Language and Imaginability [Unabridged] 1443855456, 9781443855457

Table of contents :
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
INTRODUCTION
CHAPTER ONE
CHAPTER TWO
CHAPTER THREE
CHAPTER FOUR
CHAPTER FIVE
CHAPTER SIX
CHAPTER SEVEN
CHAPTER EIGHT
CHAPTER NINE
CHAPTER TEN
CONCLUSION
BIBLIOGRAPHY
INDEX

Citation preview

Language and Imaginability

Language and Imaginability

By

Horst Ruthrof

Language and Imaginability, by Horst Ruthrof This book first published 2014 Cambridge Scholars Publishing 12 Back Chapman Street, Newcastle upon Tyne, NE6 2XX, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2014 by Horst Ruthrof All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-5545-6, ISBN (13): 978-1-4438-5545-7

For Yingchi, my wife

TABLE OF CONTENTS

Acknowledgements .................................................................................... ix Introduction ................................................................................................. 1 In the Beginning was Resemblance Chapter One ............................................................................................... 15 Resemblance in Language: Locke Chapter Two .............................................................................................. 36 Iconicity in Kant and Peirce Chapter Three ............................................................................................ 55 Meaning and Imaginability in Husserl Chapter Four .............................................................................................. 77 The Propositional Route: From Frege to Hyperintensional Semantics Chapter Five ............................................................................................ 101 A Critique of Meaning as Use Chapter Six .............................................................................................. 121 The Linguistic Route: Saussure and Vygotsky Chapter Seven.......................................................................................... 149 Cognitive Linguistics and Conceptual Blending Chapter Eight ........................................................................................... 173 Brain, Consciousness, and the Evolution of Language Chapter Nine............................................................................................ 193 Imaginability and Pragmatics: From Grice to Habermas Chapter Ten ............................................................................................. 218 Language and Imaginability

viii

Table of Contents

Conclusion ............................................................................................... 237 Language as Heterosemiotic, Intentional Construct Bibliography ............................................................................................ 239 Index ........................................................................................................ 258

ACKNOWLEDGEMENTS

In writing this book I have been indebted to a very large number of writers, and especially to those with whose arguments on the description of language I felt I had to disagree. I owe them a special gratitude because it has only been in my conversation with their positions that I have been able to put the finger on the pulse of language in a way that would allow me to carve out my path towards what I believe to be the foundational role of imaginability. I hope that in my pursuit of this explanatory avenue I have not distorted standard accounts beyond sound reasoning in order to merely score points. If an apology is needed, it is to acknowledge that my opposition to many of the dominant discourses in the study of language is the consequence of my firm commitment to intersubjective mentalism in matters of linguistic meaning. My obvious debts in philosophy are to Locke and Kant, to the phenomenology of Husserl and Ingarden, in analytical philosophy to a broad spectrum of positions from Frege to current writings on hyperintensional semantics, in semiotics above all to Peirce, in linguistics to Vygotsky and the cognitive literature. A special debt is to the ongoing work of Michael Tomasello whose empirical research has revived a number of Husserlian notions that seem to me crucial for our understanding of what goes on in meaningful linguistic communication. I acknowledge my debt to a number of brief email exchanges with Tecumseh Fitch, Michael Corballis, and David Chalmers, whose writings have had a corrective influence on my thinking, even if none of them would be prepared to take on board my emphasis on imaginability. I also extend my gratitude to my former students and colleagues. Even where intellectual disagreements were the outcome, I wish to thank Richard Hamilton, Alec McHoul and Lubica Ucnik for intensive and fruitful debates, as well as other colleagues at Murdoch University and the University of Western Australia for stimulating exchanges. A very special “thank you” goes to my old friends Professors Thomas Alexander Szlezák, eminent Plato and Aristotle scholar, and anglist narratologist Volker Schulz for the most exciting and challenging in-depth exchanges. Let me emphasize, however, that none of the faults of the book can be laid at the doorstep of any of the many scholars who have helped me improve my

x

Acknowledgements

thinking. My wife, Dr Yingchi Chu, to whom this book is dedicated, deserves my gratitude for patiently listening to my repetitive attempts at trying to clarify matters of language that seem so very obvious. Last, I acknowledge the editors of the following journals for allowing me to draw on published materials and arguments throughout the book. Philosophy Today for papers on Locke, Kant, and ‘conceptual blending’, Review of Contemporary Philosophy for the use of a programmatic paper on the role of imaginability in natural language, Linguistic and Philosophical Investigations for papers on Husserl and Wittgenstein, Rivista Italiana Filosofia de Linguaggio for papers on Saussure and Vygotsky, and Language Sciences for permission to use excerpts from my recent paper “Metasemantics and Imaginability”. Above all, I wish to thank the editorial staff at Cambridge Scholars Publishing for their prompt, polite and professional correspondence and help. HR

INTRODUCTION IN THE BEGINNING WAS RESEMBLANCE

There is a lack of clarity about the role of imaginability (Vorstellbarkeit) in our investigation … namely about the extent to which it ensures that a sentence (Satz) makes sense. (Wittgenstein, PI §395)

At the centre of this book the reader will find a suggestion that is as simply put as it is difficult to argue, for it flies in the face of most of our dominant theories about how language means. The suggestion is this. “If you are able to imagine what I am talking about and the way I am saying it, then there is meaning; if not, there is not”. And vice versa, if I am able to imagine what others are talking about and the way they do so, an event of linguistic meaning has occurred; if not, it has not. We could sum this commitment up under the umbrella of the imaginability thesis. Accordingly, the emphasis throughout the book will be on the role of imaginability in a double sense, in the sense of what must be imaginable in response to the sounds of language in order for meaning to occur and in the sense of what sort of capacity humans must have to respond in this way. With this broad goal in mind I will attempt to argue the position of an intersubjective mentalism that avoids two threatening options, the semantic privacy of solipsism and subjectivism on the one hand and, on the other, the semantics of a radical externalism, that is, a position according to which all there is to language is observable as public performance. The first option led John Locke to the aporia of not being able to reconcile language as public discourse with his private ideas, which he chose to provide word sounds with content.1 To anticipate, I will argue that to let the fly out of the bottle here is to reconcile observable public signifiers with intersubjectively sharable, indirectly public signifieds. At the same time, I want to show that externalist theories of meaning as publically observable phenomena fail because they cannot account for what happens when language deals with absent things, as the 1

See Chapter One.

2

Introduction

bulk of natural language typically does. The classic example of such a theory is Wittgenstein’s definition of linguistic meaning as use. I will try to argue a way of modifying Wittgenstein’s influential but in my view flawed notion of use to satisfy what actually happens when we engage in language.2 As such, the present project shares certain emphases with cognitive linguistics. I am sympathetic, for instance, to the two-component analysis in LCCM theory (lexical concepts and cognitive models) distinguishing between symbolic units (signifiers) and cognitive models (signifieds), 3 even if both “concept” and “models” are not simply transferable to the kind of picture I wish to draw. There are parallels in my project to be found in broad conviction with Esa Itkonen’s discussion of analogy as a central and “psychologically realistic” tool of human thinking as part of a “picture” theory of language and his Aristotelian assumption of the “ultimate unity of all types of perceptual experience”.4 Where my focus differs is in the accent I place on the all-important role of imaginative variation of realist mental scenarios. This is one reason why I regard the notion of the “image schema” limiting. Why limit our perspective to the metaphors of “container”, “balance”, “compulsion” and their associations in “conceptual blending” when the entire system of natural language stands in a resemblance relation with perceptual reality and its imaginative variations? Nevertheless I will acknowledge in some detail my debts especially to the early work of Mark Turner and his colleagues George Lakoff and Mark Johnson. 5 Likewise important to my enterprise is the 2

This will be the topic of Chapter Five. As elaborated by Vyvyan Evans in How Words Mean: Lexical Concepts, Cognitive Models, and Meaning Construction (Oxford: Oxford University Press, 1999). 4 Esa Itkonen, Analogy as Structure and Process: Approaches in Linguistics, Cognitive Psychology and Philosophy of Science (Amsterdam: John Benjamins, 2005), 85; 136f. 5 Mark Turner, Death Is the Mother of Beauty: Mind, Metaphor and Criticism (Chicago: University of Chicago Press, 1987); Reading Minds: The Study of English in the Age of Cognitive Science (Princeton: Princeton University Press, 1991); Mark Turner and George Lakoff, More than Cool Reason: A Field Guide to Poetic Metaphor (Chicago: University of Chicago Press, 1989); George Lakoff, Women, Fire, and Dangerous Things (Chicago: University of Chicago Press, 1987); George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: University of Chicago Press, 1980); Mark Johnson, The Body in the Mind (Chicago: University of Chicago Press, 1987). For a discussion of more recent works by these authors and other cognitivists, see Chapter Seven. 3

In the Beginning was Resemblance

3

work of Michael Tomasello, in which imaginability turns up in such ideas as “intention-reading”, “attention-sharing”, and the manipulation of “one another’s intentional and mental states for various cooperative and competitive purposes”. For Tomasello, as for Language and Imaginability, “it is this mental dimension that gives linguistic symbols their unparalleled communicative power”. 6 What makes Tomasello so valuable for my critical-speculative project is that it provides empirical support for my concept of imaginability. Nor can his findings be restricted to the acquisition of language in children. They apply centrally to the mastery of language by the adult native speaker. Where the book differs from the cognitive literature is that it attempts to shed light on the presuppositions that underlie some of the most influential theories of natural language and meaning. In this sense, the book is a contribution to metasemantics rather than semantics. Above all, the book is conceived as a tentative answer to the question of how our dominant theories of language in philosophy and linguistics cope with our folk-psychological opening definition of meaning as a social event requiring the successful projection of mental scenarios as a sine qua non on the part of the listener (or reader) engaged in linguistic exchanges. A cursory list of intellectual debts must also mention the philosophical background that more than just shines through the chapters of this book. In spite of my criticisms of certain positions in the analytical philosophy of language, I have always admired its rigour, clarity and elegance. The criticisms that I do make, especially in Chapters Four and Five, have to do mainly with the consequences of regarding language as a too narrowly conceived object of inquiry.7 Kant has been a deep and abiding impression on my thinking, as he has for Mark Johnson, except that I remain more firmly committed to Kant’s version of schematism and his notion of the monogram. 8 From Peirce, whose pragmatic reading of Kant I greatly admire, I take and elaborate a strong commitment to iconicity, as the central ingredient of the content of conceptuality, with the aim of revising the Saussurean signified as motivated. 9 Phenomenology has had a profound methodological influence on my thinking since the early 1960s, 6

Michael Tomasello, Constructing a Language: A Usage-Based Theory of Language Acquisition (Cambridge, Mass.: Harvard University Press, 2003), 3f.; 8; see also Chapter Seven. 7 Cf. especially the banishment of Vorstellung (understood here as “mental projection”) from meaning by both Frege and Wittgenstein. 8 Passim and specifically Chapter Two. 9 Passim and definitionally in Chapter Ten.

4

Introduction

especially in its Husserlian version and its realist extension by Roman Ingarden.10 Language and Imaginability wants to paint a broad picture of issues central to the way meaning events occur in natural language. In trying to capture, at least in principle, as many components as possible that are indispensable for a comprehensive description of language, I am taking a wide-lens perspective suitable for a generous form of pragmatics. Whatever semantic characteristics can be distilled from such a pragmatic approach, I suggest, promise a more life-like picture than formally driven approaches of what actually goes on when we use language. As to the various philosophical and linguistic positions discussed in the book, I have for the sake of convenience divided them roughly into two camps, theories that emphasize resemblance relations and theories that follow a propositional path of inquiry. Both branches can be shown to have their roots in Aristotle’s foundational remarks in De Interpretatione, De Anima, Categories, Rhetoric, Poetics and Metaphysics. Foremost, Aristotle separates arbitrary word sounds (pathema) from non-arbitrary, or motivated, resemblance relations (homoiomata) between things in the world (pragmata) and their representation via words, a resemblance which we experience as mental projections (pathemata). Word sounds are public and as such accessible to anyone. Mental projections in response to word sounds (Vorstellungen), though not publicly observable, are nevertheless “the same” for every native speaker. 11 This view is re-enforced in De Anima where Aristotle describes words as signs for Vorstellungen (pathemata, usually rendered in English as “affections of the soul”). Furthermore, pathemata should not be regarded as detailed pictures of reality but rather as concepts, sorts, or schemata (kaputon logon).12 This 10 See passim and especially Chapter Three. Arguably the most sophisticated account of the constitution of complex objects can be found in Roman Ingarden’s Time and Modes of Being, trans. Helen R. Michejda (Springfield, Ill.: Charles C. Thomas, 1964). I leave an engagement with the relevance of his ontological dependence relation for the description of language to another publication. What I take from Ingarden here is the conviction that language is an ontically heteronomous process and so escapes narrowly definitional descriptions as an object within the ontic domains of either materiality or ideality. 11 Aristotle, De Interpretatione, trans. J.L Ackrill (Oxford: Oxford University Press, 2002), 1, 16a, 3-8; Aristoteles, Metaphysik, trans. Thomas Alexander Szlezák. (Berlin: Akademieverlag, 2003). Aristotle, Poetics. Trans. Joe Sachs. Newburyport: Mass.: Focus Publishing, 2006. 12 Aristotle, De Anima, trans. R.D. Hicks (Cambridge: Cambridge University Press, 1907), II, 5, 417; 418a, 3-6.

