000). Introduction to Phenomenology. Cambridge: Cambridge University Press. Soteriou, Matthew (2013). The Mind’s Construction: The Ontology of Mind and Mental Action. Oxford: Oxford University Press Sperber, Dan (1985). Apparently irrational beliefs. In Dan Sperber, On Anthropological Knowledge. Cambridge: Cambridge University Press, 35–63. Sperber, Dan (1996). Explaining Culture: A Naturalistic Approach. Oxford: Blackwell. Sperber, Dan (2000). Metarepresentations in an evolutionary perspective. In Dan Sperber (ed.), Metarepresentations: A Multidisciplinary Perspective. Oxford: Oxford University Press, 117–137.

344

REFERENCES

Sperber, Dan, and Deirdre Wilson (1995 [1986]). Relevance: Communication and Cognition. 2nd edn. Oxford: Blackwell. Stalnaker, Robert C. (1981). Indexical belief. Synthese, 49: 129–151. Stalnaker, Robert C. (2004). Assertion revisited: On the interpretation of two-dimensional modal semantics. Philosophical Studies 118(1/2): 299–322. Steedman, Mark (1997). Temporality. In Johan Van Benthem and Alice Ter Meulen (eds), Handbook of Logic and Language. Cambridge, Mass.: The MIT Press, 895–938. Steward, Helen (2013). Processes, continuants, and individuals. Mind 122(487): 781–812. Steward, Helen (2015). What is a continuant? Aristotelian Society Supplementary Volume 89(1): 109–123. Stjernfelt, Frederic (2014). Natural Propositions: The Actuality of Peirce’s Doctrine of Dicisigns. Boston: Docent Press. Stokes, Bronwyn (1982). A Description of Nyikina: A language of the West Kimberley, Western Australia. PhD dissertation, Australian National University. Stout, Roland (1997). Processes. Philosophy 72: 19–27. Stowell, Tim. (2004). Tense and modals. In Jacqueline Guéron and Jacqueline Lecarme (eds), Time and Modality. Berlin: Springer, 621–636. Strawson, Peter F. (1952). Introduction to Logical Theory. New York: Wiley and Sons. Street, Chester S., and Lyn Street (1989). Murrinh-Patha vocabulary (electronic version). Wadeye, N.T. Ms. Suhler, Christopher, and Craig Callender (2012). Thank goodness that argument is over: Explaining the temporal value asymmetry. Philosophers’ Imprint 12(15): 1–16. Sullivan, Karen, and Linh Thuy Bui (2016). With the future coming up behind them: Evidence that time approaches from behind in Vietnamese. Cognitive Linguistics 27(2): 205–233. Sullivan, Meghan (2018). Time Biases: A Theory of Rational Planning and Personal Persistence. Oxford: Oxford University Press. Szabó, Zoltán Gendler (2000). Problems of Compositionality. New York: Routledge. Tallant, Jonathan (2007). What is B-time? Analysis 67: 147–156. Tallant, Jonathan (2013). Time. Analysis 73: 369–379. Tallant, Jonathan (2019). Presentism remains. Erkenntnis 84: 409–435. Tallant, Jonathan and David Ingram (2012). Time for distribution? Analysis 72: 264–270. Talmy, Leonard (2000). Toward a Cognitive Semantics: Concept Structuring Systems, vol. 1. Cambridge, Mass.: The MIT Press. Tatevosov, Sergei (2008). Subeventual structure and non-culmination. In Olivier Bonami and Patricia Cabredo Hofherr (eds), Empirical Issues in Syntax and Semantics, vol. 7. Paris: CSSP, 393–422. Tatevosov, Sergei (2020). On the temporal structure of nonculminating accomplishments. Linguistics 58(5): 1323–1371. Tatevosov, Sergei, and Mikhail Ivanov (2009). Event structure of non-culminating accomplishments. In Lotte Hogeweg, Helen deHoop and Andrej Malchukov (eds), Crosslinguistic Semantics of Tense, Aspect, and Modality, vol. 148. Amsterdam/Philadelphia: John Benjamins Publishing Company, 83–130. Thomason, Richmond H. (1970). Indeterminist time and truth-value gaps. Theoria 36: 264–281. Tho¨nes, Sven, and Kurt Stocker (2019). A standard conceptual framework for the study of subjective time. Consciousness and Cognition 71: 114–122. Thornton, Agathe (1987). Maori Oral Literature as Seen by a Classicist. Dunedin: University of Otago Press.

REFERENCES

345

Todd, Steven J. (2009). A difference that makes a difference: Passing through Dennett’s stalinesque/orwellian impasse. The British Journal for the Philosophy of Science, 60(3): 497–520. Tonhauser, Judith (2011). Temporal reference in Paraguayan Guaranı´, a tenseless language. Linguistics and Philosophy 34: 257–303. Tooley, Michael (1997). Time, Tense, and Causation. Oxford: Clarendon Press. Tooley, Michael (1999). The metaphysics of time. In Jeremy Butterfield (ed.), The Arguments of Time. Oxford: Oxford University Press, 21–42. Tooley, Michael (2000). Time, Tense, and Causation. Oxford: Oxford University Press. Torrengo, Giuliano (2014). Ostrich presentism. Philosophical Studies 170: 255–276. Torrengo, Giuliano (2017a). Feeling the passing of time. The Journal of Philosophy 114(4): 165–188. Torrengo, Giuliano (2017b). The myth of presentism’s intuitive appeal. Phenomenology and Mind 12: 50–55. Torrengo, Giuliano (2018). Perspectival tenses and dynamic tenses. Erkenntnis 83(5): 1045–1061. Traugott, Elizabeth Closs (1978). On the expression of spatio-temporal relations in language. In Joseph H. Greenberg, Charles A. Ferguson and Edith A. Moravcsik (eds), Universals of Human Language, vol. 3. Stanford: Stanford University Press, 369–400. Truswell, Robert (2011). Events, Phrases and Questions. Oxford: Oxford University Press Tse, Peter Ulric, James Intriligator, Josée Rivest, and Patrick Cavanagh (2004). Attention and the subjective expansion of time. Perception and Psychophysics, 66(7): 1171–1189. Tsunoda, Tasaku (1981). The Djaru language of Kimberley, Western Australia (Pacific Linguistics. Series B 78). Canberra: Pacific Linguistics. Tye, Michael (2002). Representationalism and the transparency of experience. Noûs 36(1): 137–151. Tye, Michael (2003). Consciousness and Persons: Unity and Identity. The MIT Press. Tye, Michael (2008). The experience of emotion: An intentionalist theory. Revue Internationale de Philosophie 243: 25–50. Vafaeian, Ghazaleh (2018). Progressives in Use and Contact: A Descriptive, Areal and Typological Study with Special Focus on selected Iranian Languages. PhD dissertation, Stockholm University. Van Linden, An, and Jean-Christophe Verstraete (2008). The nature and origins of counterfactuality in simple clauses: Cross-linguistic evidence. Journal of Pragmatics 40(11): 1865–1895. Van Olmen, Daniël (2018). Reproachatives and imperatives. Linguistics 56(1): 115–162. Van Valin, Robert D. (2001). An Introduction to Syntax. Cambridge: Cambridge University Press. Velleman, J. David (2006). So it goes. The Amherst Lecture in Philosophy 1: 1–23. Vendler, Zeno (1957). Verbs and times. Philosophical Review 66(2): 143–160. Verhulst, An, and Renaat Declerck (2011). Constraints on the meanings of modal auxiliaries in counterfactual clauses. In Tanja Mortelmans, Jesse Mortelmans and Walter De Mulder (eds), In the Mood for Mood (Cahiers Chronos 23), 21–42. Leiden/Boston: Brill. Verstraete, Jean-Christophe (2005). The semantics and pragmatics of composite mood marking: The non-Pama-Nyungan languages of northern Australia. Linguistic Typology 9(2): 223–268. Verstraete, Jean-Christophe (2006). The nature of irreality in the past domain: Evidence from past intentional constructions in Australian languages. Australian Journal of Linguistics 26(1): 59–79.

346

REFERENCES

Vlach, Frank (1993). Temporal adverbials, tenses and the perfect. Linguistics and Philosophy 16(3): 231–283. Voegelin, C. F., and F. M. Voegelin (1969). Hopi /ʔas/. International Journal of American Linguistics 35(2): 192–202. Vuillermet, Marine (2012). A Grammar of Ese Ejja, A Bolivian Language of the Amazon — Grammaire de l’ese ejja, langue tacana d’A mazonie bolivienne. PhD dissertation, Université Lyon 2. Vuillermet, Marine (2018). Grammatical fear morphemes in Ese Ejja - Making the case for a morphosemantic apprehensional domain. In Maïa Ponsonnet (ed.), Morphology and Emotions across the World’s Languages. Amsterdam: John Benjamins Publishing Company, 256–293. Wallach, Hans, and Lucretia Floor (1971). The use of size matching to demonstrate the effectiveness of accommodation and convergence as cues for distance. Perception and Psychophysics 10: 423–428. Wearden, John (2015). Passage of time judgements. Consciousness and Cognition 38: 165–171. Wearden, John, Alan O’Donoghue, Ruth Ogden, and Catharine Montgomery (2014). Subjective duration in the laboratory and the world outside. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press, 287–306. Weingard, Robert (1977). Space-time and the direction of time. Noûs 11(2): 119–131. Westerlund, Torbjo¨rn (2015). A Grammatical Sketch of Ngarla (Ngayarta, Pama-Nyungan). Anu, A.C.T: Asia-Pacific Linguistics. Whitehead, Alfred North (1929). Process and Reality: An Essay in Cosmology. Cambridge: Cambridge University Press. Whiteley, Wilfred Howell (1960). The Tense System of Gusii. Kampala: East African Institute of Social Research. Wiener, Norbert (1914). A contribution to the theory of relative position. Proceedings of the Cambridge Philosophical Society 27: 293–304. Wierzbicka, Anna (1996). Semantics: Primes and Universals. Oxford: Oxford University Press. Wilkins, David (1989). Mparntwe Arrernte (Arand): Studies in the Structure and Semantics of Grammar. PhD thesis, Australian National University. Williams, Donald. C. (1951). The myth of passage. The Journal of Philosophy 48(15): 457–472. Wilson, Jessica M. (2021). Metaphysical Emergence. Oxford: Oxford University Press. Wilson, Jessica M. (2022). Some thoughts on the emergence of spacetime. Paper presented at the Chronos Workshop, Rethinking and Communicating Time project, Universitat Autònoma de Barcelona. Wittmann, Marc (2014). Embodied time: The experience of time, the body, and the self. In Valtteri Arstila and Dan Lloyd (eds), Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge, Mass.: The MIT Press, 507–524. Zakay, Dan, and Richard A. Block (2004). Prospective and retrospective duration judgements: An executive control perspective. Acta Neurobiologiae Experimentalis 64: 319–328. Zaslavsky, Noga, Jennifer Hu, and Roger P. Levy (2020). A Rate-Distortion view of human pragmatic reasoning. DOI: 10.48550/arXiv.2005.06641 Ziegeler, Debra (2000). What ‘almost’ can reveal about counterfactual inferences. Journal of Pragmatics 32(12): 1743–1776.

REFERENCES

347

Ziegeler, Debra (2015). Calamities and counterfactuals: A historical view of polarity reversal. Anglophonia: French Journal of English Linguistics 19(19). DOI: 10.4000/anglophonia.520 Zimmerman, Dean (2005). The A-theory of time, the B-theory of time, and taking tense seriously. Dialectica 59(4): 401–457. Zlatev, Jordan (2016). Turning back to experience in Cognitive Linguistics via phenomenology. Cognitive Linguistics 27: 559–572.

Index 2D framework 5, 60–84 abduction 299, 303, 306, 312–13, 315 ablative 138 absolute tenses 185–7, 189, 281–7, 311 absolute time 251, 265, 267, 272 Acceptability operator 268–9, 273–4 accuracy conditions 220 actional control 114, 116, 119, 154 action dependent ability 160 action sentences 289 actuality entailments 165–7, 169 adverbials 94n10, 119, 136, 143, 260, 264, 265, 269 adversatives 112, 145, 147, 148 affective effects 104, 105, 106 see also emotion African languages 112 ‘Aged Man Remembers April, An’ (Fischer, 1984) 37–8, 40 agency see also ego degrees of commitment 270–1 deictic centre 45, 261 evidentiality 161 godlike 27 misidentification of agent 255 passage of time 45, 171, 266 private states 173 time directionality 204n7, 217 volitionality 143, 159, 160, 167 agentive ordering of time 171, 216, 217, 261 agentive verbs 138, 145, 167 Aiken, Conrad 45–6, 57 Aikhenvald, Alexandra Y. 255 Aktionsart 114, 118, 153 Albert, David Z. 216 alethic expressions 93–4, 98, 100, 268 Alexandrova, Anna 112 all-powerful time 42–3

Amazonian languages 110, 112 see also Pirahã Amis, Martin 244, 252 ‘An Aged Man Remembers April’ (Fischer, 1984) 37–8, 40 anaphora 149, 270n39, 292n15 Ancient Greek 67, 271 animals and time 249n8 Anindilyakwa 121, 122, 124, 144, 150–1, 168 anisotropy 197n26 Anscombe, Gertrude E. M. 74n11, 218n14 anterior future 102 anticipation of future events 23, 214 antirealist perspectives 202, 205 antiresultatives 113, 116, 148 apparent motion/phi-motion phenomenon 231 apprehensionals 109n3, 113, 170 Aquinas, Thomas 247 Arapana 128, 143, 146 Arawak languages 255 Arc Pair Grammar 292n15 argument from design 171 argument structure 119 Arrernte 128, 143, 146 Arstila, Valterri 40, 52, 55, 56, 57, 58, 233, 233n23 aspect avertive/frustratives 6, 113–16, 118, 153–4 as means to express temporal reference 264 philosophy 67–8, 72–8 Pirahã 9, 278, 281–7, 293, 315 processes versus states 67 tense logic 62–3 ‘When am I?’ 5, 60, 61, 70–1, 73, 74–5 atelicity 73–4, 75, 78, 118 atemporality 62 A-theory