In the Beginning was Resemblance

5

suggests at least these two things: concepts are iconic and at the same time schematic. As to schematic mental images, Aristotle adds later that they “are like sensuous content but without matter”. At the same time, he separates thought from images by suggesting that though the former should not simply be equated with images, they “cannot occur without images”. 13 Yet Aristotle by no means celebrates the human capacity to imagine things. He soberly notes that “while perceptions are always veridical, imaginings are for the most part false”. 14 Nevertheless, even imagined worlds, whatever their truth status, can be usefully expressed by language, as he amply demonstrates in his Poetics. So it seems appropriate to cast our theoretical net widely. For, as Paul Grice noted, “language serves many important purposes besides those of scientific inquiry”.15 This strikes me as a wise reminder that in an age of science and technology we are likely to have developed a somewhat lopsided horizon of expectations, favouring a reductive attitude towards language both as definiendum and definiens. Positions in the history of language theorization that foreground resemblance relations can be represented by Locke, Kant, Peirce, Husserl, and writings in cognitive linguistics. Because of their emphasis on likeness in mental representations, it should not be surprising that they are committed to some sort of mentalism, known also somewhat misleadingly as “ideational” approach to meaning. On the other hand, language theories foregrounding truth and falsity, that is language viewed in terms of propositions, comprising those by Frege and the entire post-Fregean tradition up to current writings in hyper-intensional semantics, can glean support from Aristotle’s realism and his observation that “it is because the actual thing exists or does not exist that the statement is said to be true or false”.16 With respect to this opposition in the philosophical theorization of language, linguistics appears more ambivalent. While the Saussurean structuralist heritage has tried to realize its goal of a scientific description of natural language, the psychological school of Lev Vygotsky has retained its interest in the mentalist side of meaning, an emphasis which current cognitive linguistics is attempting to consolidate by proposing “cognitive models” in an attempt to corral mental iconicity and its abstractions. What links all these theoretical enterprises is the Aristotelian 13

De Anima, III, 8, 432a, 11-17. De Anima, III, 3, 428a, 19f. 15 Paul Grice, Studies in the Way of Words (Oxford: Clarendon Press, 1989), 23. 16 Aristotle, Categories, trans. E.M. Edghill (Adelaide: University of Adelaide, 2013), 4b, 8f. 14

6

Introduction

primacy of schematic resemblance of which analytical notions of the “concept” must take note as much as do phenomenological notions of “typification”. For only formal systems, such as various forms of mathematics, escape the constraint of resemblance. Without it, natural language would not be what it is, that is, natural. If in the West language has been theorized mainly along the Aristotelian principles of resemblance and proposition, this is by no means a necessary global starting point. In sharp cultural contrast, Chinese views of language for instance have been consistently anchored to this day in what one could call a “normative nominalism” and the tradition of the “rectification of names” from Confucius to Xunzi and beyond. Normative, because language in the Chinese tradition was primarily conceived as prescriptive in terms of social rules. A prince must behave like a prince, a father like a father, according to the Confucian Analects. One can call this conception of language also a “nominalism” in the sense that the terms “prince,” “father,” “wife,” “elder brother” denote social particulars rather than ideas. “Ming” in the Analects and other early writings still means “name,” “title,” and “rank” all at the same time. As Confucius is recorded as having advised, “Let the prince be a prince, the minister a minister, the father a father and the son a son”. 17This very much chimes with what is known as the source of the idea of the “rectification of names”. In answer to the question “Why should language be corrected?” the Master in the Analects says, “If language is incorrect, then what is said does not accord with what was meant; and if what was said does not concord with what was meant, what is to be done cannot be effected”.18 The main function of language in the Chinese classical tradition, then, has been what contribution it can make to the stabilization of the existing social order. Given the two very different points of departure in the Western tradition, it should not be surprising to find one recent branch, language philosophy from Frege to current writings in hyperintensional semantics, favouring formalisation as capturing best the propositional principles discoverable in or imposable on language, while the other branch, exemplified most prominently by Locke, Kant, Husserl, and much of cognitive linguistics, has remained committed to the way our shared pathemata about the world are transformed in language. Even if resemblance plays a different role in each of Locke’s successors, they all acknowledge its foundational role, 17

Confucius, The Analects, ed. Arthur Waley (Hertfordshire: Wordsworth Editions, 1996), Book 12, 11. 18 Confucius, The Analects, Book 13, 3.

In the Beginning was Resemblance

7

reflecting as it does the function of language as complement to, replacement of, and economizing hierarchization of its gestural and perceptual precursors. 19 From such an abstract perspective, two important shifts appear to characterize the evolution of language. One is the gradual semantic disembodiment of signifiers, a process of which onomatopoeia is a minimal leftover. The other shift underlying the development of language is the transformation of nonverbal concepts into more precise linguistic signifieds. So in order not to lose sight of the content of the signified, I emphasize throughout its two components, iconic mental materials and their conceptual regulation. This is why I deviate from the practice in both analytical philosophy and cognitive linguistics of conflating content and regulation under the term “concept”. As I will try to show, a consequence of this conflation is that it discourages us from specifying what the content consist in and what sort of regulatory mechanisms that content is governed by.20 Here, the regulated, motivated signified is claimed to be activated by imaginability in two ways, one by the reconstruction of what a speaker is talking about, in short, aboutness, immediately modified by the reconstruction of speech intention, implied deixis carried by voice. I conceive ‘voice’ broadly, allowing for Roland Barthes’ personal “grain of voice”, as well as for Foucault’s pragmatic notion of “enunciative modalities”. 21 Accordingly, the meaning event is regarded as result of the realization of both in the mental states of the language user. Throughout the book, natural language, as for example Chinese, French, Navajo, Turkish, or English, is understood as “a set of social instructions for imagining and acting in a world”. That imaginability in this formulation precedes “acting” is no accident, for without the projection of some relevant mental scenario (at very high speeds in habitual speech) we cannot meaningfully act on linguistic cues. The proposed primacy of the imaginability of linguistic content then applies to these two fundamental ingredients of all natural languages: (1) aboutness, 19

Not everyone agrees. David McNeill, for instance, in How Language Began: Gesture and Speech in Human Communication (Cambridge: Cambridge University Press, 2012) argues against gestural communication as a precursor of language; cf. however Adam Kendon, Gesture: Visible Action as Utterance (Cambridge: Cambridge University Press, 2004) and “Language’s Matrix,” Gesture 9, 3 (2009), 352-372. 20 The reader who requires a more detailed description of the components of the redefined linguistic signs is invited to go straight to Chapter Ten. 21 Roland Barthes, The Pleasure of the Text, trans. Richard Miller (London: Jonathan Cape, 1976), 66; Michel Foucault, The Archaeology of Knowledge (London: Tavistock, 1986), 55ff.

8

Introduction

or what a discourse, sentence, phrase, or word is about; and (2) voice, or the manner in which expressions are spoken. Aboutness, as I will argue later is conceived broadly (swan, walking), including such special applications as in realist reference (Venus, this swan) and fictional reference (electric sheep). At the same time, I will emphasize the iconic nature of aboutness as something we must be able to imagine, even though we appear to do so schematically and often at the speed of a camera shutter.22 In choosing the term voice I sum up a broad conceptual field, comprising Austin’s technical “illocution,” 23 Foucault’s pragmatic and political “enunciative modalities,” 24 as well as explicit and especially implicit deixis.25 The notion of voice, broadly conceived, is also meant to remind us that natural languages are primarily spoken, their written forms being a fairly recent invention. Terms like “language,” “linguagio”, “lingua,” and “yu yan” all remind us of the essential role of the tongue, while terms such as the German noun “Sprache” or the Swedish “språk“ still carry the idea that languages are above all spoken forms of communication. But my main reason for anticipating the fundamental distinction between aboutness and voice here is that I want to alert the reader from the outset to a principle of natural language that distances it sharply from formal languages. Put simply, formal languages have neither aboutness nor voice in any comprehensive sense. Applied formal languages have only aboutness but no voice. This is why approaches to natural language that take as their ideal the case of a formal sign system will from the start be misled in vital respects. In contrast, all natural languages are characterized by the inextricable interaction of these two components. The fact that we can, by the use of an ironic tone, reverse the propositional content of an expression into its opposite only indicates the tip of the iceberg that is the domain of enunciative modalities which is crucial to the ontological status of a natural language as a process rather than an object. In order to help clarify such issues, each chapter will focus on one or more theoretical positions on language from the perspective of the way they handle the imaginability of aboutness and voice.

22

Cf. “Shutter-Speed Meaning, Normativity, and Wittgenstein’s Abrichtung,” Philosophical and Linguistic Investigations (2014; in press). 23 John Austin, How to Do Things with Words (Oxford: Oxford University Press, 1962). 24 Michel Foucault, Archaeology of Knowledge, 55ff. 25 Horst Ruthrof, The Body in Language (London: Cassell, 2000), 48-53.

In the Beginning was Resemblance

9

Chapter One “Resemblance in Language” revisits Locke’s philosophy of language, asking the question of what sort of arguments would be required to not merely defend his linguistic meaning as a mental event but to claim that Locke, in spite of the shortcomings of his semantics, was basically on the right path towards a satisfactory description of natural language. In foregrounding mentally realized resemblance relations, or Vorstellung, as the mainstay of Locke’s semantics, the chapter argues that Locke’s modest mentalism is closer to the mark than our dominant externalist solutions in their sentential, propositional, and formal varieties. Next, Locke’s notion of immediacy is analysed in terms of speed, exclusivity, absence of mediation, and as a case of strict rule following. Locke’s theorization of abstraction is then viewed in terms of Peirce’s hypoiconic generalization, before I address Locke’s rejection of radical externalism, his arguments against meaning identity, and his four constraints on meaning: human biology, world, language as system, and the speech community. The chapter concludes by dissolving Locke’s seeming paradox, the contradiction between private meaning and public discourse, by the notion of linguistic meaning as indirectly public. Chapter Two “Iconicity in Kant and Peirce” explores the ways the two philosophers conceive of resemblance relations. Although Kant did not leave us a theory of language, I suggest that we can distil from his Critiques a number of concepts that help us shore up an intersubjective mentalism conducive to arguing for a semantics of imaginability. Foremost amongst those concepts are his notions of schemata and the monogram, the role he grants Vorstellung, and the power of the productive imagination (Einbildungskraft) in cognition. The chapter further claims that Kant’s emphasis on the mechanism of schematization allows us to lift Locke’s empiricist notion of “idea” to a level of abstraction that is conducive to non-subjectivist arguments about mental resemblance relations and so can play an important role in a semantics of imaginability. Kant is also shown to be useful for such an enterprise in terms of his perspectivism, in which language could be said to occupy a prominent position. The chapter then addresses Peirce’s insistence on iconicity as essential for the comprehension of non-formal sign systems, such as a natural language. I want to show that the Aristotelian and Kantian schematic resemblance relations that I argue are vital for linguistic meaning, receive in Peirce a systematic placement in his theory of semiotics on which we can draw substantially for a semantics of imaginability.

10

Introduction

Chapter Three “Meaning and Imaginability in Husserl” takes its cue from the intriguing evolution of Husserl’s phenomenology from idealist semantic convictions towards a view of linguistic meaning in tune with his later focus on the life-word. The chapter argues for a way of bringing Husserl’s semantics in Logical Investigations up to date by drawing on a variety of critical tools gleaned from his later writings. My argument proceeds in two steps. First I offer a summary of the main ingredients of Husserl’s theorization of natural language, with an emphasis on his description of linguistic meaning as a Platonic ideal species. The chapter then gathers a number of concepts from Husserl’s later works up to Experience and Judgment for the kind of repair work that could make his semantics once more a competitive candidate in the arena of natural language semantics. For this to be possible, his early commitment to meaning ideality, which proved incompatible with Husserl’s later writings on the Lebenswelt, has to be replaced by iconic resemblance, which is precisely the ingredient we find in such later concepts as appresentation, quasi-perceiving, non-essential typifications, and semblance acts. With this re-emphasis the chapter reformulates Husserl’s approach to linguistic meaning in terms of a semantics of imaginability. Chapter Four “The Propositional Route: From Frege to Hyperintensional Semantics” offers a non-technical introduction to the most powerful tradition in the philosophy of language, its analytical branch.26 Mindful of our overall theme of imaginability, the chapter addresses first Frege’s influential and yet untheorized conflation of two kinds of sense, the formal sense we find in geometry and arithmetic and the non-formal sense of meaning in natural language. Mindful of not wishing to torture the reader with details of formal semantics, the chapter traces the presuppositions of the post-Fregean tradition up to some recent publications in hyperintensional semantics. In so doing, the chapter argues that as a result of formalisation 26

There can be little doubt that the dominant discourse about natural language has emerged from analytical philosophy in the wake of Gottlob Frege’s seminal paper “On Sense and Reference” (“Ueber Sinn und Bedeutung”) of 1892. Nor is there any doubt as to the pre-eminence of his influence in a recent wave of collections of papers, such as Ernie Lepore and Barry C. Smith (eds.), The Oxford Handbook of Philosophy of Language (Oxford: Clarendon Press, 2006); Michael Devitt and Richard Hanley (eds.), The Blackwell Guide to the Philosophy of Language (Oxford: Basil Blackwell, 2006); Stephen Davis and Brendan S. Gillon (eds.), Semantics: A Reader (Oxford: Oxford University Press, 2004); and Peter V. Lamarque (ed.), Concise Encyclopedia of Philosophy of Language (Oxford: Pergamon, 1997).

In the Beginning was Resemblance

11

the object of inquiry, natural language, has been fundamentally transformed from a sign system in which iconicity is vital to a sign system reduced to syntax. This transformation is argued to be visible in the reduction of aboutness to empty place holders and the elimination of voice. Chapter Five “A Critique of Meaning as Use” turns to the later Wittgenstein and his highly influential definition of linguistic meaning as use. 27 In asking the impertinent question of what “use” consists in, the chapter attempts to show that in its Wittgensteinian form “use” can only partially account for what actually occurs in the process of meaning making. In particular, the chapter focuses on Wittgenstein’s elimination of Vorstellung (mental projection of relevant scenarios) from meaning by relegating it to the side show of a mere playing of tunes on the keys of our Vorstellungsklavier, our piano of the imagination. Instead, the chapter argues that Wittgenstein’s radical semantic externalism only captures a relatively small portion of the total of linguistic meaning events. After all, the argument goes, the bulk of meaning as “use” is not open to public inspection. The chapter concludes by suggesting that Wittgenstein’s notion of meaning as “use” can be improved upon by reintroducing the banished Vorstellung as an essential ingredient of the meaning event. Chapter Six “The Linguistic Route: Saussure and Vygotsky” opens with confronting two central tenets in the General Course in Linguistics of 1916, Saussure’s conviction that without language we would live in a “nebulous” world and his radical arbitrariness thesis of the linguistic sign. Neither hominids nor tigers can be conceived to live in a woolly world. Rather, it makes sense to regard their worlds as clearly distinct and no doubt simpler than ours, but not as less precise. Otherwise, survival would be unlikely. As to the alleged arbitrariness of the linguistic sign, it is crucial for a semantics of imaginability, as it is for all cognitive approaches, to insist on the theorization of the signified as motivated, leaving the signifier much as it is defined by the father of modern linguistics. As I put it elsewhere, “at the level of the signified we are iconic beings.”28 The chapter also points out that the arbitrariness thesis as its stands rests on a quite straightforward pars pro toto form of reasoning. What remains unresolved in the chapter are the consequences of these criticisms, which will be resumed in detail in Chapter Ten. The second part of Chapter Six is dedicated to an analysis of Lev Vygotsky’s 27

Ludwig Wittgenstein, Philosophical Investigations, trans. G.E.M. Anscombe, P.M.S. Hacker and Joachim Schulte (Oxford: Blackwell, 2009). 28 Horst Ruthrof, The Body in Language (London: Cassell, 2000), 152.