INDE X directed nowness 202 directionality of time 199, 200, 205, 208, 211n12 and dynamic time 271n42 flow of time 247n5, 249n10, 253–4 poetics 39, 58 tensed theory 4, 11–12, 270n40 Atlas-Kempson thesis 250n11 attitudinalism 256n25 Austin, J.L. 80 Australian indigenous languages 6, 107–73 Austronesian languages 112 autobiographical narratives 57 automatic processing 37 autotelic activities 257 auxiliaries 95, 139 avertives/frustratives 6, 107–73 awareness, perceptual 220, 227–9, 235, 242 Aymara 41n12, 251n14, 271 Bachelard, Gaston 35 background information 97, 105, 278, 296, 304–5 Baltic languages 112 Bar-el, Leora 119, 149 Barham, Larry 277, 307 Bar-Hillel, Yehoshua 314 Barnes, Elizabeth 81 Baron, Sam 271n42 Bednall, James 121n18, 122, 151 beliefs epistemic modality 93 flow of time 245, 247–50, 255, 258, 259, 261, 273 temporal transparency 242 time as complex concept 263–74 universality of 266 Belnap, Nuel 178 Bergen, Benjamin 54 Bergson, Henri 35 Bernard, Timothée 165 Billinara 127, 133n58 Bininj Gun-wok 125, 132–5, 139–41, 142, 143, 144, 145, 146, 148, 149, 155, 161, 162 Blackburn, Patrick 65, 66, 68, 79 Blake, Barry J. 131 Block Universe 205, 247, 260, 261

349

bodily experience see embodied cognition; sensory experience Bohnemeyer, Ju¨rgen 265 Bordini, Davide 18n9 Bourne, Craig 70, 71, 254n20 Braddon-Mitchell, David 70 brain 55n16, 208, 249, 303 see also cognition brain time 226, 230, 233–4 branch attrition 178, 181–4 branching future 6–7, 176, 177–98 Briggs, Rachael 81 Broad, C.D. 79n16 Brown, Jessica 82 B-theory attitudinalism 256n25 date theory versus token-reflexive theory of the past 15–16 directed passage 202 directionality of time 197, 199, 200, 202, 203, 204, 205, 207, 208, 211n12 egocentricity 4, 14–15 first-person redundancy 27 flow of time 4, 7, 11, 245, 246, 247, 254 metaphysics of time 11, 254 modelling the future 189–94 ordering of events 3, 4, 11 order of events 12–13 poetics 39, 58 static time 7 ‘tenseless’ 11 truth conditions 15–17 ‘Burnt Norton’ (Eliot) 39 Bybee, Joan 155 calendar time 37–8, 62 see also dates Callender, Craig 214, 216, 252, 253, 273 Cameron, Ross 81, 254n20 capacitative 152, 153, 160, 167 Cappelen, Herman 13, 25, 256 Caquinte 118 Carol, Javier 117, 160 Caudal, Patrick 118n15, 121, 122, 141, 147, 148, 165 causal indexicality 21n14 causation 187, 204, 214, 253 centring 77–9, 83 Chadha, Monima 256n25 Champollion, Lucas 165

350

INDE X

change directed change 210 directionality of time 204 as feature of temporal experience 177 illusion 208 logic 68–9 of meanings 81–3 passage of time 252 versus series of objects 248 temporal qualia 212–13 veridicalism 217 change-of-state 114, 153, 160 chaotic systems 191, 197 Chief, Lian-Cheng 150 Chomsky, Noam 175, 304n27 Chorote 117 Cinque, Guglielmo 89n6 circadian physiological changes 55 clause chaining 148, 151 Clay, E.R. 39 Clendon, Mark 119 clitics avertive/frustratives 118, 124–6, 131, 135, 136, 143, 147–9, 161 Pirahã 298 clock time 31, 35, 48, 52, 53, 58 cognition absolute tenses 311 avertive/frustratives 155–69 and culture 278 degrees of temporal experience 207–10 embodied cognition 5, 33, 54 epistemic possibility 189 interaction of time-cognitive categories 110 learning from ‘tenseless’ languages 309–10 sensationalism 221 time as form of thought 175 transparency 236 cognitive linguistics 31, 33, 35n3, 41–2, 47–8, 53–4, 252 cognitive psychology 31, 33, 51, 55 cognitivism/cognitive error theory 208, 211, 214–16 coherence 24, 176, 185, 187, 192, 218–19 Collins, J.M. 82 commitment, degrees of 266, 268–9, 272–3, 274

complex concepts 263–74 complex event structures 109, 164–5 complex systems 191, 253, 275 compositionality 9, 42, 277, 294–310 compression of time 51, 57, 257, 258 computational economy 248 Comrie, Bernard 66, 67, 72, 73, 75, 101, 102–3, 278n5 concepts of time 248–50, 258, 262, 265–6, 273, 275 conceptual integration networks/blends 47–8, 53–4 conceptual semantics 263–74 conditional 65, 85–106, 170, 286 Condoravdi, Cleo 89n5, 92n8, 96, 99n14 consciousness duration 35 and flow states 257 inference 303 internal time 34 temporal processing 37 time as flow of 247 conscious pragmatic inference 269 contingency 69, 178–80, 189, 190, 193 contradictions 79–81, 178–9 Copley, Bridget 115, 117, 118, 168, 169 corpus studies 32–3, 274 co-speech gestures 41n11 counterfactuals 2D framework 77, 78–9 avertive/frustratives 108, 138–9, 140, 157, 161–4 as part of volition-avertive continuum 117 periphrastic modals 143 as pragmatic inference 5–6, 85–106 covert tenses 264n31 creativity 42, 44, 53 Csikszentmihalyi, Mihaly 257 C-theory 3, 7, 199–219 culture compositionality in meaning 295 cultural influence on semantics 9 dark matter of the mind 310–11 directionality of time 271 inferential meanings 304, 310–11, 315 language-culture symbiosis 311 linguistic relativity 267

INDE X metaphors for time 41 semi-propositional representational beliefs 261 socio-cultural theory of time 169–72 taboos and constraints 311 tacit knowledge 280 tense comprehension in Pirahã 278, 279–80, 296 timeline organizations 272 cutaneeous rabbit illusion 231 Danto, Arthur 81 dark matter of the mind 280, 310–11 Da-sein (‘Being in-the-world’) 34 dates 62–3, 64, 66, 68 see also calendar time Davidson, Donald 289 Davis, Henry 119 ‘Dead Pope, The’ (Meredith) 42 Deasy, Daniel 254n20 death 43, 44, 49 decessives 118 Declerck, Renaat 86n3, 105, 289 deduction 174, 190, 293, 299, 303, 306, 312–13, 315 Default Semantics 9, 247, 250n11, 266–8, 273, 274 definite/indefinite statements 61–4 deflationism 236, 238, 239, 240, 249–50 degree of futurity 23 degrees of commitment 266, 268–9, 272–3, 274 degrees of temporal experience 207–10 deixis see also ‘now’ agentitive ordering of time 261 avertive/frustratives 110 deictic centre of agency 45, 261 deictic shift 90–1, 99, 101, 103 future-in-the-past 102 ‘here’ 12, 14, 27 metaindexicality 251 modal verbs 90 reported thought/speech 90–1 speech time 89, 93, 101–2, 104, 287–9, 291, 292 speech time as anchor 289 temporal marking in languages 264n31 Delancey, Scott 161 Demirdache, Hamida 132