12

Introduction

arguments on the transformation of perceptual thinking into linguistic concepts by way of pedagogy. This is important in light of the aim of the book to argue in favour of meanings not only as schematically iconic mental entities but also as indirectly public and so sharable to a high degree by the members of a speech community. In drawing on Vygotsky’s findings I extrapolate that language pedagogy must have two sides, one that secures knowledge of public signifiers and their standard syntactic constellations, the other, which aligns these with indirectly public signifieds as conceptualized mental iconic materials. Chapter Seven “Cognitive Linguistics and Conceptual Blending” first identifies a few important differences between some of the axioms we find in cognitive linguistics and the position advocated here, in spite of many parallels. A fundamental difference, for example, is shown to be the distinction drawn by Vyvyan Evans between “lexical concepts” and “cognitive models” and my return to the basics of Saussure’s arbitrary signifiers which, as no more than arbitrarily sounds, do not “contain” any meaning whatsoever, but are meaningful only to speakers of the language as a result of having been trained to associate those sounds with the mental iconic scenarios of motivated signifieds. In contrast, lexical concepts are meaning entities independent of speakers to be activated by individuals via conceptual models. The chapter selects a number of positions in cognitive linguistics, rehearsing their presuppositions from the position of the theme of imaginability. By way of conclusion, I look at the influential theory of “conceptual blending” as proposed by Gilles Fauconnier and Mark Turner (2002), asking how its principles relate to graded iconic schematization. In particular, the chapter asks how conceptual blending deals with non-visual, non-spatial mental entities. Chapter Eight “Brain, Consciousness, and the Evolution of Language” opens with the sobering finding that current research results in brain language research are anything but grand and that we must stay content for the time being with a paucity of discoveries in support of any kind of semantics. The chapter takes issue with the common confusion of brain and person by insisting that a deep chasm exists between the two. After all, a brain cannot be “embarrassed,” only a person can be. After discarding the Chomskyan claim of a “language organ” in the brain, the chapter embraces the idea of the plasticity of the brain and the speculative likelihood that all perceptual and sensory-motor neural pathways are involved in the processing of natural language. This is argued to support the idea of a fundamental link between semantic performance and

In the Beginning was Resemblance

13

iconicity which can draw not only on brain cells needed for perception but also, and more importantly, on the massive presence in the brain of neurons not directly involved with perceptual activities but rather with their internal monitoring. The extraordinary asymmetrical ratio between perceptual input and neurally driven internal monitoring neurons, I suggest, strongly supports the emphasis on the primacy of imaginability. The chapter then turns to a leading researcher on consciousness, David Chalmers, whose impressive work on “conceivability” I argue underestimates the basis of iconic awareness on which propositional conceivability supervenes.29 The chapter concludes with a discussion of some findings in research on language evolution with a view towards buttressing some of the evidence in brain and consciousness studies in support of the role of imaginability in natural language. Chapter Nine “Imaginability and Pragmatics: From Grice to Habermas” addresses standard pragmatics in terms of imaginability. Gricean and other “presuppositions”, like truth-conditions, are shown to be parasitic on imaginary scenarios, that is, clusters of Vorstellungen. The chapter argues that pragmatics should incorporate the bridging process of imaginability, understood as “mental iconic transformations of perception”. In such a broadened conception of pragmatics, imaginability must be acknowledged as a formative component not only of logical presuppositions, but of all meaning events in natural language. The chapter includes a critical reassessment of the kind of pragmatics we find in Deleuze and Guattari, especially in A Thousand Plateaus, and in Foucault’s Archaeology of Knowledge. The chapter concludes with what I regard as the most inclusive pragmatic vista, the notion of “communicative action” argued by the social philosopher Jürgen Habermas. Chapter Ten “Language and Imaginability” is a comprehensive statement of an iconically driven theorization of meaning in natural language. At its centre stands the notion of imaginability, understood as mental, iconic transformation of perception. As such, imaginability is what determines the way word sounds are given iconic content in the mental acts of native 29 David Chalmers, The Conscious Mind: In Search of a Fundamental Theory (Oxford: Oxford University Press, 1996); cf. also his The Character of Consciousness (Oxford: Oxford University Press, 2010) and “Does Conceivability Entail Possibility?” in Conceivability and Possibility, ed. T. Gendler and J. Hawthorne (Oxford: Oxford University Press, 2002), 145-200; see Chapter Eight below.

14

Introduction

speakers which are regarded as indirectly public as a result of pedagogy. In pulling together the observations made in the preceding chapters, Chapter Ten presents a series of redefinitions of standard notions in our dominant discourses on language in philosophy, semiotics, and linguistics. Foremost amongst these redefinitions are the Saussurean linguistic sign, with an emphasis on the motivated signified. Reference is revised as a special case of aboutness, while deixis is shown to require a stronger emphasis on voice, or implicit deixis. The chapter redefines the concept as regulatory component of the signified, governing its mental iconic materials by way of directionality, quality, quantity, and degree of schematization. Other notions to be revisited are meaning and metaphor, referential and deictic background, as well as the constraints imposed on natural language by sufficient semiosis, which replaces meaning identity and allows for a more realistic description of language than does truthconditional semantics. The Conclusion “Language as Heterosemiotic, Intentional Construct” begins with the summary observation that in the event of meaning, language, in combining the symbolicity of arbitrary signifiers with the iconicity of motivated signifieds, is fundamentally heterosemiotic. This contrasts sharply with the widely held belief that language is a symbolic system. Language is argued to be made up of two interacting systems, the public system of observable linguistic expressions, including their specific syntactic sequences and standardized phonetics, and the indirectly public system of motivated signifieds consisting of iconic mental materials within the conceptual boundaries of directionality, quality, quantity, and degree of schematization. This picture of natural language, the Conclusion claims, is compatible with phenomenological ontology, according to which traditional ontologies cannot account for ontically compound objects because they confine themselves to the domains of materiality and ideality. I end the book by suggesting the possibility of extending the phenomenological idea of ontic heteronomy to include natural language.

CHAPTER ONE RESEMBLANCE IN LANGUAGE: LOCKE

Introduction John Locke’s semantics of natural language has not been well received in the literature mainly because of its empiricist psychologism, its private notion of linguistic meaning, and the deficiencies of its theory of abstraction. However, re-reading the Essay Concerning Human Understanding today, especially in the context of cognitive linguistics, tells us that it is still making a substantive contribution to a theory of natural language semantics. 1 Locke has a good deal to say about the speech community, the distinction between signifiers and signifieds, the arbitrariness of the signifier, a shared perceptual world, and speakers with a common biological make-up suitable for speech production, mental acts for meaning-conferment and meaning-fulfilment, and the fundamental communicative purpose of language. So why should Locke’s semantics be regarded as a grand failure? There is broad agreement that his “private” notion of meaning prevents him forging the vital connection between speaker, public language, and speech community. One solution would be to argue language strictly along externalist lines, as public performance. This is what Wittgenstein’s notion of meaning as use is trying to accomplish. Another solution is to revise Locke’s concept of semantic privacy, a path I will explore below. I want to show how the connection between public discourse and intersubjectively shared mental meanings can be argued in order to resolve Locke’s “central paradox”. This I feel is justified in light of the Essay Concerning Human Understanding still 1

John Locke, An Essay Concerning Human Understanding, ed. John W. Yolton (London: Dent). For balanced summaries, see Norm Kretzmann, “The Main Thesis of Locke’s Semantic Theory” in Locke on Human Understanding: Selected Essays, ed. Ian Tipton (Oxford: Oxford University Press, 1977), 123-140; or Paul Guyer, “Locke’s Philosophy of Language.” In The Cambridge Companion to Locke, ed. Vere Chappell (Cambridge: Cambridge University Press, 1994), 115-145.

16

Chapter One

posing a serious challenge to many of our post-Fregean theories of linguistic meaning in their formalist, externalist, and syntactic guises. It does so by addressing biological and social constraints on ideas as meaning, by its insistence on the impossibility of identity of meaning in natural language, the rejection of radical semantic externalism, and as a firm alternative to the Fregean treatment of natural language meaning as formal sense. To be sure, the kind of semantics envisaged by Locke stands in stark contrast to a plethora of recent publications in the philosophy of language. 2 But Locke’s Essay has to be appreciated for its insistence on natural language carrying Aristotelian resemblance relations. 3 Locke differs radically from our dominant theories of language in that he resumes the Aristotelian theme of pathemata and homoiomata by which we are able to signify aspects of the world in their absence not only in their radically formalised transformations, but as iconic schematizations.4 If we foreground resemblance relations as a vital ingredient of natural language, a point resurrected in cognitive linguistics, Locke’s Essay once again looks like a serious candidate in the arena of natural language semantics.5 2

Cf. Alex Barber and Robert J. Stainton (eds.), Concise Encyclopedia of Philosophy of Language and Linguistics. Oxford: Elsevier, 2010; Keith Allan (ed.), Concise Encyclopedia of Semantics. Oxford: Elsevier, 2009; Ernie LePore and Barry C. Smith (eds.), The Oxford Handbook of Philosophy of Philosophy of Language. Oxford: Clarendon Press, 2006; Michael Devitt and Richard Hanley (eds.), The Blackwell Guide to the Philosophy of Language. Oxford: Blackwell, 2006; Stephen Davis and Brendan S. Gillon (eds.), Semantics: A Reader. Oxford: Oxford University Press, 2004; and Peter V. Lamarque (ed.), Concise Encyclopedia of Philosophy of Language. Oxford: Pergamon, 1997. 3 Benjamin Hill, “”Resemblance” and Locke’s Primary-Secondary Quality Distinction,” Locke Studies 4 (2004), 89-122; Michael Jacovides, “Locke’s Resemblance Thesis” Philosophical Review 108, 4 (1999), 461-496; and Roland Hall, “Locke and the Senses,” Locke Newsletter 26, (1995), 13-27. Hall asks the question: “Where did Locke find his basic materials?” before tracing a number of connections between the Essay and Aristotle, especially in De Anima III, 7, a passage that suggests a tight link between consciousness and imaginative resemblance. This supports my emphasis throughout on homoiomata. 4 A compatible pathway towards the theorization of schematic resemblance relations can be found in the work of Eleanor Rosch, as or example in her “Principles of Categorization” in Concepts: Core Readings, ed. Eric Margolis and Stephen Laurence (Cambridge, Mass.: MIT Press, 1999), 189-206. 5 Victor Gallese and George Lakoff, “The Brain’s Concepts: The Role of the Sensory-Motor System in Reason and Language,” Cognitive Neuropsychology 22, (2005), 455-479; Mark Johnson and George Lakoff, Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books, 1999. What must remain doubtful is the notion of the “neural concept,” which is

Resemblance in Language: Locke

17

Whatever flaws have been discovered in Locke’s semantic project, its strength lies precisely in his attempt to provide an explanation for that fundamental relation. However, instead of taking the avenues opened up by recent work in cognitive linguistics, I will approach Locke’s central paradox from a related but somewhat different angle: the perspective of imaginability.6 Crudely, we can sum up such an approach by saying that “If I can imagine what you are saying, and the manner in which I say it, there is meaning; if I cannot, there is not”. To see the relevance of imaginability for Locke we first need to find out what it means to know a language. This is Locke’s starting point: a speech community roughly shares ideas about the world, as well as their basic relations, and has created a language as a vehicle to facilitate at least rough communication amongst its members. Speakers have inherited their ideas about the world by way of a shared biology and their transformations into verbal ideas by culture. Higher-level terms owe more to culture; lower-level terms more to biology and the natural world. Even if painted crudely, this picture points to the powerful intervention Locke’s semantics makes: unlike the “rear view mirror” approach characteristic of our dominant forms of semantics,7 Locke attempts to bridge the mysterious gap between perception and language by way of “ideas” or Vorstellung, rendered here as “mental iconic transformation of perception”. To demonstrate these claims, I want to discuss some of the fundamental issues underpinning Locke’s enterprise, (1) his limited form of mentalism; (2) the four components of immediacy; (3) arguments concerning abstraction; (4) the rejection of radical externalism; (5) semantic latitude and the non-identity of meaning; and (6) his fourfold set of semantic constraints: human biology, world, language as system, and the speech community. difficult to reconcile with both primitive concepts and linguistic concepts as social rules. There also remains the traditional worry about the theorization of the relation between brain and consciousness in the sense that while the neural basis of the mind is not in question, what remains entirely opaque is how bio-chemicalelectrical relations turn into awareness. Cf. Kant, Critique of Pure Reason, trans. Norman Kempt Smith (New York: St. Martin’s Press, 1965), CPR A141/B181 and Wittgenstein’s “unbridgeable gulf,” in Philosophical Investigations, trans. G.E.M. Anscombe, P.M.S. Hacker and Joachim Schulte. Oxford: Blackwell, 2009), § 412; See also David Chalmers’ arguments in Chapter Eight. 6 Horst Ruthrof, “Semantics of Imaginability – Vorstellungssemantik: 13 Theses,” Review of Contemporary Philosophy 10 (2011), 165-183. 7 Cf. the emphasis on logical relations in the ontogenesis of language, as for example in Stephen Crain, The Emergence of Meaning (Cambridge: Cambridge University Press, 2012).

18

Chapter One

Idea as Vorstellung and aboutness Mentalism takes many forms, some of which should be laid to rest, while others remain competitive candidates in the quest for a satisfactory explanation of how language works. The charge that Locke’s semantics is imagist belongs to the group to be discarded, as has been argued by David Soles, on the grounds that to reduce notions such as “succession” and “duration” to images is untenable.8 (Essay, Book II, xiv, 3; cf. II, xvii, 7f) One could add olfactory, gustatory, thermal, gravitational and other nonverbal readings of the world for which the notion of “image” is inappropriate. Following Locke, we should also resist the temptation to think that any mental abstraction would primarily have to be verbal. Some other form of schematization may have to be stipulated. A convenient entry into Locke’s mentalism is the question of the semantic scope of his notion of “idea”. Locke’s “idea” is conceived as a mental representation mediating between perceptual world and language. It is “something in the mind between the thing that exists, and the name that is given to it”. (Essay, Book II, xxxii, 8) This “something in the mind” is “whatsoever is the object of the understanding when a man thinks”. (I, i, 8) As such, the term “idea” also covers resemblance in and as Vorstellung, “a memory of an actual perception being “made an actual perception again”. (I, iv, 20) This suggests a realist conception of imaginability. The affinity of perception and its mental modification comes across clearly also in Locke’s assertions that “having ideas, and perception” are “the same thing” (II, I, 9) and that ideas are “nothing but bare appearances or perception in our minds”. (II, xxxii, 1) We can see why Husserl regarded Locke as an important forerunner of phenomenology in spite of his reservations. 9 If ideas function as the mental mediation between perceptual reality and language, their main task is their transmission of aboutness from world to word and vice versa. Whether we call this “intentionality” and its specific object of

8

See David Soles, “Is Locke an Imagist?” in John Locke: Critical Assessments by Leading Philosophers, Series II, ed. Peter Anstey (London: Routledge, 2006), 4781; 70. 9 Edmund Husserl, Psychological and Transcendental Phenomenology and the Confrontation with Heidegger (1927-1931), ( Dordrecht: Kluwer, 1997), 97; cf. also Edmund Husserl, Phantasy, Image Consciousness, and Memory (1898-1925), trans. John B. Brough, Collected Works, ed. Rudolf Bernet, v. XI. Dordrecht: Springer, 2005.

Resemblance in Language: Locke

19

thought the “intentional object,” as does Landesman,10 or characterize it in some other way, what is pivotal in Locke’s account is his insistence on a process from percept via concept to language (and vice versa) if aboutness is to be preserved in some way. Furthermore, whatever the degree of abstraction which aboutness undergoes in the transition from perceived world to language, at the moment of the mental event that Locke terms “idea” we have a “representation of the thing” in the mind (IV, xxi, 4) with “greater or nearer correspondence” with actual objects “than any words or sounds whatsoever”. (IV, iii, 19) Resemblance is likewise in the foreground when Locke writes, “our simple ideas are clear, when they are such as the objects themselves, from whence they were taken … in a wellordered sensation or perception”. (II, xxix, 2) Here Locke identifies what we might now call iconic resemblance relations as the point of departure for mental abstraction as a necessary step towards linguistic meaning. At a certain level of generalisation of aboutness we are in a position to conceive of “diagrams drawn on paper”11 as “copies of the ideas in the mind” which Locke thinks are “not liable to the uncertainty that words carry in their signification”. (IV, iii, 19) On the way to language, Locke distinguishes ideas that are merely “mental” from those that are “verbal”. (IV, v, 2) He also distinguishes merely “mental” propositions which permit us to decide on true and false relations from “verbal propositions” formulated in affirmative or negative sentences. (IV, v, 5) When Locke speaks of ideas as “thoughts,” he is drawing our attention to the fact that they remain “hidden from others”. (III, ii, 1) Whether Locke’s “postulation of unobservables” is a necessary consequence of his empiricism remains a moot point.12 What is pivotal to his semantics is that an important part of the meaning event is not open to public inspection. Here Locke’s semantics defies any simple form of externalism. Instead, the accent is on the performance of “meaning conferring” and “meaning fulfilling” acts, as Edmund Husserl is to describe them much later in his Logical Investigations.13 The threatening solipsism in Locke’s insistence 10

Charles Landesman, 222f; cf. also Michael Ayers, “Are Locke’s “Ideas” Images, Intentional Objects or Natural Signs?” Locke Newsletter 17 (1986), 3-36. 11 Cf. the notion of “schema” in Kant, CPR A141 and Charles Sanders Peirce, 1984: Writings of Charles S. Peirce: A Chronological Edition, vol. 2. 1867-71. Peirce Edition Project (Bloomington: Indiana University Press, 1984), EP 2.273. 12 As discussed by David Soles in “Locke’s Empiricism and the Postulation of Unobservables,” Journal of the History of Philosophy, 23 (1985), 339-369. 13 Edmund Husserl, Logical Investigations, trans. J.N. Findlay (London: Routledge and Kegan Paul, 1973), §7f.

20

Chapter One

on the concealment of the linguistic meaning event appears to be reinforced by his characterization of “thought” as fundamentally reflexive: “thinking consists in being conscious that one thinks,” (II, I, 19). What considerably weakens the charge of solipsism, however, is Locke’s introduction of a number of constraints, as we shall see. One such constraint is his concept of “immediacy”. The split-second event of comprehension, which I have discussed under the term “linguistic linkage compulsion,” defines the fully trained speaker of a language.14

Immediacy A forceful argument in favour of Locke’s notion of mental immediacy is that our other semantic explanations (sentential, propositional, truthconditional, and formal) are all parasitic on our ability to imagine what we are talking about and comprehend when we listen to the words of others. When we imagine what others are saying, we first appear to construct aboutness, qualified by our construal of voice, or modality in a broad sense, the manner of the utterance, before we are in a position to play such language games as paraphrasing, forming propositions, entertaining truth checks, or translating natural language into formal strings. Locke’s immediacy of the meaning event is the baseline on which any languagelike operation can be performed: without the performance of imaginable scenarios in the minds of the speakers of a natural language there would be no such things as meanings at all. Where Locke speaks of “words in their primary or immediate signification” standing for the ideas in the mind, immediacy has been addressed in the literature as a feature of his mentalism,15 a characteristic of linguistic signification,16 and in terms of propositions. 17 Yet none of these objections account for the fact that immediacy is so much in the foreground of the Essay, and rightly so. What seems understated in the literature is that Locke was quite aware of the astonishing speed, force and exclusive compulsion that characterize the combination of public language and mental event in response to the sound of words. In stressing the importance of immediacy in linguistic 14

Horst Ruthrof, “Semantics of Imaginability,” 168. Hannah Dawson, “Locke on Private Language” in John Locke: Critical Assessments by Leading Philosophers, Series II, vol 2, ed. Peter Anstey (London: Routledge, 2006), 412-439; 413. 16 Landesman, 229. 17 David E. Soles, “Locke on Ideas, Words, and Knowledge,” in Locke: Epistemology and Metaphysics, ed. Udo Thiel (Dartmouth: Ashgate, 2002), 411433; 433. 15

Resemblance in Language: Locke

21

communication, Locke puts his finger on a fundamental constraint, the “linguistic linkage compulsion”. Try as we may, we cannot think the meaning of “toe” at the very split second of the comprehension of the signifier “democracy”. Viewed from the angle of the speech community, immediacy cannot be adduced to undermine Locke’s mentalism. Quite the contrary, grounded as it is externally, it makes his mentalism not only respectable but unavoidable. For what is public are signifiers, not signifieds. Signifieds can be established by ostension, that is, an association of word sound and perceptual grasp, but in the bulk of language use no such physical tethering is taking place: meanings have already been internalised. And when they are, they occur at high speed, at least in habitual speech. This is not to say, of course, that internalised meanings must be private; they are more likely indirectly public. If this point can be argued persuasively, then instead of regarding immediacy as part of Locke’s reduction of language to psychology, it turns out to be as robust a component of meaning as it is indispensable for any natural language semantics. Indispensable it may be, but the robustness of the immediacy of a speaker’s Vorstellung in response to the sounds of language is not of the kind argued for instance by Kretzmann who dissolves the difference between signifier and signified by declaring that “every word is an idea”.18 “Only if we know the language,” we should add, which reveals the begging of the question that is involved here. Kretzmann’s move of collapsing the signified into the signifier works for formal placeholders and their syntactic relations; it does not work for “justice,” “feng shui,” “jouissance” or “Zweckentfremdung”. While Kretzmann’s proposal would neatly solve the problem of the holy grail of the public nature of linguistic meaning by identifying signifiers with signifieds, natural language is more complicated. It is not a homosemiotic sign system; language is fundamentally heterosemiotic. The aboutness which the speaker recognizes in language is not a property of its words but a compulsory association established and regulated by the speech community. The robustness of Locke’s immediacy is the result of a binding social agreement for a certain time. Locke rejects Kretzmann’s solution by insisting that the “constant connexion between the sound and the idea” is firmly embedded in our memories “by familiar use from our cradles”. (III, ii, 7) Yet this connection is not natural but conventional as a consequence of a “voluntary imposition”. (III, x, 5) Words then cannot be ideas by themselves, nor can they be the result of 18

Kretzmann, 135.

22

Chapter One

“any natural signification,” nor indeed should Locke’s “voluntary imposition” be read as an individual act. Rather, it is the speech community that has imposed the “constant connexion,” an imposition that produces a powerful sense of immediacy when linguistic meaning occurs. Once we move to a more general level of analysis of what occurs at the moment of meaning under the “linguistic linkage compulsion,” we note that there are four aspects to consider. One is its temporal aspect captured in the notion of immediacy. Speed is the feature of the habitual meaning event characterized in Wittgenstein by the phrase “in a flash”.19 Another aspect of immediacy is what Husserl was to call “exclusive directionality”. 20 In standard meaning events aboutness is close to automatic. We do not have to ponder where a set of signifiers is pointing us to; it directs us immediately in a certain direction to the exclusion of any other. This we could call the spatial aspect of immediacy. A third aspect in Locke’s description is the absence of mediation. Habitual meaning events are immediately grasped by speakers of a language without the mediation of other Vorstellungen. Fourth, meanings are immediate because they are performed by the speakers of a language as a matter of strictly following the instructions of the speech community. In this last sense, immediacy is the consequence of comprehension as the performance of a rule.21 This turns Locke’s allegedly private meaning event into a social necessity.

Forms of abstraction An important stay in the rigging of Locke’s semantics is the way he argues the process of abstraction from particular to general ideas and, concomitantly, from particular to general terms.22 (III, iii) As an empiricist nominalist, Locke concedes that certain ideas, that is, basic ones, can be produced by the “quality of the object”. (II, iix, 8) Yet “universality belongs not to things themselves” (III, iii, 11) and so general ideas as “fictions and contrivances of the mind” (IV, ii, 9) depend on our capacity “of signifying or representing many particulars”. We produce general ideas in order to cater for sorts as “representatives of many particular things”. (III, iii, 11) Locke is aware of the need for some kind of 19

Ludwig Wittgenstein, PI §§ 139, 191, 197, 318, 319. Edmund Husserl, Logical Investigations I §10. 21 Ludwig Wittgenstein, PI §§185-242. 22 Cf. Jonathan Walmsley criticisms of Locke’s proposal in “Locke, Ideas and Abstraction: A Reply to Yasuhiko Tomida,” Locke Studies 7 (2007), 173-205; and “The Development of Lockean Abstraction,” in John Locke, ed. Peter Anstey (London: Routledge, 2006), 111-135. 20

Resemblance in Language: Locke

23

theorisation of abstraction. To “heap up names of particular things,” he says, would fail communication with others. (III, iii, 3) So, generality is embraced by Locke as a mainstay of language as the result of a process of abstraction. Starting with “particulars,” we focus on similarities amongst them and by reduction into “sorts” arrive at “general names”. (III, iii, 4) “General words” are produced by abstraction from “the circumstances of time and place” and other features determining “particular existence”. (III, iii, 6) By climbing the ladder of abstraction towards more and more “comprehensive extension” of terms, we form classes of classes, as well as “genus and differentiae”. (III, iii, 10) Locke also appears to anticipate Husserl’s “essence” as unavoidable “spontaneous ideation”.23 His “abstract idea” is “something in the mind between the thing that exists and the name that is given to it”. (II, xxxii, 8) This means that even the most concrete “idea” is always already abstract and can be abstracted further to sorts, species, and higher level classes. The Essay also treats abstraction in four different forms as “verbal or sentential generalisation,” “propositional abstraction,” “formal abstraction,” and schematized abstraction, or “hypoiconic” abstraction, to use the Peircean term.24 It is this last form of abstraction that is the most important of the four in Locke’s Essay. An important difference between the first three kinds and hypoiconic abstraction is that the latter is the precondition of the former. This is because ideas about the world and ourselves, that is, Vorstellung rather than perception, are the base line of all thinking. Even if Locke is unable to resolve the problem of treating “mental and verbal propositions separately”, (IV, v, 4) he is adamant that all forms of abstraction take their starting point from imaginability before reducing resemblance relations in different ways and to differing degrees. The four species of abstraction can now be employed to look after four different generalized notions: kinds of terms and sentences, varieties of propositions, kinds of formalisation, and sorts of nonverbal phenomena abstracted to hypoiconic realizations, that is, non-propositional, schematized states of affairs. Although Locke emphasizes that these four generalities are being played out in our minds as ideas, it would be a misrepresentation of his semantics to say that he envisaged them as divorced from their necessary social dimension. In fact, it is the social in the form of the cultural community which Locke declares the origin of our ability to perform ideas mentally in relation to language the way we do. This is why he calls language “the great instrument and 23

Edmund Husserl, Ideas, §§22f. Charles Sanders Peirce, Writings of Charles S. Peirce: A Chronological Edition, vol. 2, 1867-71 (Bloomington: Indiana University Press, 1984), EP 2.273. 24

24

Chapter One

common tie of society”. (III, I, 1) What is undeniable is the lesser charge that Locke failed to make cogent this all-important connection between imaginable abstraction and the speech community. Yet it is Locke’s fourth form of generalisation, non-verbal mental schematization, which is the most controversial because its performance cannot be publicly shown. After all, his “abstraction,” is a mental process “whereby ideas taken from particular beings become general representatives of all of the same kind”. (II, xi, 9) Unlike many of our contemporary philosophers of language, Locke considers this mental process as non-linguistic. What Locke did not pursue in the Essay is the question of gradation or degrees of schematization of resemblance relations within the process of abstraction. He likewise spent little effort in arguing the salience of such relations, in the sense of salient similarity and salient difference, except as retention of what is “common”. Nevertheless, Locke’s four avenues of schematization can be harnessed to rehabilitate the fundamental contribution which his semantics has made to the possibility of an iconically based theory of meaning.

Eschewing externalism By semantic externalism we mean a view of linguistic meaning as open to “public inspection”.25 Externalist views of language are useful as long as they restrict themselves to the observables of language without claiming that the public face of language is all there is to it. 26 Signifiers, their standardized phonetics and syntactic combinations can be described in this way. But when it comes to meaning and everything in language that depends on it, externalism as defined proves wanting. This is why Locke avoids externalism as the precondition of meaning in three forms, in its syntactic, formal and behaviourist guises. As to syntactic externalism, Locke draws a firm demarcation between meaning and definition, the latter being “nothing else, but the showing the meaning of one word by several other not synonymous terms”. (III, iv, 6) In other words, definitions are syntactic substitutions rather than meanings. 27 Verbal definitions, for Locke, are “only the explaining of one word, by several others, so that the meaning, or ideas it stands for, may be certainly known”. 25

William Alston, Philosophy of Language (Englewood Cliffs, N.J.: Prentice-Hall, 1964), 25f. 26 Cf. David Soles on “unobservables” in “Locke’s Empiricism and the Postulation of Unobservables,” Journal of the History of Philosophy, 23 (1985), 339-369. 27 Cf. Quine, Willard van Orman, Pursuit of Truth (Cambridge, Mass.: Harvard University Press, 1993), 56ff.