351

demodals 166 Dendale, Patrick 98 Deng, Natalja 211n12 Dennett, Daniel 232n20, 232n22, 253, 256n25 density of experience 257 deontic expressions avertive/frustratives 133, 139, 152, 160, 166 counterfactuals 87, 88, 91, 92, 94, 95, 99, 100 depression 52n14 depth perception 21–2 de se 254–6 desideratives 287n11 determinism 252 Detges, Ulrich 148 development paths 155–64 Dever, Josh 13, 25, 256 Dickinson, Emily 58–9 dimension 5 3 directed block 202 directed change 210 directionality of time 7, 36n4, 186, 196–7, 200–6, 251, 271–2 directionless time 199–219 directionless vection 212 discourse 108, 148, 151–2, 244–5, 263–74, 300 Discourse Representation Theory 295n18, 300 discursive presuppositions 86, 87, 97 distortion of time perception 40, 46–7, 48, 50, 51–3, 57, 257–8 see also protracted time Dobson, Henry Austin 43–4 Donnellan, Keith 82 Double Access Readings 289–90 double indexicality 255–6, 259 Dowty, David 68, 79, 108, 117, 294 doxastic meanings 152–3, 159–60, 161, 162, 163, 167 Droit-Volet, Sylvie 46, 50, 52, 55 ‘Dry Salvages, The’ (Eliot) 48–9 dual system views 249 dubitatives 162 Dummett, Michael 193 duration and aspect 73

352

INDE X

duration (Continued) durative intonations 132, 144, 147 illusions 231 philosophy 35 poetics 50–3 protracted time 52, 257, 258 sensationalism 221 and states 67 subjective time 36 temporal transparency 7–8, 224, 225–31, 238 duration judgement tasks 50 Dyke, Heather 18, 19, 26, 29 dynamic modals 152, 167 dynamic time 8, 68, 77, 241n37, 244–75 Earman, John 202, 219n16 echoic language 89n5 ecological realism 57, 96 ego dynamic time 271 egocentric inferential isolation 25–6 egocentric mental states 4, 13, 20–5 egocentric predicative terms 14–15 emergent ego 245 enduring self 214, 248, 256n25 flow of time 8, 245, 246, 248, 250–5, 257, 259, 266, 267 metaindexicality 270 tense and emotion 11–29 Ego-moving metaphor of time 41 Einstein, Albert 107, 176, 178n5, 197, 241n37, 251, 253 elaboration 47 Elias, Norbert 35 Eliot, T.S. 38n8, 39, 40, 48–9, 57 ellipsis 145, 148, 166 El Refaie, Elisabeth 52n14 embedded clauses 289, 292 embedded sentences 297–8, 301n24, 302–3 embodied cognition 5, 33, 54 embodied semantics 54–5 embodied subjectivity 34 embodied temporality 53–6 emergent ego 245 emergent nature of time 2–3, 8, 246, 250–3, 260, 266, 267, 273

emotion affective effects and counterfactuality 104, 105, 106 distortion of time by emotions 51–2 distortion of time perception 257 past conditionals 104 and tense 11–29 time-emotion paradox 46–7 emphatics 168 enduring self 214, 248, 256n25 enemy construals of time 43 English aspect 315 counterfactuals 157 evidentiality 255 future-in-the-past 101 modal verbs 89 past conditionals 85, 86, 105 perfect tenses 287–8 similatives 161 time words 281 English PoetiCog 32 entailment 63, 108n2, 143n73, 165–6 entropy 201, 203, 214, 215–16, 217, 257 epistemic adverbs 94n10 epistemic asymmetry 176–7, 179, 188, 197 epistemic modality 87–96, 167, 245, 246, 266, 267–70, 273 epistemological priority 289 Ese Eija 115–16 essential indexical 12–13, 14 eternalism 68–9, 190, 254n20 eternity 280 evaluations 29n21, 93n10 evanescence of time 46 Evans, Nicholas 133, 134, 138, 139, 140, 141, 142n71, 143, 144, 145, 146, 148, 149, 150, 151, 161, 162, 271 events avertive/frustratives 108, 109, 164–5 duration 67 event inertia 117–18 event time 287–8, 291, 292 past conditionals 101 perceptual experiences 241 pluractional imperfective 144–7 eventualized postmodal meanings 166

INDE X Everett, Daniel L. 276, 277, 278–9, 281, 284n8, 285, 287, 292, 293, 293n15, 294, 295, 296, 297, 298, 299n21, 301n24, 302, 303, 307, 310, 311, 314 evidentiality 98–9, 140, 151–2, 161–4, 255, 260, 266, 281 evolution 4, 17–20, 175, 233, 262 evolving frame 254, 260–1 experience of time passing see passage of time; perceptions of time experimental research 37, 50, 53, 249, 304 extension (metaphorical) 44–5 extensionalism 229, 230, 231, 247n4 facial expressions 55 ‘failure’ 110, 120, 142, 159, 163, 166 Farr, Matt 200n2, 202, 203n6, 206, 207, 209 Fauconnier, Gilles 35n3, 48, 53 features 220n2, 221, 223, 238 feelings see also emotion flow of time 248, 250, 257 not the same as emotions 28 figurative construals 32, 33, 43, 46, 47, 51–2, 54 see also metaphor Fine, Kit 185, 186, 187 first-person perspectives 13, 14, 15, 255, 261see also ego first-person redundancy 20–5, 26–7, 29 Fischer, Luke 56 Fischer, Raymond 37–8, 40 Fischoff, Baruch 181 Flaherty, Michael 35n3, 36, 41, 57, 250, 257 flow (state of ) 257 flow of events 108, 111, 113, 171, 173, 237 flow of time A-theory 247n5, 249n10, 253–4 awareness of 3, 236 beliefs 245, 247–50, 255, 258, 259, 261, 273 B-theory 4, 7, 11, 245, 246, 247, 254 consciousness 247 directionality of time 199, 210–11, 218 ego 8, 245, 246, 248, 250–5, 257, 259, 266, 267 enduring self 214 feelings 248, 250, 257 as illusion 248

353

incoherence of time direction 218 indexicality 254 inner flow 235–43, 247 metaindexicality, metarepresentation and basic concepts 32, 244–75 metaphors for time 44, 45, 261 metaphysics of time 247 metarepresentation 256, 260–3 and the ‘now’ 254 temporal transparency 220–43 Forbes, Graeme A. 60, 61, 72, 73, 74, 81 foreclosed counterfactuality 85n2 formal semantics 89n5, 107, 108, 165, 269 formulaic expressions 51 Forrest, Peter 60, 72, 73 fragmented models of time 185 free indirect speech 90–1, 106 Frege, Gottlob 23, 68, 294, 295 French 85–106, 115, 160n90, 166–7, 170 front/back axis for future/past representations 41 frustratives 313 see also avertives/frustratives fully automatic high-quality translation (FAHQT) 314 functional indexicals 256 functionalism 248 functional propositions 268–9 future B-theory 190 double indexicality 255 future conditional 99n12 future contingents 69–70 future indeterminacy 69, 74, 80, 82 marking in Pirahã 282, 286 modelling the 174–98 poetics 49 tense logic 63 Wari’ 298 future-in-the-future 65 future-in-the-past 65, 101–4, 120 futurelessness 49 future tenses 65, 95, 103–4, 139, 180n6, 190 future-to-past ordering 201, 202–4, 214, 217 Galton, Antony 251n15 Garcı´a Salido, Gabriela 159