Resemblance in Language: Locke

25

(III, iii, 10) Importantly, meanings are not to be found in words and sentences but in the way they are realized by speakers who share the “voluntary imposition” whereby “a word is made arbitrarily the mark” of a specific “idea”. (III, ii, 1) If this is so, then semantic externalism fails and we need an explanation that relieves the notion of “idea” of its privacy, a task to be taken up below. Second, when we are engaging in formal operations, our mental acts are strictly defined by syntax rather than meaning. In the formal domain semantics reduces to syntax. In the absence of access to mathematical logic, Locke describes the principle of such operations in terms of “trifling propositions”. (IV, viii) Since here syntax is all there is and since it can be demonstrated publicly, recourse to the ideas we entertain in their performance is redundant. No such redundancy, however, can be transferred to natural language as a whole. Third, the behaviourist path even in its most respectable form of meaning scepticism, Wittgenstein’s “meaning as use,” cannot be taken by Locke because of his insistence on mental acts playing a central role in linguistic meaning. Locke rejects externalism as a foundation of meaning, but at the expense of having to defend semantic privacy, a task for which he lacked the kind of equipment we now have at our disposal. Today, we can argue that my meanings are private in the trivial sense that they cannot be inspected by others; and yet that they are public in the important sense that they are shared as a result of training by the speech community. As such, meanings can be said to be indirectly public.

Semantic latitude and the non-identity of meaning “The most burning argument of Book Three, which quivers throughout the Essay as a whole,” writes Hannah Dawson, “is Locke’s deep anxiety about the actual miscommunication which bedevils our daily intercourse”.28 And it is the daily fact of miscommunication that confirms for Locke that linguistic meaning cannot be identical in a way externalist explanations would have it. (III, ii, 2; also III, ii, 4; III, ii, 5) Instead, he based his entire theory of language on the principle of semantic latitude. 29 Instead of regarding this as a fundamental flaw, as for example do Ian Hacking and

28

Hannah Dawson (2006), 413. Rather than “semantic promiscuity,” a phrase employed to effect by Hannah Dawson in her paper “A Ridiculous Plan: Locke and the Universal Language Movement,” Locke Studies 7 (2007), 137-158; 157; where she argues that Locke’s theory of language seriously undermines the assumptions underlying the universal language programme. 29

26

Chapter One

Roy Harris, 30 the flaw may indeed lie in the import of the notion of identity into our inquiry. Certainly, if we embrace an intersubjective mentalism, as Locke is attempting, the assertion that speakers must hold identical meanings for communication to be possible is unwarranted. Only a tiny group of semiotic systems, that is, formal systems obeys the rule of identity. In contrast, communication by gestures does not follow the expectation of identity conditions. Gestures are more or less the same: approximation suffices. Much the same can be observed about natural language. The second shortcoming is a tendency to be dazzled by the ideal of the “crystalline purity” of logic introduced as a condition of our investigation.31 Locke avoids both pitfalls and has been blamed for doing so. Though “common use” is a “very uncertain rule” and a “variable standard” for natural language, it nevertheless guarantees sameness of meaning “as near as may be”. There is, amongst the speakers of a language, “an agreement in the signification of common words, within some tolerable latitude”. (III, xi, 25) Semantic identity is not required. Meaning approximation is the necessary and sufficient condition of linguistic communication. As Locke was well aware, “languages constantly change”. (II, xxii, 7) Given these provisos, Locke can be confident in asserting that public discourse, understood as “common use,” “regulates the meaning of words pretty well for common conversation”. (III, ix, 8) Fundamentally, language is governed by use rather than definition, (III, xi, 25) as long, that is, as “use” does not exclude imaginability. One reason for Locke’s non-identity of meaning is that though signifiers are public, the ideas or Vorstellungen that make them semantic are not, at least not in any sense of identity. (III, ii, 2) Another reason for Locke’s rejection of meaning identity is the socially shared assumption that we mean roughly the same when we use the same terms on the grounds of the “common acceptation of language”. (III, ii, 4) This argument is reinforced a little later in the Essay when its author speaks of the “connexion between certain sounds and the ideas they stand for” as a result of “constant use,” a relation however that we cannot assume to produce identity in any strict sense. (III, ii, 6) Yet another reason is that when we use language as a reflection of perceptual reality, even if only indirectly so via Vorstellung, as members of a speech community we do so in a similar way. (III, ii, 5) As Locke seems to be saying, in natural language the communicative rule 30

Ian Hacking, Why Language Matters to Philosophy (Cambridge: Cambridge University Press, 1975), 44; Roy Harris, The Language-Myth (London: Duckworth, 1981), 88. 31 Ludwig Wittgenstein, PI §108.

Resemblance in Language: Locke

27

is semantic latitude within the constraints of what I have called sufficient semiosis.32

Locke’s semantic constraints In Locke’s Essay we find four fundamental constraints on linguistic meaning: human biology, perceptual world, language as system, and the speech community. Because humans share the same sense organs and most likely process sense information in the same way, Locke is “very apt to think that the sensible ideas produced by an object in different men’s minds are most commonly very near and indiscernibly alike”. (II, xxxii, 15) This is a resumption of Aristotle’s pathemata and homoiomata about the world which are more or less the same for all humans.33 Locke did not wish to “trouble” the reader with the “many reasons” he could offer for what he thought to be a justified assumption. Today, Locke would be able to shore up these observations with the help of evolutionary biology and neural science. What we share as humans are “our animal spirits, variously agitated by external objects”. (II, viii, 4) At least roughly, and allowing for misrepresentation, Locke feels entitled to assume “it evident that our simple ideas can none of them be false in respect of things existing without us”. (II, xxxii, 16) We are such, Locke seems to be saying, that we typically process the world around us in roughly the same way, as a result of our common biological make-up, which is responsible both for the mental representation of the world via simple ideas and the cognitive processes involved in what he calls complex ideas produced by a combination of simple ones. Contrary to Kant, at the level of simple ideas we are dealing with a primary passivity of the mind. We begin with “simple ideas” which “the understanding can no more refuse to have, nor alter” than “a mirror can refuse, alter, or obliterate the images or ideas, which, the objects set before it, do therein produce”. (II, I, 25) Though error is always possible, “simple ideas are least of all liable to be mistaken; because a man, by his sense and every day’s observation, may easily satisfy himself what the simple ideas are which their several names that are in common use stand for”. (II, xxxii, 9) The second major constraint relevant to Locke’s semantics is the world we bump into and realize mentally. The world, says Locke, very much 32

Horst Ruthrof, Semantics and the Body: Meaning from Frege to the Postmodern (Toronto: University of Toronto Press, 1997), 48f.; 258. 33 Aristotle, De Interpretatione 1, 16a, 3-8.

28

Chapter One

dictates to us what we experience. The “texture in the object, by a regular and constant operation” produces “the same idea of blue in us, its serves us to distinguish, by our eyes, that from any other thing”. (II, xxxii, 14) It is the “appearances” of “external objects” that produce in us “true ideas”. (II, xxxii, 14; cf. III, xi, 14) According to Locke’s causal epistemology, the role in determining what we see is very much in the hands of the world. For any doubts as to agreement about the ideas about the world amongst members of a community are easily rectified “by the objects they [ideas] are to be found in”. (II, xxxii, 9) It is primarily the “outward object” that provides the mind with the appropriate idea. (II, xxix, 4) For an empiricist nominalist like Locke, only direct experience of the world is real (IV, iii, 28) in the sense of providing us with real essences, reflected in simple ideas, whereas all complex ideas are manufactured by human cognition out of simple ideas and so can give us only nominal essences. What exists in the world, trees and rocks, determine both simple ideas and their linguistic expressions, while abstraction such as species and higher level classifications receive their nominal essence from language.34 In the first case, we are dealing with particulars such that “experience here must teach me, what reason cannot,” (IV, xii, 9) including our “intuitive knowledge of our own existence”; (IV, ix, 3) in the latter case, we are dealing with an “abstract idea” tied to a “name” signifying not” any particular being” but rather “any sort of things”. (III, iii, 16; my emphasis) What is important for our argument here is that the external world itself functions as a significant constraint on imaginability and its role in the meaning process. As a third constraint on Locke’s private notion of meaning, the careful reader will discover in the Essay the view of language as well ordered public discourse, both syntactically and semantically. This judgment is corroborated by Locke’s insistence on the constancy of the association between signifiers and signifieds within a speech community. “As far as words are of use and signification, so far there is a constant connexion between the sound and the idea”. (III, ii, 7) If Locke had intended linguistic meanings to be private, in the sense of being solipsistic and merely subjective, he would most certainly not have committed himself to this sort of characterization. While leaving open the possibility of any nonpublic use and non-signification of language as a form of misuse, what is crucial here is the persistent association between words and their meaning whenever we employ speech for public purposes. What we should avoid 34

Margaret A. Crouch, “Locke on Language and Reality,” Locke Studies 1 (2001), 87-104.

Resemblance in Language: Locke

29

though, says Locke, is to assume that because signifiers and signifieds have conventionally been associated in a certain way for a long time that there exists a “natural connection between them”. Instead, they have been associated as a result of a “perfectly arbitrary imposition,” to be understood as “common use, by tacit consent”. Earlier, Locke had already defined “voluntary imposition” as a convention “whereby such a word is made arbitrarily the mark of such an idea”. (III, ii, 1) Anyone who does not respect this social agreement “does not speak properly”. (III, ii, 8; my emphasis) For if we do not respect this fundamental and “constant connexion between the sound and the idea, and the designation that the one stand for the other,” then language is nothing but “insignificant noise”. (III, ii, 7; my emphasis) Apart from Locke’s emphasis on shared linguistic meanings on top of conventional public signifiers, we must also briefly pay attention to his notion of arbitrariness. 35 Locke’s arbitrariness thesis must be carefully distinguished from its radical version in Saussure in that Locke always treats the signified as motivated. Such partial arbitrariness allows Locke to argue the systemic nature of language in terms of its double directionality, pointing as it does to the world and back to the speaker as bound by the speech community. In order to avoid the “parrot” use of language, we must not, says Locke, pay attention only to the words themselves but also always to the things they are made to represent via Vorstellung. (III, ii, 7) There is the justified assumption about language that its speakers “would not be thought to talk barely of their imaginations,” their merely private fantasies, “but of things as really they are”. Speakers are justified to assume that by and large words “stand also for the reality of things”. (III, ii, 5) That imaginability is a sine qua non, though not in its merely private use, should be obvious since typically linguistic meanings do not require the actual presence of objects.36 Meaning largely works, as it were, with the world in perceptual absentia. (III, ii, 6) At the same time, we must not forget the main purpose of language, which Locke says is communication, “the chief end of language,” (III, iii, 3) and a “very short and expedite way of conveying [our] thoughts to one another”. (II, xviii, 7) For every speaker, “the end of speech is that those sounds, as marks, may make 35

In its non-controversial sense as restricted to the signifier, we find “arbitrariness” in Locke discussed also by John J. Jenkins, Understanding Locke (Edinburgh: University of Edinburgh Press, 1983), 156. 36 Though a necessary component of the meaning process, the imagination is not coherently argued by Locke, as pointed out by Roland Hall in “Some Uses of the Imagination in the British Empiricists: A Preliminary Investigation of Locke, as Contrasted with Hume,” Locke Newsletter 20 (1989), 47-62; 57.

30

Chapter One

known his ideas to the hearer” (III, ii, 2) on the reciprocal assumption that speakers “use the words as they imagine in the common acceptation of that language”. (III, ii, 4) The systemic nature of language, then, is also shored up by use as shared imaginability.37 The constraint imposed on meaning by the systemic nature of language in Locke is consolidated further by the tasks he allocates to the various components of speech. The following distinctions amongst expressions can be found in the Essay, together forming an interacting network of linguistic functions. Imaginability is regulated by such signifiers as (1) conjunctions; (2) disjunctions; (3) simple terms (natural kind terms); (4) complex terms (culturally saturated terms); (5) reference classes; (6) sentences (7) clusters of sentences (discourse); (8) propositions (for the purpose of distinguishing truth and falsity); (9) formal expressions (mathematics; logic); (10) terms suggesting intentional objects; and (11) terms for purely intentional objects (objects for which there is no equivalent in the actual world). In the Essay, general names as classes derived from particular names and proper names furnish aboutness. Particles guarantee the kind of connections that make both the signified world and language into interrelated, systemic webs. (III, vii) Proper names signify particular substances, each with co-existing qualities. Verbs signify “the different modifications of motion” such as “to slide, roll, tumble, walk, creep, run, dance, leap, skip,” expressions suggestive of “the different modifications of motion”. (II, xviii, 2) Negative signifiers contribute to the systemic character of language in that they all signify the absence of a quality (rather than a negative quality). Negative qualities do not exist in Locke’s scheme of things; there are only positive particulars with positive qualities. (III, I, 4) The arbitrary names of “sorts” accomplish linguistic reductions, that is, abstractions into species from a series of 37

In Lowe’s version of Locke’s ideational theory of meaning, we find an “adverbialist explanation” to the effect that instead of talking of “inner mental objects” Lowe prefers the notions of “modes of sensing” and “modes of quasisensing”; Jonathan E. Lowe, Locke on Human Understanding (London: Routledge, 1995), 148. Attractive as this intervention is, what it leaves unexplained are the “contents” of such modes. Because Lowe does not want to give up on standard conceptions of meaning, he comes to the conclusion that Locke’s theory of language should not be called a “theory of meaning” at all but rather a theory of “expressive relations,”150. I take the opposite route. By salvaging “a good deal more” from Locke than is usual, I question some standard descriptions of “meaning”; Lowe, 165.; cf. also Lowe, Jonathan (2010) “Ideational Theories of Meaning,” Concise Encyclopedia of Philosophy of Language and Linguistics, ed. Alex Barber and Robert J. Stainton (Oxford: Elsevier, 2010), 299-301.

Resemblance in Language: Locke

31

selected particulars and so too contribute to language as network. (III, iii, 4) Yet Locke’s “sorts” not only consolidate language but also strengthen our coherent conception of reality. (III, vi, 46) Here sorts and general names interact systemically. (III, vi, 29f.) But Locke did not only address the systemic nature of language in terms of nouns, verbs, adjectives, and adverbs, but also for syncategorematic terms under the term “particles,” (III, vii, 1-5) which has found recent support in arguments for the nonformal origin of conditionals, conjunctions and disjunctions.38 Whatever boundaries the systematicity of language imposes on the way we project mental scenarios in response to verbal cues, they are themselves under the control of the speech community. This is why Locke makes the speech community his fourth and most important constraint on linguistic meaning. It is the speech community which, “by a tacit consent, appropriates certain sounds to certain ideas in all languages” and so “limits the signification of that sound”. (III, ii, 8) Earlier in the Essay we read, “Men one amongst another,” Locke’s phrase for society, invented names for “complex ideas,” so that they “might the more easily record and discourse of those things they were daily conversant in” and “that the things they were continually to give and receive information about might be the easier and quicker understood”. (II, xviii, 7) This is why, Locke warns, “the signification of words are very warily and sparingly to be altered”. (III, vi, 51) After all, language is “no man’s private possession”; it is “not for anyone at pleasure” to “alter the ideas they are applied to”. (III, xi, 11) We can, but should not change the community sanctioned relation between linguistic expressions and Vorstellung. In other words, we should not undermine the “linguistic linkage compulsion” imposed on each speaker by the speech community. At the same time Locke is aware of a double directionality of constraint in language, from world to word and simultaneously from society to language. “Our names of substances,” says Locke, “being made use of ultimately to represent things” are such that “their signification must agree with the truth of things as well as with men’s ideas”. (III, xi, 24) Though hidden from view, people “ideas” are not conceived here as individually private but as part of signification at large. Taking Locke’s paradox of a “hidden” meaning event and its functioning within public discourse seriously, one way to resolve this aporia is to argue that though meaning events are hidden from view, the speech community nevertheless controls 38 As for instance in Eve Sweetser, From Etymology to Pragmatics (Cambridge: Cambridge University Press), 1991.

32

Chapter One

what is going on in the minds of its speakers. This is the avenue I recommend for the solution of Locke’s paradox between meaning and discourse. What are needed are reasons for how the control by the speech community over meaning as mental event can be made plausible. One such reason addressed by Locke is the intervention by the speech community in terms of what it regards as linguistic abuse, which the Essay describes as the “voluntary imposition” of signification deviating from community use. (III, x, 5) It is abuse because it fails to respect the “constant connexion between sound and idea,” (III, ii, 7) a constancy guaranteed by the speech community on the one hand and, on the other, by certain constraints imposed by nature on the way we experience the world. Locke’s emphasis on linguistic meaning as “my idea” throughout Book III does not invite the interpretation of linguistic meaning as ruled by any private licence. If we wish to speak “properly,” (III, ii, 8) we must respect the linguistic conventions of the community, including the kinds of Vorstellungen we have been trained from childhood to activate on cue. In this sense, the performance of meaning as an all-important part of language is inherited as a coherent system from society and as such is binding. What Locke rightly draws our attention to in his insistence on the personal mastery of a public language is that meanings are not in the first instance publicly visible. Meaning is nevertheless indirectly public if it is not to be “abuse”. This emphasis has been neglected in our theories of natural language since Frege’s elimination of Vorstellung from meaning as private, replacing it by fully public formalisation, and further eroded by behaviourist tendencies of externalisation restricting meaning increasingly to observables. That linguistic abuse can be argued even within a perspective eschewing the all or nothing of fully public meanings is a special and important feature of Locke’s semantic. Abuse, in Locke, includes “wilful faults” and “neglects” such that signification fails its “natural” clarity and distinctness, (III, x, 1f.) inconsistency of “use” (III, x, 5), “affected obscurity” (III, x, 6), and the conflation of words and things (III, x, 14). To avoid abuse, “it is sometimes necessary for the ascertainment and the signification of words, to declare their meaning”. (III, xi, 12) This, presumably, is done by paraphrase or “definition,” loosely conceived, a theme that occupies the last chapter of Book III in the Essay, persuasively discussed under the term “rectification” by Michael Losonsky.39 39

Michael Losonsky, “Locke on Meaning and Signification,” in Locke’s Philosophy: Content and Context, ed. G.A.J. Rogers (Oxford: Clarendon Press, 1996), 123-141; 134.

Resemblance in Language: Locke

33

Locke’s emphasis on the proper use of language in both its syntactic and semantic aspects seriously qualifies his insistence on the “hidden” character of the meaning event. Yet rather than regarding this as a major flaw in the Essay, we should acknowledge that Locke is aware of this as a tension in his semantics. Even though individual meaning performance is hidden, when speakers of a language mentally project what the speech community has taught them in response to signifiers, they do so in conformity with “the common acceptation of that language, in which they suppose that the idea they make it a sign of is precisely the same to which the understanding men of that country apply that name”. (III, ii, 4) For “by use or consent, the sound I make by the organs of speech, excites in another man’s mind, who hears it, the idea I apply it to in mine, when I speak it”. (III, iii, 3) Under this communal agreement it is possible to “make known one man’s thoughts or ideas to another”. (III, x, 23) Thus “common use regulates the meaning of words pretty well for common conversation”. (III, ix, 8; my emphasis) In spite of the fact that language is a “very variable standard” (III, xi, 25) and that “there are few names of complex ideas which any two men use for the same just precise collection,” (III, x, 22) Locke leaves no doubt as to the ordinary functioning, indeed the “necessity of communication by language”. (III, xi, 25) Language remains throughout the Essay, “the great instrument and common tie of society”. (III, i,1) But have we done enough to argue the “tie” between hidden meanings and social expectations? Has the chapter been able to apply Wittgenstein’s rule that “an ‘inner process’ stands in need of outward criteria”?40

Conclusion: dissolving Locke’s paradox As far as Locke’s “ideas” in the mind of a speaker are a component of language, they are not private in the sense of being a private invention. They are “hidden” only in the sense that they are not observable. They are nevertheless indirectly public and so social because of their multiple constraints of publicly observed world, human physiognomy, the systemic nature of language and, above all the speech community. This is why Ashworth rightly emphasizes what Locke “has to say about … how it is that our ideas conform both to the ideas of other men and to external objects”. 41 The dissolution of the paradox between Locke’s mentalist 40

Wittgenstein, PI §580. Earline J. Ashworth, “Locke on Language,” in Locke: Critical Assessments, ed. Richard Ashcraft (London: Routledge, 1995), 235-258; 235. 41

34

Chapter One

account of meaning and his insistence on the governance of language by the speech community requires a review of what we mean by “idea” as Vorstellung and its constitution in the event of meaning. Traditional attempts at such a reconciliation of seemingly irreconcilable opposites has foundered on the assumption that all Vorstellungen are private and so cannot be employed in any definition of linguistic meaning. No doubt, there are private Vorstellungen. They are what Wittgenstein is castigating when he uses the colourful expression of Vorstellungsklavier. They are also what Frege thought had to be eliminated from the description of meaning, and so not surprisingly they have since been rejected by the bulk of natural language semantics on the grounds of psychologism. Crucially, though, at the moment of linguistic comprehension, imaginability does not function as imaginative free play. Rather it is strictly administered by social rules transmitted via pedagogy from the cradle to the grave. We recall that in the Essay Locke’s refers to a “boy” who imagines “shiny” when he hears the signifier “gold”. (III, ii, 3) In adult language use this would fall under “abuse”. (III, x) But this criterion does not apply to someone who is not yet in full command of the language. Pedagogy is what makes the essential difference. Once fully trained, the boy’s category mistake is avoided as a result of speakers of English having been trained to entertain specific mental states, exclusive hypoiconic, generalised schematizations. In such a picture we need to come to grips with abstraction as hypoiconic generalisation. Locke seems to have been right in wanting to have it both ways, meaning as a hidden event and as publicly controlled. The accusation, then, that Locke cannot hold on to both mental events and speech community rests on false assumptions about imaginability. From the perspective of a semantics in which the mediation of world and community is achieved via Vorstellung in the sense of “mental, iconic transformation of perception,” the two, mental performance and social constraint are a necessary double condition of language. In such a perspective, meaning as mental event is to be viewed as indirectly public. As such, Locke’s foregrounding of the “immediacy” of ideas during meaning events is not just respectable but necessary. After all, a more or less automated response to signifiers is bound to occur at very high speed beyond the control of the individual. That at the centre of the communicative act there should occur certain mental processes performed by speakers, then, legitimately remains a central feature of Locke’s theory of natural language. In such a theory, individual variations of meaning events can be accommodated as a typical rather than aberrant characteristic of language, one of those that make it “natural”. Locke was

Resemblance in Language: Locke

35

right then it would seem in putting his finger so firmly on the one thing about language that is not observably public, but merely indirectly public, the event of linguistic meaning.

CHAPTER TWO ICONICITY IN KANT AND PEIRCE

Introduction Two philosophers who have made significant contributions to the way we can account for Aristotle’s homoiomata, or resemblance relations, in human cognition are Immanuel Kant and Charles Sanders Peirce. Kant did so in his analysis of the conditions that make cognition possible in terms of Anschauungen, their schematization via concepts and a priori presuppositions, as well as by his differentiation of seven different forms of reasoning; Peirce in his theory of iconic, indexical and symbolic signs and especially in his subdivision of iconicity. What links the two is not just Peirce’s considerable and broad debt to Kant, as Peirce himself happily acknowledges, but also a shared commitment to the all-important role resemblance relations play in the human mind. And although neither philosopher focuses primarily on language, Kant because his critical enterprise is broadly epistemic and Peirce because his interest lies primarily in general semiosis beyond human mental acts, both hold positions which fundamentally shore up arguments in favour of the presence of imaginability as a necessary condition of meaning events in language.

Kant: from idea to schema “All our knowledge begins with experience”, writes Kant in agreement with Locke, but “it does not follow that it all arises out of experience” (CPR B1). This announces the care Kant will take in distinguishing between the fact of objective reality and its perspectival representation by the cognizing organism. At the same time, this disjunction in the process of the emergence of resemblance relations puts the accent on Kant’s unwavering emphasis on external reality affecting our senses as the origin of all knowledge. Being “affected by objects … alone yields us Anschauungen”. In turn, Anschauungen are processed in the understanding by concepts (CPR A19). Objectivity as the ground of experience and

Iconicity in Kant and Peirce

37

knowledge is not in question. The external world transcends its representation. 1 As Westphal observes, Kant was “the first great antiCartesian” and a “staunch externalist avant la lettre”. 2 But now the externalist picture is complicated by Kant’s systemic concept of the faculty of Vorstellung. In order to leave Kant’s take on imaginability intact, I leave the German term Vorstellung, since its English translations as “representation”, “mental presentation”, “conception”, “idea”, or sometimes “imagination” introduce a conceptual murkiness that may have led as sharp a mind as Gilbert Ryle’s astray when he rejected it out of hand in the Concept of Mind.3 For Kant, Vorstellung covers anything we can become aware of, as for instance iconic versus abstracted conceptual grasp, “Anschauung” versus “Begriff”, which he regards as two different “Vorstellungsarten” (kinds of imaginability).4 Another reason for staying with the Kantian vocabulary is to emphasize the difference between Vorstellung as perceptual modification and Einbildungskraft as its productive sub-faculty. Evolutionary biology has now provided us with observations which can be adduced if we wish to re-emphasize Vorstellung in the sense of neurally based modification of perception by drawing attention to the asymmetrical relation between the relative paucity of perceptual input cells and the quite remarkable abundance of billions of neurons monitoring that input, as well as manipulating it for nonperceptual, imaginative purposes.5 According to Kant, imaginability is essential to the way we process perceptual input. The first base for cognition is what affects the senses, sensation being defined as “the effect of an object upon the faculty of Vorstellung“(CPR A 20/B 34). Our capacity to be so affected, Kant calls “receptivity” (CPR A 26/B 42; A 50/B 74). What we receive in this manner has a material and a formal side. The matter of appearance 1

Karl Ameriks is correct in my view that in spite of Kant’s emphasis on the human cognitive process, objective reality as our reality is beyond doubt. See his Kant’s Theory of Mind: An Analysis of the Paralogisms of Pure Reason (Oxford: Oxford University Press), 1982. 2 Kenneth Westphal, Kant’s Transcendental Proof of Realism (Cambridge: Cambridge University Press, 2009), 7. 3 Gilbert Ryle, The Concept of Mind (London: Hutchinson, 1949), 232-263. 4 In a letter to J.S. Beck of 20. July 1792; quoted in Ralph Meerbote, “Rickert’s Auseinandersetzung mit dem Riehlschen Realismus,” in Kant-Studien, 86 (1995), 346-362; 348. 5 Tecumseh Fitch, “Nano-Intentionality: A Defense of Intrinsic Intentionality,” Biology and Philosophy 23 (2008), 157-177; 170; this important paper will be addressed in detail in Chapter Eight.

38

Chapter Two

“corresponds to sensation”, while its form “so determines the manifold of appearance that it allows of being ordered in certain relations” (CPR A 20/B 34). In other words, objective reality strikes twice, once in dictating what we are able to sense and, a second time, by determinately prefiguring the way in which the mind is able to experience what has been sensed as ordered. At the level of receptivity, then, reality calls the shots. “Cognition of any kind requires an object somehow given”. 6 Certainly, Anschauung is insufficient for cognition and requires the “cooperation of the spontaneity of understanding”. 7 What matters here, is to establish Kant’s perspectival realism as the basis of his pragmatism. We can do so by partially adopting Westphal’s observation that “the ultimate implication of Kant’s own transcendental analysis of the a priori conditions for the possibility of human experience is a transcendental proof of realism sans phrase about physical objects”, except that we cannot call Kant’s approach a “transcendental realism”.8 Kant’s perspectival realist überzeugung can be successfully argued only, it would seem, via his method of transcendental idealism. The second level of cognition is the imaginative variation of perceptual input achieved by the sub-faculty of Vorstellung, the Einbildungskraft. If our senses make up our “outer sense” and “the sum of all Vorstellungen” (CPR A 177/B 220) which is our “inner sense” effecting a “modification of the mind” (CPR A 98), then the Einbildungskraft is the faculty in which perceptual modification takes on a life of its own. It manifests itself as “the faculty of representing in intuition an object that is not itself present” (CPR B 151), in its empirical form serving the lower levels of cognition and in its overarching capacity acting as “indispensable power of the soul, without which we should have no knowledge whatsoever” (CPR A 78/B 103). As such, Vorstellung “frames out of the relations abstracted from experience something that does indeed contain what is general in these relations, but which cannot exist without the restrictions which nature has attached to them” (CPR A 40/B 57; my emphasis). So constrained is the cognitive process that even the most abstract relations which we can imagine in response to our Empfindungen (CPR A 28/B 44) via Anschauung and the hierarchy of conceptualization would not be possible if it were not for the way nature dictates what we can do with it. This is 6 Henry Allison, Kant’s Transcendental Idealism: An Interpretation and Defense (New Haven: Yale University Press, 2004), 77. 7 Ibid. 8 Westphal, Kenneth Kant’s Transcendental Proof of Realism (Cambridge: Cambridge University Press, 2009), 8.