354

INDE X

Garde, Murray 133n59 Gayral, Françoise 294 Generativism 289 gesture 296, 315 Gibson, Edward 302 Gil, Sandrine 46, 55 Goddard, Cliff 137, 138, 145, 160 godlike perspectives 27–9, 69, 192–3, 280 Gold, Thomas 203 Goonyiandi 126, 133, 134, 144, 147, 150, 162, 163 Gow, Laura 220n1, 223, 238 grammaticalization avertive/frustratives 6, 107–73 dynamic time 263, 265, 267 grammaticalization theory 116–17, 171 ‘tensed’ languages 264n31 tenseless languages as non-grammaticalized 278n5 Green, Rebecca 134, 135, 136, 157n87, 163, 164, 170 Grey, James 137 Grice, Paul 268, 313 Growing Block theory 254n20 Gurr-goni 126, 133, 134, 135, 136, 157n87, 163 habits of inference 299 habitual aspect 72, 145, 158, 315 Hacquard, Valentine 167 Haiman, John 109, 116 Hale, Kenneth 111, 115, 148, 153n83 Hall, Edward T. 279–80 Harley, Heidi 115, 117, 118, 168, 169 Hartle, James B. 216 Hartshorn, Charles 299 Harvey, Mark 151n82, 152 Hausa 265 Hawking, Stephen W. 273 hearsay evidentiality 98 Hedden, Brian 182n8 Heidegger, Martin 34, 247, 261n28 Heine, Bernd 117, 156, 157n88 Hercus, L. 143, 146 ‘here’ 12, 14, 27 hierarchy of simplicity 265 Higginbotham, James 289 Hindi 149 hindsight bias 181–2

Hintz, Daniel J. 158 Hoerl, Christoph 209, 223n8, 230n17, 230n18, 241n36, 249, 249n9, 249n10 Hollebrandse, Bart 302 Hopper, Paul J. 156 Hornstein, Norbert 276, 289, 292 hortatives 129n48 ‘House of Dust, The’ (Aiken) 45–6 Hua 116 Hume, David 174 Husserl, Edmund 34, 241, 247, 249n10, 272 Huxley, Aldous 47–8 hyperbole 51 ‘Hyperion’ (Keats) 50–2, 57 iconicity 44, 45, 306 idioms 294 illocutionary force markers 296 illusions 58, 199, 207–8, 211–14, 231–4, 248, 253 Immediacy of Experience Principle 311 immunity to error through misidentification (IEM) 255–6 imperatives 133n58 imperfective aspect 2D framework 72–8 avertive/frustratives 132, 144–7, 154 conditional 98, 103–4 development paths 155–64 imperfective paradox 108 implicature avertive/frustratives 108n2, 131, 142, 143, 157, 159, 166, 168 semiotic inference 294–5, 313 inceptive 129n25 inceptive-iterative meanings 146, 158 inchoative verbs 137, 153 inconsequential frustratives 113–14, 116 indexicality see also deixis; metaindexicality de se 256 dynamic time 270 and egocentric mental states 13 epistemic modality 89 essential indexical 12–13, 14 first person perspectives 245 and the flow of time 254 pronouns and egocentric predicates 13–15

INDE X reference points 102 self-location 26–7 indicative mood 102 indirect evidentiality 98 indirect speech acts 296 induction 174, 190–1, 281, 299, 303, 306, 312–13, 315 ineffability 195 inertia 108, 117–18, 160 inexplicability of humanity 175 inferential isolation of egocentric thoughts 25–6 inferential meanings 38, 58–9, 85–106, 267–9, 276–318 inheritance 227–9, 231, 234 innate knowledge 299n21 inner flow 235–43, 247 intensionalism 247n4, 295 intentionality 74, 292 Intentional State Constructions 297–9 internal clock 52–3 internal time 34, 235 see also inner flow intersubjective modality 93 intervention, asymmetry of 216 introspection 229, 233, 235, 238–9, 240, 300, 304, 306 intuition 300, 303, 306 irrealis 108, 221 irrealis-avertive inflections 120–3, 126–34, 155–72 irreversible processes 201, 253 Ismael, Jennan T. 214, 216, 252–3, 254, 256, 260, 273 iterativeness 124–6, 144, 145, 146, 157–8, 170 Iwaidja 121, 122, 124, 133, 136, 138, 141, 142, 147, 148, 161, 162, 168n101 Jackman, Henry 82 James, William 39, 40, 52, 55, 230n17, 247 Jaminjung 121, 122, 123, 125, 136, 142, 147, 168n101 Jaszczolt, Kasia M. 32, 58, 235n25, 235n26, 241n37, 245, 255, 261n28, 265n33, 266, 267, 268, 273, 277, 295, 296, 303, 307n29 Jørgensen, Klau Frovin 65, 66, 68, 79

355

Kahneman, Daniel 249 Kalkatungu 131 Kamp, Hans 68, 79, 108, 114, 295n18 Kant, Immanuel 34, 175, 176, 247, 248 Kapitonov, Ivan 133, 141 Kaplan, David 26 Kappa effect 257 Karlsson, F. 307 Kayardild 127, 138, 139, 142, 143, 146, 150, 151, 161, 168, 169, 170, 308 Keats, John 50–2, 57 Kern, Barbara 297, 298 Kinsbourne, Marcel 232n20, 232n22 Koenig, Jean-Pierre 150 Kothari, Anubha 149 Kozlov, Alexey 158 Kroeger, Paul 119, 119–20, 168 Kronning, Hans 93, 94 Kunbarlang 124, 133, 141–2, 156n85, 161 Kutach, Douglas 216 Kuteva, Tania 112, 113, 116, 117, 118, 159, 159n89 Lakoff, George 42, 44, 47 Landman, Fred 108, 117–18 language communities 77 Larrimore, Bonnie 110 Larrivée, Pierre 85–6 Lauer, Sven 99n14 Laughren, Mary 161 laws of nature 189, 190, 197, 201, 203, 252 see also physical laws of the universe Leder, Harry 289 Leeding, Velma J. 144 Legate, Julie Anne 129n44, 129n45 ‘L'Envoi’ (Strobel) 52 Le Poidevin, Robin 210–12 Levinson, Stephen C. 265, 275 Lewis, David 195 lexicalization 110, 117, 169, 263, 265, 267 lexicon-grammar-pragmatics trade-offs 267–8 lexification 117 linguistic relativity 264–7 Lithuanian 117n13 Lloyd, Dan 40, 52, 55, 56, 57, 58 Loewer, Barry 216 logic 61–70, 75, 189, 287, 290

356

INDE X

Ludlow, Peter 272n44 Lyons, John 87, 93 MacFarlane, John 69, 178, 179–80, 181, 182, 188, 192–3, 194, 195 machine translation 314 Maclaurin, James 18, 19, 26, 29 Mailhammer, Robert 121, 141, 147, 148, 162n95 Malchukov, Andrej L. 112, 116, 148 Mandarin 149, 150 Maori 271 Mari, Alda 160 markers of time 226, 233, 238 Martin, Fabienne 132, 160 Martin, Michael G. F. 222 mathematical principles 313 mathematical representations of time 219n16 Matthewson, Lisa 119 Maudlin, Tim 202, 205 Mawng 124, 132–5, 149, 150 McArthur, Robert P. 61, 62, 64 McCall, Storrs 178, 182, 185, 197n26 McConvell, Patrick 161 McCormack, Teresa 249 McGregor, William 134, 143n73, 144, 147, 150, 163 McTaggart, John M.E. 3, 4, 7, 39, 177, 200, 203, 247n5 Meakins, Felicity 129n47, 130n51 Mẽbêngôkre 117 Mellor, D. H. 16, 202 memory continuity of the ego 256n25 duration judgements 50, 52 enduring self 214 misidentification of agent 255 order of events 272 and the passage of time 187, 215, 217 temporal processing 38 mental representations of time 35–6, 175–6, 220, 237, 239 mental simulation 54, 55–6 Meredith, Owen 42 Merleau-Ponty, Maurice 34 metaindexicality 8, 245, 250–60, 261, 263, 265–7, 270–5 meta-metaphysics 248n7

metaphor cognitive semantics 252 composition of conceptual metaphors 42 degrees of temporal experience 209 elaboration 47 extension (metaphorical) 44–5 flow of time 261 metaindexicality 251 passage of time 41–9 poetics 30–1, 33, 41, 53–6 questioning 44 sequential markers in Pirahã 284 spatial metaphors 41–2 timelines 261, 279 metaphysics of time 2D framework 76–9 aspect 61, 67, 72–8 branch attrition (future modelling) 181 B-theory 11, 254 Default Semantics 266 directionality of time 202–4 dynamism 68 flow of time 8, 247 and linguistics 245 metaphysical emergence 2–3 metarepresentation 261–3 order of events 39, 272–3 passage of time 248 phenomenology of time 34 possible worlds discussions 78 tense logic 61–70, 75 transparency 223, 224, 226–8, 230, 237–8, 239 ‘When am I?’ 70–1 metarepresentation 246, 250, 251, 256, 258, 260–3, 267, 271, 273, 274–5 metatemporality 249 micro-laws of physics 3, 7, 252 Miller, Izchak 208 miratives 161, 167 misidentification errors 255 ‘mistaken thought’ 129n46, 137, 161, 162, 164, 167, 171 Modal-Contextualist view 273 modality Acceptability operator 268 Australian languages 110

INDE X avertive/frustratives 116–23, 131, 132–5, 137, 151–2, 156 counterfactuals 5–6, 108 epistemic modality 87–96, 167, 245, 246, 266, 267–70, 273 events in the process of taking place 108 and evidentiality 266 habitual aspect 72 as means to express temporal reference 264 modal realism 195 modal supervenience 268, 271, 272 modal verbs 85–106 non-verbal modalities 41n11 periphrastic modals 142n72 postmodality 120, 137, 160n90, 166 prospective aspect 74 Reichenbach’s temporal theory 287n11 subjective-epistemic modality 87, 93, 97, 101, 103 Moens, Marc 68, 79 Moksha Mordvin 158 monadic predicates 20–1 Montague, Richard 295 Moore, George E. 238n31 Moore, Kevin Ezra 271 Morpurgo Davies, Anna 271 motion apparent motion/phi-motion phenomenon 231 directionality of time 194, 204, 217, 218 flow of time 45, 54 motion illusions 212 motion in space 261 motion verbs 47, 260 passage of time 48 poetics 31 motor system 55, 56 Moving Spotlight theory 254n20 Mozersky, Joshua M. 188, 248n7 Mucha, Anne 265 Multiple Drafts model 232n22, 233n23 Murrinh-Patha 121, 125, 143, 156 Nagel, Thomas 192 naïve realists 221 naïve theory of time 249 narrative present 75, 90, 103 Nash, David George 129n46, 138, 162

357

Natural Semantic Metalanguage 263n30 negation 120, 132, 143, 145, 148 negative events 111, 118, 152, 157, 164–6, 172 neo-Whorfianism 265–6, 275 neuroimaging studies 274 neuroscience 55n16, 208, 249, 303 see also cognition Ngarla 127, 136–7, 145n76, 160 non-being, realm of 198 non-cognitivism 209, 211, 216–19 non-culminating accomplishments 119, 149–52, 156, 168 non-verbal modalities 41n11 Nordlinger, Rachel 121, 129n31, 129n47, 130n51 ‘now’ absolute present 185–6, 187, 189 agentitive ordering of time 261 centring 77–8 directionality of time 202 essential indexical 12, 13–14 and the flow of time 254 indexicality 12, 13–14, 256 perceptions of 187 presupposition of common 19n11 self-location 27 specious present 40 time slices model 185–6 Nuyts, Jan 93, 94n10 Nyangumarta 136n64 Nyikina 126, 133, 135, 136, 143, 146, 164 objective covert qualifiers 235n26, 259–60, 264, 269 objective-epistemic 93 objective views of time 34, 35n3, 38, 47–8, 235–6 oddball effect 231 O’Hagan, Zachary 118 online awareness 258, 263 ontic indeterminacy 69, 74, 80, 82 ontological autonomy 2–3 ontological indeterminacy 177, 180, 183–4, 188, 191–2, 197 ontology 172 order of events see also B-theory C-theory 200 cultural differences 272

358

INDE X

order of events see also B-theory (Continued) directionality of time 199–219 future-to-past ordering 201 modelling the future 186, 187 overt marking in languages 264–5 poetics 35, 38, 39, 58 temporal transparency 225–31 transparency 238 Orwellian rewriting 232–3 Osgarby, David 139 Overall, Simon E. 110, 112, 116 Pahonțu, Beatrice 157 Palmisano, Stephen 212 Papafragou, Anna 92, 94n10 Papago 111, 117, 118, 168 Papineau, David 221n3, 232n21 ‘Paradox of Time, The’ (Dobson) 43–4 Parsons, Josh 272n45 Partee, Barbara 68, 79, 269, 270, 277, 304 partial orders of time 186 particles 135, 136–7, 139, 140, 142, 143, 147–9 partitive culminations 119, 132, 149–52, 156, 168 ‘part-of ’ semantics 155 passage of time see also distortion of time perception; flow of time A-theory 205–6 antirealist perspectives 205 autobiographical narratives 57 beliefs and concepts 247–50 branching future model 178, 182–3 degrees of temporal experience 207–10 directed passage 202 directionality of time 7, 199–219 dynamic time 271 and the ego 252 experiences of 248 feelings of 257–8 human experiences of 207 as illusion 58 ‘losing track of time’ 242 mental representations of time 35–6 metaindexicality 251 metarepresentation 262 passage phenomenology 207–12 perceptions of 3

phenomenal passage cognitivism 3, 208 phenomenal passage eliminativism 3, 208 phenomenal passage illusionism 3, 207–8 poetics 41–9, 53 rejection of 71 relative passage of time 186 shifts in reality 187 temporal transparency 224, 234–43 past 2D past 60–84 continuity of the ego 256n25 double indexicality 255 fixity of 177, 178–80, 193 marking in Pirahã 281–2, 286 past progressive 75 poetics 49 truth conditions 15–17 past conditionals 85–106 ‘past hypothesis’ 216 pastism 254n20 past perfect tense 65, 190 Patard, Adeline 85–6 Pathak, Tanay 45 Paul, Laurie A. 206n8, 207, 208, 213–14 Peirce, Charles Sanders 9, 277, 279, 280, 294–310, 311–14, 315 perceptions of time see also flow of time; passage of time; sensory experience absolute present 187 contextual factors 57 culture 278–9 directionality of time 210–12 hindsight bias 181 poetics 34, 35–7, 40, 41, 54 transparency 220–1 Whorfian 171n102 perceptual awareness 220, 227–9, 235, 242 perceptual inaccuracy 232 perceptual transparency 222–4 perdurance 248, 250, 256n25 perfective aspect 2D framework 61, 72–8 avertive/frustratives 132, 144, 144n75, 150–1, 154, 170 perfect tenses 65, 66, 99, 287–93 periphrastic irrealis/avertives 135–44 permanentism 254n20