Iconicity in Kant and Peirce

39

why Kant, not unlike Locke, so stresses the immediacy of the relation between Anschauung and object (CPR A 320/B 377). 9 The same relation of immediacy and dependency applies to the third level of cognition, “apperception”, the Vorstellung of the “I” (CPR B 68). Vorstellung in this respect embraces everything we can be conscious of. As Kant notes, “all possible appearances, as Vorstellungen, belong to the totality of a possible self-consciousness” (CPR A 113; my emphasis), which guarantees “the thoroughgoing identity of the self in all possible representations” (CPR A 116). Kant insists that “it must be possible for the ‘I think’ to accompany all my representations” (CPR B 131). Ultimately, however, Kant’s subjective generality is parasitic on objectivity, for “we must derive the subjective succession of apprehension from the objective succession of appearances” (CPR A 193/B 238; my emphasis). Speaking of a ship drifting downstream, Kant says, “the order in which the perceptions succeed one another in apprehension is … determined [by reality], and to this order apprehension is bound down” (CPR A 192/B 237; my emphasis). The necessity of the sequence of perceptions is not our addition to reality, though we need Vorstellung to apprehend it (CPR A 176/B 218). In sharp contrast, “the subjective succession by itself is altogether arbitrary” (CPR A 193/B 238). Yet even this kind of subjective experience as well as “inner experience in general is possible only through outer experience in general” (CPR B 277). Our senses are themselves part of nature, so much so that they are incapable of erring, “not because they always judge rightly but because they do not judge at all” (CPR A 293/B 350). Error has its source in something else. “Error is brought about solely by the unobserved influence of sensibility on the understanding” (CPR A 294/B 350). That the objectivity of reality outside ourselves has primacy over human cognition results in “inner sense” being overruled by “outer empirical Anschauungen” (CPR B 293). Objective reality in Kant reveals itself to us within “the unalterable limits of nature” (CPR B 294). It is objective succession rather than its subjective realization that is the master. Much the same can be argued for Kant’s employment of the notion of space. Kant does not deny that there must be a condition of space from

9

Henry Allison I think rightly rejects Hintikka’s reading of immediacy as “a mere corollary”; in Kant’s Transcendental Idealism, 80f.; quite the contrary, “immediacy” is crucial in underlining the absence of mediation. At this most basic level of cognition, the understanding has no say in the matter: objectivity rules.

40

Chapter Two

which the a priori of human space derives.10 Paul Guyer makes the point that space in general may indeed have been Kant’s reason for distinguishing it as transcendental demand from space as a category of human cognition.11 Kant sums up the three levels of cognition thus: “sense represents appearances empirically in perception, imagination in association (and reproduction), apperception in the empirical consciousness … of recognition” (CPR A 115). While Kant carefully attends to the cognitive process, he leaves no doubt that it is objective nature that is in charge of the way by which that process unfolds. Humans are part of the very same world they so experience. It is this base condition of being both world and perspectivally cognizing creature that significantly qualifies the notion of pragmatism in the Critiques. What is realistically pragmatic here, according to Peirce, is that Kant recognizes what is “mental” in our concepts about the real.12 Kant’s remarks on the natural restrictions of cognition introduce a serious specification of his observation that senses intuit, understanding thinks, and apperception reflects. Now, all three are subject to the rule of natural constraint, the limits set for knowledge by “objektive Realität” (CJ §65). But even if we were to say that cognition frees itself from nature to the extent that the understanding “spontaneously” conceptualizes affective content, since Kant insists on their unity produced in apperception, the strictures of nature survive at all levels of cognition. After all, “the understanding can intuit nothing, the senses can think nothing. Only through their union can knowledge arise” (CPR A 51/B 75). Even though “concepts are based on the spontaneity of thought”, while “sensible intuitions” rely “on the receptivity of impressions”, and apperception on self-conscious awareness, all three are subject to a reality check (CPR A 68/B 93). Though “thought” is spontaneous, it “can become knowledge for us only in so far as the concept is related to objects of the senses” (CPR B 147; my emphasis). Yet no matter how far the human mind pursues the increasingly abstracting hierarchization of Vorstellungen, Kant’s “function” (CPR A 68/B 93), they could not exist at all “without the restrictions which nature has attached to them” (CPR A 45/B 57). In other words, Vorstellung as conceptualized modification of perceptual input is part of 10

Henry Allison, in Kant’s Transcendental Idealism, 55, rejects the idea that “the expression appearance is parasitic upon, or at least correlative with, the expression thing-in-itself”. 11 Paul Guyer, Kant and the Claims of Knowledge (Cambridge: Cambridge University Press, 1987), 430. 12 Peirce, CP 5.525.

Iconicity in Kant and Peirce

41

the reality of nature in as far as it sustains “the relation in all our empirical concepts to an object” (CPR A 109), which supports the anthropic thesis that we shouldn’t be surprised to find that “natural or physical reality is essentially not alien to our minds”. 13 The “objective reality of our empirical knowledge” includes us (CPR A 110). And beyond our immediate empirical grasp of reality, even the “purposiveness” of nature which we cannot but infer from our observations of “organized and selforganizing” organisms is “objektive Realität” (CJ §65). These remarks in the Critique of Judgement are foreshadowed in the first Critique where Kant speaks of “the systematic unity, order, and purposiveness of the arrangement of the world” as “an idea which reason is constrained to form as the regulative principle of its investigation of nature” (CPR A 697/B 725). Yet, what appears to be an argument from idealism reveals its realist ground since our inferences about the “order and purposiveness in nature must themselves be explained from natural grounds and according to natural laws” (CPR A 772f./B 800). The pragmatic implication is that if we violate this sequence, we cannot successfully function in the real. Kant’s perspectival realism is further specified by his schematism according to which the resemblance relations between objective reality and human cognition are no straightforward mimesis of percepts. “It is schemata, not images of objects, which underlie our pure sensible concepts” (CPR A 141/B 180). Generalizing, Kant writes, “the schema of a reality, as the quantity of something in so far as it fills time, is just this continuous and uniform production of that reality in time” (CPR A 143). This should not be read to mean that we subjectively construct a “reality” that is not there. Nothing could be further from Kant’s insistence on the constraints imposed by nature on our faculties. The human schematic construction of “actuality is existence in some determinate time”; just as the “schema of necessity is the existence of an object at all times”; and all within the realist boundaries of what nature allows us to realize both in epistemic and ontological terms (CPR A 145/B 184). Kant’s mind-independent reality is shored up also by two other devices in his critical toolbox, his notion of what is actual and his characterizations of truth. Actuality is intimately associated in Kant with perception. “Whatever is connected with perception in accordance with empirical laws is actual, even if it is not immediately perceived” (CPR A 231/B 284). This qualification strengthens the impressions that Kant is talking about an objective reality 13

Robert Hanna, “The Kantian Revenge: On Forster’s Kant and Scepticism,” Kantian Review, 17, 1 (2012), 33-45; 39.

42

Chapter Two

independent of our cognitive faculties, for whenever there is a connection between objective reality and “perception according to empirical laws”, we are dealing with the “actual” (CPR A 376). Object, sensation, perception and actuality, then, form an order dictated by nature. If the object “stands in connection with perception, that is, with sensation as material supplied by the senses, and through perception is determined by means of the understanding, the object is actual” (CPR A 234/B 286). In this sense, “all outer perception … yields immediate proof of something real (Wirkliches) in space, or rather is the real itself” (CPR A 375). Resemblance relations in Kant, then, have a firm, primary footing in the objective reality of the world we bump into. On this basis we can now address Kant’s controversial notion of schematization. In the schematism chapter of the Critique of Pure Reason, Kant writes the following wellknown sentence. “This much only we can assert: the image (Bild) is a product of the empirical faculty of the reproductive imagination (Einbildungskraft); the schema of sensible concepts (such as figures in space) a product and, as it were, a monogram, of pure a priori imagination, through which, and in accordance with which, images themselves first become possible”. Imaginability and schematization are intricately linked, for images can be connected with concepts only by means of the schemata which the images indicate (bezeichnen). Nor are images ever fully “congruent” with their concepts. (CPR A141/B181) In response to the charge of obscurity that has been laid against Kant’s “schematism” section in the first Critique, including Schopenhauer’s celebrated denunciation, one could say that the principle of schematization nevertheless stands. When Schopenhauer compares Kant with someone who derives the height of a tower from measuring its shadow, when he would simply put the measuring tape on the tower itself, he nevertheless defends the schema as a viable principle. Indeed, the measuring tape is precisely such a Kantian schema. “The purpose of schemata in empirical (real) thinking refers solely to the material content of such concepts” which are “abstracted from empirical perception”. 14 Whether construed via a top-down inferentialism or by a bottom-up process of abstraction, as far as its function is concerned, the schema works, unaffected by its genesis. Moreover, a Kantian, in agreement with Locke, may very well ask how Schopenhauer would explain our peculiar ability of deriving schemata from apperception. What is needed to answer such a question is an 14 Arthur Schopenhauer, Die Welt als Wille und Vorstellung (Zürich: Diogenes, 1977), 555; 552.

Iconicity in Kant and Peirce

43

evolutionary framework which neither Locke nor Kant, nor Schopenhauer had access to. Kant could do no more than offer speculative answers by way of transcendental reduction. The “concept of dog,” he writes, “signifies a rule according to which my imagination can delineate the figure of a four-footed animal in a general manner”. (CPR A141) But what can we make of “allgemein verzeichnen”? Draw, sketch, trace, or outline? A figure like a diagram or like a child’s drawing? Which of Peirce’s three hypoicons is applicable here? Iconic signs by virtue of direct resemblance (images), by way of abstraction and idealisation (diagrams), or by displacement and parallelism (metaphors)?15 Kant specifies that this kind of Vorstellung is not to be limited to any “single determinate figure such as experience … actually presents,” nor “any possible image that I can present in concreto”. (CPR A141) At the same time, images can be connected with concepts “only by means of the schema which they indicate”. (CPR A142) In the Norman Kemp Smith translation, Kant’s “welches sie bezeichnen” is rendered as “to which they belong”. I prefer the literal translation of “indicating” because it suggests that Kant chose his phrasing in recognition of a directionality emanating from the image towards its possible schematizations. “We cannot think a line without drawing it in thought,” writes Kant. (CPR B154) Does this mean that we proceed likewise when using language? Are we producing some iconic analogue of an olfactory realisation when we think we hear a linguistic expression about a certain fragrance? And do we proceed likewise when we read the phrase “a hint of orange jasmine”? There are two competing traditions inclined to give contrary answers to such questions, one of which emphasizes nonverbal, quasi-perceptual residues in schematized Vorstellung, again understood as mental projection and perceptual modification. This view, perhaps inspired by Plato’s eidos, can be found in John Locke’s “ideas” which are communicated via general terms that reduce particulars “into sorts” by way of “abstraction” and “comprehensive extension”. Importantly, general ideas and their names have their foundation in the “similitude our understanding observes amongst” things.16 Resemblance relations inform Kant’s “monogram”, Husserl’s type, appresentation, as well as his noema,

15

Charles Sanders Peirce, Writings of Charles S. Peirce: A Chronological Edition, vol. 2. 1867-71. Peirce Edition Project (Bloomington: Indiana University Press, 1984), EP 2.273. 16 Locke, Essay, 231-236.

44

Chapter Two

and the “mental corporeality” (geistige Leiblichkeit) of language, 17 the characterisation of language as “sensory images awakened” in William James,18 Cassirer’s “sensuous abstractions” and “schematised percepts,”19 and Vygotsky’s notion of language as a “generalised reflection of reality,”20 not to mention a host of similar explanations in psychology. The second tradition appears to be strongly indebted to Aristotle’s propositional path along which the mental content of sentences was gradually reduced to Frege’s “pure thought” and by some of his successors, as for instance Carnap and Quine, to an intensional notion of meaning which, as I will discuss later, has culminated in what has become known as hyperintensional semantics. The two traditions are still visible in contemporary debates wrangling over the definition of the concept. 21 Clearly, in terms of our basic cognitive faculties Kant stands squarely in the first tradition of Aristotelian resemblance relations. But instead of saying that images belong to certain schemata, Kant seems to be saying that images direct us as to the kind of schematization we are capable of performing and require for certain purposes. The literal translation in this case allows for a feed-back relation between percept and schema. Yet these observations still leave in abeyance the degree and manner of schematization we typically employ. Importantly, as quoted above, an image is “never completely congruent with the concept”. (CPR A142) Presumably, this is so because the concept is the result of a schematization at a specific level of abstraction and of a certain kind, each of which renders identity as well as thoroughgoing congruence impossible. Nevertheless, the Vorstellung of the object “must be homogeneous with the concept”. When we look at a plate, says Kant, we can recognize the geometrical concept of a circle in it. “The roundness 17

Edmund Husserl, Formale und Transzendentale Logik, ed. P. Jenssen. Husserliana XVII (The Hague: Martinus Nijhoff, 1976), 25. 18 William James , Principles of Psychology (New York: Dover, 1950), 265. 19 Cassirer, Ernst, Philosophy of Symbolic Forms, v.3, trans. R. Manheim (New Haven: Yale University Press, 1957), 331. 20 Lev Vygotsky, Thought and Language, trans. G. Vakar and E. Hanfmann (Cambridge, Mass.: MIT Press, 1962), 153. 21 Donald Davidson , Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984); John McDowell, Mind and World (Cambridge: Harvard University Press, 1996); Jerry Fodor, Concepts: Where Cognitive Science Went Wrong (Oxford: Clarendon Press, 1998); Eric Margolis and Steven Laurence, Concepts: Core Readings (Cambridge, Mass.: MIT, 1999); Ray Jackendoff , “What Is a Concept That a Person Can Grasp It?" in Concepts: Core Readings, ed. Eric Margolis and Stephen Laurence (Cambridge, Mass.: MIT Press, 1999), 305-333.