INDE X Perry, John 12, 21n14, 26 perspectivalism 254 perspectival meaning of Past Conditional 101–4 perspectival shifts 5–6 perspective taking 87, 89n5, 90–2, 193 phenomenal modifier theory (PMT) 237–43 phenomenal passage cognitivism 3, 208 phenomenal passage eliminativism 3, 208 phenomenal passage illusionism 3, 207–8 phenomenology dynamism of 237 flow of time 258 inner flow 247 phenomenal character of time 30n1, 31n2, 56–7, 223, 235n25 phenomenological experience 32, 37, 57 poetics 31–5, 56, 58 primitive phenomenal modifier 249, 257–8, 259, 264, 273, 274 transparency 223, 226–8, 234, 235, 238 Phillips, Ian 221n4, 224n10, 229n16, 230n18, 239n42, 249n10 Phillips, Joshua 120n17 physical laws of the universe 2–3, 7, 8, 200–6, 246, 252 see also laws of nature Piata, Anna 51, 53, 57 Pietarinen, Ahti 307n29 Pinson, Mathilde 161 Pirahã 9, 276–318 Plato 280 Plungian, Vladimir 113, 116, 148 pluperfect tense 65 pluractional imperfective 144–7 PoetiCog 32 poetics 4–5, 30–59 Polanyi, Michael 280 Po¨ppel, Ernst 36 Popper, Karl 174n1 Portner, Paul 133n58, 160 possible worlds discussions 24, 77–8, 108, 117, 272–3 Postal, Paul 292n15 postmodality 120, 137, 160n90, 166 posture 55 Potts, Christopher 168 Pragmatic Principle of Compositionality 307n29

359

pragmatics avertive/frustratives 132, 137, 151–2, 154, 157, 168 conscious pragmatic inference 269 counterfactuality as pragmatic inference 85–106 and Default Semantics 273 inferential meanings 85–106, 267–8 as means to express temporal reference 264 propositions 250 Predelli, Stefano 19n11, 26 predictions 74, 95–101, 105, 163–4, 214 present absolute present 185–7, 189. see also ‘now’ gestalt of perceptual moments 55 imperfective aspect 73 living present 34 marking in Pirahã 286 narrative present 75, 90, 103 perceptions of time 40 relativity of 185–6 specious present 39–40, 49, 52, 55, 241 ‘When am I?’ 70 presentational phenomenology 223 present conditional 98 presentism 71, 254n20 present perfect 102 presuppositions 97, 108n2, 157, 168 Price, Huw 204, 252, 256, 273 primitive phenomenal modifier 249, 257–8, 259, 264, 273, 274 Principle of Presentational Concurrence (PPC) 229n16 Principle of Simultaneous Awareness (PAS) 229n16 Prior, Arthur 11, 12–13, 16, 17, 26, 29, 61, 65, 66, 68, 198 probabilistic reasoning 174n1, 181, 197 processes 67, 201, 203, 213 processes versus states 72–8 process philosophy 67–8, 72–8 progressive aspect 61, 67, 72–8, 79, 108, 117, 117n12 projection 227–9, 230 propositions 68–9, 77, 250, 261, 266 proprioception 278–9 prospective aspect 74

360

INDE X

prospective judgements 50, 52, 154 prospective tense 65 Prosser, Simon 13, 30, 56, 58, 205, 207, 211, 214, 215, 248, 249n8, 278 protasis 87, 105 protention 272 protracted time 36, 51, 52, 57, 257, 258 proximatives avertive/frustratives 112, 113, 117, 119, 120, 136, 138, 143–4, 145, 152, 154, 167, 287n11 development paths 155–64 psychic entropy 257 psychology 20n13, 33–4, 36, 38, 46–7, 58 Pullum, Geoffrey K. 307, 308 Pustejovsky, James 313 Putnam, Hilary 83n20 Pym, Noreen 110 quantum gravity 2 quantum systems 191–2, 195, 245 Quechua 158 questioning 44 Quine, W.V. 67, 68 radical pragmatics 250n11 rationality 24, 175, 261, 299n21 realistic settings for research 57 Recanati, François 21, 256 recursive syntax 277, 292, 300–2, 307 reduced negative clauses 148, 150 reductionism 3, 209, 214, 240, 249n10, 271 reduplication 132, 144–7, 157–8 reference points 65, 69, 101, 108, 190, 277, 287–9, 291, 292 referential opacity 13 Reichenbach, Hans 9, 65, 66, 101, 108, 174n1, 190n19, 201, 202, 203, 204, 215, 276, 280, 284n8, 287–8, 289, 290–1, 292 relative markers in languages 264–5 relativistic time 176–98, 218–19, 241n37, 250–1, 272 Relativity, theory of 176, 189n17, 197, 250–1 relevance 86, 97, 104, 250n11 relief after events see ‘thank goodness’ reported thought/speech 90–1, 106

representationalism 8, 30n1, 220, 221, 249, 258 reprochatives 152, 160, 166 resultative aspect 281 retensionalism 229, 272 retrospective time judgements 50, 52, 195 Reyle, Uwe 114 rhetorical routines 148 rhythm 48 Richardson, Louise 223n7, 241n35 Role and Reference Grammar 297, 308 Romanian 157 Rovelli, Carlo 33, 56, 274 Rule for Temporal Connectives (RTC) 289 Russell, Benjamin 304, 313 Russell, Bertrand 107 Russian 158 Salanova, Andrés Pablo 117, 160 Salish languages 112, 119 Sansò, Andrea 157 Sapir, Edward 278 Sattig, Thomas 230n18 Sauerland, Uli 300–2 Saussure, Louis de 86, 89n5, 97, 104 Schultze-Berndt, Eva 121n18, 122, 123, 137, 147 seasons of the year 38 Second Law of Thermodynamics 201, 203 self, and the passage of time 3 see also ego self-governing systems 252–3, 270–1 self-location 26–7 semantic externalism 83–4 semantic indeterminacy 178, 180, 181–4, 188, 193 semantic primes 263n30, 267 semiotic inference 276–318 semiotics 9, 294–310 semi-propositional representational beliefs 261, 262, 263, 267, 274 Seno, Takeharu 212 sensationalism 221, 237 sensory experience 37, 40, 55, 56, 176, 187, 252, 279, 280 sentential recursion 277, 292, 300–2, 307 Sequence of Tense (SOT) structures 289–90, 292 sequential markers in Pirahã 284–5 Shoemaker, Sydney 255

INDE X Sider, Theodore 248n7 similatives 139, 140, 143, 161–4 similes 46, 51 Simon, Herbert 311 Simpson, J. 129n45 simultaneity 36, 40, 189n17 simultaneity judgement tasks 37 Singer, Ruth 150 Sino-Tibetan languages 112 Sketch Engine 32 Skow, Bradford 218n15, 254n20 Slavic languages 112 Slobin, Dan I. 263 slowing down of time 52 see also protracted time Smith, Carlota S. 113–14 snapshot views 229 social construct, time as 35 social psychology 35n3 social time 35 socio-cultural theory of time 169–72 Sokolowski, Robert 34 spacetime dynamic time 250, 256 flow of time 261 four-dimensional spacetime theory 7 phenomenology of time 33 spacetime regions and aspect 68 static time 251 where time is found 2–3 spatialization of time 41 spatial metaphors 41–2 spatial ordering 186 specious present 39–40, 49, 52, 55, 241 speech acts 65, 80n17, 86, 97, 99, 104, 255, 296 speech time 9, 89, 93, 101–2, 104, 287–9, 291, 292 Sperber, Dan 97, 250n11, 261, 262 stage-based theory 108 stage theory of persistence 17–20 Stalinesque delay 232, 233 Stalnaker, Robert C. 77, 78 states 67 static time 7, 244–7, 251, 253, 254, 255, 260, 262, 266, 273, 275 see also B-theory static universe 3, 245, 251, 253, 261–2 statistical reasoning 174n1, 181, 304

361

statives 72, 73, 75, 76, 118, 153 Steedman, Mark 68, 79 stereopsis 22 Steward, Helen 67 Stocker, Kurt 33, 35, 36, 37, 41, 50 Stokes, Bronwyn 135, 146 stopping time 45 Strawson, Peter F. 174n1 Strobel, Marion 52 subjective-epistemic modality 87, 93, 97, 101, 103 subjective flow and metaindexicality 256–60 subjective overt qualifiers 235n26, 259–60, 264, 269 subjective time 30–59 subordinate clauses 151, 285–6, 290, 301 succession of events 210, 213, 215, 217, 221, 225–31 supervaluations 194–8 synechism 314 Szabó, Zoltán Gendler 295, 307 tacit knowledge 280 Tallant, Jonathan 254n20 Talmy, Leonard 31 teleology 153 telicity 73–4, 75, 118, 153–4, 156, 170, 285 temporal auxiliaries 95, 139 temporal betweenness 200 temporal clauses 285 temporal experience see experience of time passing temporal externalists 82 temporal IEM 255–6 temporalism versus eternalism 68–9, 71 Temporally Evolving Point of View 260–1 temporal modifier theory 222 temporal-order judgement tasks 37 temporal processing 35–6, 37–40 temporal qualia 207, 209–10, 212 temporal transparency 8, 220–43 tensed statements 62 tenseless languages 264, 278, 315 tenseless reality 62–4, 265 see also B-theory tense logic 61–70, 287, 290 tense realism 189–90 tense-time mismatches 269, 270 tensism 4, 270n40

362

INDE X

Tepehuan 159 ‘thank goodness’ 11, 12, 15–17, 19–20, 29 ‘The Dead Pope’ (Meredith) 42 ‘The Dry Salvages’ (Eliot) 48–9 ‘The House of Dust’ (Aiken) 45–6 Theory of Relativity 176, 189n17, 197, 250–1see also relativistic time ‘The Paradox of Time’ (Dobson) 43–4 thinking for speaking 263 third-person perspectives 13, 24–5, 69 Thomason, Richmond 178, 194, 195 Tho¨nes, Sven 33, 35, 36, 37, 41, 50 Thornton, Agathe 271 thoughts distinction of different thoughts 23–4 feeling sequential 279 time as form of thought 175 time-asymmetry 215 time-emotion paradox 46 ‘Time Flows’ (Pathak) 45 timelines 7, 41, 178, 189–90, 196–7, 259, 261, 270n39, 272 Time-moving metaphor of time 41 Time Orientation Metaphor 41 time slices 184–9 time-telling devices 48, 49 Tohono O’odham 111, 115, 118, 120 token-reflexive truth conditions 15–16, 19 tolling bells 49 Tooley, Michael 198, 254n20 TOR (time as its own representation) model 226, 230, 233–4 Torrengo, Giuliano 18n9, 241n37, 249–50, 254, 256n25, 257–8, 259, 264, 273, 274 transcendence 34, 254, 256, 260 transientism 254n20 transparency 222–4, 294, 299, 300, 307–10 traumatic events 26 Truswell, Robert 67 truth conditions Acceptability operator 268 contextual factors 250 evolutionary utility of reactions to events in past/future 19 functional propositions 268–9 and pragmatics 273 predictions 97 tense logic 62 thoughts about the past 15–17 timelessness of 68

truth values 2D framework 5, 76–7, 78 branch attrition (future modelling) 181 branching future 179–80, 188, 193 change 68–9 contradictions 79–80 epistemic modality 266 ontic indeterminacy 69 precision and salience 81 relativity to temporal context 179–80 relativization of 272n45 subjective-epistemic modality 103 supervaluations 194–8 truth-value links across time 193 Turner, Mark 35n3, 42, 44, 47, 48, 53 Twin Earth 76–7 Tye, Michael 234n24 typology based on compositional transparency 307–10 uncertainty 89, 100, 281 undirected time 200 universal grammar 264n31 universal time (versus human time) 49 unpredictability of the future 174–98 Uto-Aztecan languages 111–12 Van Valin, Robert D. 308 vector, time as a 219 vector fields 197 Velleman, J. David 214–15 Vendler, Zeno 67, 68 Verhulst, An 86n3, 105 veridicalism 217–18 Vietnamese 41n12 ‘Villiers de l’Isle-Adam’ (Huxley) 47–8, 57 visual accommodation 21–2 volitionality 111, 117, 141n68, 143n73, 145, 152, 156, 159–60, 167 Vuillermet, Marine 115–16 Waltereit, Richard 148 Wambaya 125 Wari’ 297–9, 310 Warlpiri 127, 137, 138, 161, 162, 308 ‘watchmaker’ analogy 171 Wearden, John 35n3, 36, 50, 52, 57 Weingard, Robert 197, 219n16 Weiss, Paul 299 Westerlund, Torbjo¨rn 137, 145n76

INDE X Wharton, Tim 104 ‘When am I?’ 60, 70–1, 73, 74–5 Whitehead, Alfred North 68, 107 Wiener, Norbert 107 Wierzbicka, Anna 263n30, 267 Wildman, Nathan 61, 82 Wilkins, David 146 Williams, Donald C. 205 Wilson, Deirdre 97, 250n11

363

Witek, Maceij 255 Wittgenstein, Ludwig 218n14 Wittmann, Marc 40, 55, 56 worldviews 255 Yankunytjatjara 127, 136, 137, 138, 145, 160, 166 Yoruba 75 Yukatek Maya 264–5