Iconicity in Kant and Peirce

45

which is thought in the latter can be intuited in the former”. (CPR A137/B176) And “the only manner in which objects can be given to us is by modification of our sensibility” and as a result of “certain formal conditions of sensibility, namely those of inner sense”. It is this “formal and pure condition of sensibility to which the employment of the concept of understanding is restricted” that Kant calls “the schema of the concept,” while the process by which we understand with the assistance of schemata he calls “the schematism of pure understanding”. (CPR A139f. /B179f.) But why should there be schemata at all? Why can we not simply proceed from Anschauung to concept? Kant’s answer is that the two are so different that we could not explain the transition from one to the other. The only way to align such different entities as specific percepts and generalised concepts is a tertium comparationis, “some third thing”. A special kind of Vorstellung fulfils this role in the Kantian scheme: a “mediating Vorstellung” which is “void of all empirical content” yet still “sensible” and so “homogeneous” with both percept and concept, a Vorstellung which is at the same time intellectual and sensual. This Vorstellung he calls the “transcendental schema”. (CPR A138/B177) Since like in “conceptual blending”, 22 Vorstellungen are “ordered, connected, and brought into relation,” they are not only static entities but also dynamic bridges between different cognitive domains. (CPR A99) What matters to our discussion is the transition from Anschauung to schema, concept, and linguistic concept. What matters are the various degrees to which images and their non-visual equivalents such as olfactory, gustatory, thermal, gravitational or other nonverbal readings are transformed, resulting in schematised concepts at different levels of generality and kind. In cases where we cannot have an iconic Vorstellung as a starting point, we go straight to a schema. Kant here sharply distinguishes between understanding the number five as an image of five consecutive points from understanding a very large number for which no image would be a sufficient representation. In this case, he says, we are dealing with the Vorstellung of a method; we imagine a schema rather than an image. Thus, “pure sensible concepts,” that is, concepts without iconic mental materials, are not based on images but on schemata alone. Applied to our discussion of natural language, this encourages us to think of the schema not as a static entity in between Anschauung and concept, but rather as a faculty of schematising, producing different degrees of 22

See my discussion of Gilles Fauconnier and Mark Turner’s notion of “conceptual blending” in Chapter Seven.

46

Chapter Two

schematization and different manners of schematization, according to social and linguistic context. The summary schema of this dynamic ladder, with its rungs from sense impression to concept by way of differentiating schematization, is the schema of reality, its “continuous and uniform production” in time. (CPR A143) At this point, Kant tries to dissuade his readers from pursuing the matter further by telling us that the “schematism of the understanding … is an art concealed in the depths of the human soul, whose real modes of activity nature is hardly likely ever to allow us to discover”. (CPR A141/B180f.) This sentiment is echoed in Wittgenstein’s Philosophical Investigations where he speaks of “an unbridgeable gulf between consciousness and brain-process”. (PI §412) But even if the actual causal nexus between neurons in the brain and acts of consciousness is not a philosophical problem but rather a question for science to resolve, whatever answers will be given in the future will have profound philosophical repercussions. What we must ask here is this. If the act of schematising produces elliptic results, as it must, what is being left out and what is left over? If the signs generated in the schematising process reduce the content of mental images and their non-visual equivalents, what kind of reductions are we talking about? And what of Vorstellung is being so reduced? What has been taken away (aphairesis)? Is it its totality, leaving intact certain salient resemblance relations and perhaps a specific manner and degree of iconicity? Such questions were part of the kind of thinking that occupied Charles Sanders Peirce, which I will address below. Before I do so, I want to add another facet of Kant’s contribution by taking a broad perspective of the three Critiques, Kant’s seven forms of reasoning and the different conceptual schemata they require.

Imaginability and Kant’s seven forms of reasoning Looking at Kant’s variety of reasoning from the angle of imaginability we gain an additional vantage point from which to consider the role of mental iconic transformations of perception in language. The concepts with the least involvement of imaginability are Kant’s mathematical schemata, employed in the construction of formal systems. As he says about formalised reasoning, even where it is constitutive as it is in all forms of mathematization, it has as its main task the “prevention of error”.23 The 23

“The greatest and perhaps the sole use of all philosophy of pure reasoning is therefore negative; since it serves not as an organon for the extension but as a

Iconicity in Kant and Peirce

47

syntactic relations amongst its concepts prescribe both subject and predicate of its propositions as definitionally ruled and stripped of imaginability. Here imaginability is reduced to the performance of syntactic relations. As such, Kant’s formal reason lies outside the boundaries of the intersubjective mentalism involved in linguistic meaning events. Mathematical or formal reason has to be carefully distinguished in Kant from what he calls pure reason. Whereas the former covers stipulated, axiomatic and definitional reasoning procedures, the latter deals with concepts abstracted from Anschauungen. While Kant regards them both as empty in terms of content, they differ sharply in their genesis. In mathematical and all formal reasoning we proceed from stipulated concepts to intuition, whereas in the concepts of pure reason are the result of generalisation from intuitions to high level abstraction (CPR A160/B199) and so to pure representations, conceived by Kant as representations “in which there is nothing that belongs to sensations”, and their imaginable mental variations one could add. (CPR B34) Likewise, Kant defines pure thought as “the form of thought in general”. Thus the a priori thought of space and time in general are arrived at by abstracting from the specifics of the discursive concept to the point where we are left with two empty notions, stripped entirely of imaginability. According to Kant, we arrive at the iconic purification of sensible concepts “when there is no mingling of sensations” (CPR A50/B74) in our concepts and when they are no longer “mixed with anything extraneous”. (CPR A11/B24) Whether this is indeed what occurs when we consider time and space as such or whether we introduce them as stipulated entities must remain a moot point. Suffice it here to leave Kant the last word on his differentiation: stipulated formal concepts allow us to see “the universal in the particular”; discursively derived, philosophical, pure concepts reverse this relation in that they permit us to see “the particular in the universal”. (CPR A713f./B741f.) So much appears to be clear: since formal particulars are empty by definition and since the universal of pure concepts are fully abstract, imaginability is radically reduced in both. Kant’s third kind of reason refers to reasoning with empirical concepts, which he sharply distinguishes from the schemata of pure and especially formal reason. While formal concepts are ruled by definitions, empirical concepts are shown to be fuzzy in two important respects. For one, the empirical concept “cannot be defined at all, but only made explicit”. It discipline for the limitation of pure reason, and, instead of discovering truth, has only the modest merit of guarding against error,” Kant CPR A795/B823; cf. also A730ff./B758ff.

48

Chapter Two

follows that therefore, the “limits of the concept are never assured”. Two, even if we cannot but apply empirical concept all the time in spite of this deficiency, while trying to refine its applicability to phenomena, “the completeness of the analysis of my concept is always in doubt”. This is why Kant prefers to characterize the description of empirical concepts as an “exposition”. (CPR A727ff./B755ff.) From the perspective of imaginability, we notice a fundamental difference between formal and pure concepts one the one hand and empirical concepts, on the other. Whereas the former can convey neither aboutness nor voice, both components are essential for the empirical concept in linguistic expressions. Hence in the empirical concept imaginability enjoys a powerful presence. In the comprehension of a linguistic concept, which is always an empirical schema or its imaginative extension, interpretive meaning constitution must contain mental iconic projections. Without them we would not be in a position to explore the concept’s meaning potential. In habitual speech this meaning process appears compressed into an event characterized by the “linguistic linkage compulsion”, the inescapable, community driven mental linkage of word sound and motivated signified. Kant’s fourth form of reasoning, practical reason, in combining social phenomena with stipulated rules for moral conduct requires imaginability in two senses.24 Society must be able to imagine the kinds of cases a moral maxim is supposed to cover and its individual members must be able to imagine how their conduct relates to any such rule. In this sense moral maxims and their formulation in legislation are a special case of natural language. They highlight the need of imaginatively varying the scenarios covered by moral rules in the very act of comprehension. Kant’s fifth form of reasoning, reflective reason, typically occurs in the judgment of complex situations. It is characterized by reflective concepts which Kant says require us to search for a law under which we are able to subsume a specific phenomenon for which we do not already have a determining set of concepts. The general principles required for judgment, therefore, have to be invented in the act of interpretation. In searching for such principles we rely on imaginability in the sense of having to use imaginative variation of suitable larger scenarios under which the phenomenon that caused us interpretive difficulties can be subsumed. Such a larger principle is provided by Kant’s sixth variety of reasoning under the term teleological reason, referring to the stipulated interpretive umbrellas we 24

Immanuel Kant, Kritik der praktischen Vernunft (Hamburg, Felix Meiner, 1967).

Iconicity in Kant and Peirce

49

cannot but invent in complex judgments.25 As such, teleological reasoning is indispensable in judgments about the purposiveness of nature and what Kant regards as the purposiveness without purpose of art. Without a large dose of imaginability the invention of interpretive frames could of course not get off the ground, an interpretive process that entails a mental, experimental playfulness. Applied to language, this means that any formulation of the unavoidable dialectic between bottom-up reflective concepts and top-down teleological concepts depends on the human ability to activate word sounds semantically by mentally varying relevant iconic scenarios. Kant’s seventh and final form of reasoning, transcendental reason, does not easily align itself with the six previous ones in that it is methodological. As such, it occupies a meta-level of reasoning and has been used by Kant in his differentiation of the preceding six forms. Nevertheless, transcendental reasoning as invented by Kant does not escape the deep constraint of imaginability. On the contrary, it could not be performed without it. What imaginability contributes is the mental material contents that have to be reduced towards their summary, the reduced conditions that make specific phenomena possible. For example, in Kant’s application of his transcendental procedure to moral reasoning he starts with the surface rules of specific prohibitions and then proceeds to their increasingly general formulations until he arrives at the categorical imperative, which is no longer as much a moral rule as a definition of the social. It is often called “empty” for lack of moral specifics. But leaving aside Kant’s controversial and in my view unwarranted condition of free will, even at the level of abstraction of the social from all cultural specifics, we require imaginability as a sine qua non if we are to comprehend the imperative in its most general form. Imaginability, then, plays a central role in Kant’s broad picture of human cognition and in his differentiation of a series of kinds of concepts by which we construct knowledge of the world, ourselves, and of formal sign systems. I use the term imaginability relevant to Kant’s vocabulary from Vorstellung (imagination) to Einbildungskraft (productive imagination), Sinnlichkeit (sensibility), Gemüt (mental state; self-affection), Rezeptivität (receptivity) and Spontaneität (spontaneity). In other words, imaginability in its dual sense of what can be imagined and our capacity to imagine partakes of much of Kant’s conceptual apparatus in the three Critiques. Because imaginability so conceived is about mental resemblance relations,

25

Immanuel Kant, Kritik der Urteilskraft (Hamburg, Felix Meiner, 1968).

50

Chapter Two

it is also closely related to the notion of iconicity theorized by Charles Sanders Peirce.

Peirce and the primacy of iconicity In spite of the fact that some scholars have played down his debts to Kant,26 there seem to be good reasons to think that Peirce was significantly influenced by Kant’s philosophy.27 This is so especially in the attempts by both Kant and Peirce at differentiating between forms of reason, in the relation of the definition of “reflective reason” with that of “abduction” and the involvement of hypothetical reason in the most basic empirical forms of experience. From the position advocated here, I want to emphasize also their shared commitment to mental iconicity. As Peirce famously wrote, “Every assertion must contain an icon or a set of icons, or else must contain signs whose meaning is only explicable by icons”.28 It is hardly possible to overstate the importance of this statement, especially since in human semiosis it refers specifically to consciousness. Peirce clearly commits himself to the presence of iconicity in much of our mental acts beyond perception and the importance of the translation amongst different sign systems in the process of understanding. For Peirce, signs and Vorstellung, even when generalised as “quasi-mind,” never completely part company, although he is far from arguing a subjectivist position. After all, the community, Kant’s sensus communis, is specifically acknowledged as the necessary background without which human sign processes could not take place. Most importantly, from the perspective taken here, Peirce updates Kant by providing tools with which we can pursue further Kant’s schematism and the monogram, especially in his subdivision of the iconic sign. These achievements can be read at least in part as a development of the Kantian picture of human cognition, including the role of sensus communis, and as a methodological transformation of transcendental reasoning into a pragmatic system. 29 26

E.g., John Deely, The Four Ages of Understanding (Toronto: University of Toronto Press, 2001). 27 Dinda L. Gorlée, Semiotics and the Problem of Translation (Amsterdam: Rodopi, 1994), 39f. 28 Charles Sanders Peirce, Collected Papers of Charles Sanders Peirce, 1931-1966, ed. C. Hartshorne, P. Weiss, and A.W. Burks, (eds.) 8 vols. (Cambridge, Mass.: Belknap Press, 1974), CP 1.158. 29 Nicholas Rescher, “Kant on the Limits and Prospects of Philosophy – Kant, Pragmatism, and the Metaphysics of Virtual Reality,” Kant-Studien 91, 3 (2000), 283-328.

Iconicity in Kant and Peirce

51

With the benefit of recent neurological and neurolinguistic findings, we are in a position to place Peirce’s iconicity in a larger, evolutionary picture in which Kant’s Anschauungen and Vorstellungen, reformulated in terms of iconic signs, can play their part in cognition and language. Any mental act is redefined by Peirce as a sign in the sense that it is an aliquid pro aliquo. This is at the heart of his influential definition of the sign of 1897. A sign, or representamen, is something that stands for something in some respect or capacity. It addresses somebody, that is, creates in the mind of that person an equivalent sign, or perhaps a more developed sign. The sign which it creates I call interpretant of the first sign. The sign stands for something, its object. It stands for that object, not in all respects, but in reference to a sort of idea, which I have sometimes called the ground of the representamen.30

The representamen corresponds to Saussure’s signifier in that it is arbitrary and has no meaning by itself. As such, anything whatsoever can be declared a representamen, though the representamens of natural languages are attached to their meanings by a community of speakers, and so are stabilised at least for a certain time. We might call them “semi-rigid designators”.31 Meaning occurs when the representamen is connected with something else, that is, another sign, in someone’s Vorstellung. This “more developed sign” Peirce calls the interpretant, (CP 2.228) roughly equivalent to Saussure’s signified. Contrary to the Saussurean picture however, the representamen “mediates between the interpretant sign and its object”. (CP 8.343) While in structuralist linguistics, objects and referents are subsumed within an idealised system of language and so play no methodological role in the dyadic linguistic sign, Peirce offers a much broader explanation. Peirce’s triadic definition allows for the perceptual grasp of objects in the world and their processing via sign relations as meaningful within a semiotic community. Not only is the object therefore a necessary component of the sign itself, Peirce also builds into his sign the rationale for or purpose of any particular use of a sign, its ground. By this Peirce guides us to think in terms of “a sort of idea” that we express by using a sign in a certain way. I read this as including the kind and degree of schematization we subject something in sign practice. As such, Peirce’s ground determines the selection of traits, the degree of equalisation, and the extent of foregrounding of the features of an object, 30 31

Peirce, Collected Papers, CP 2.228. Saul Kripke, Naming and Necessity (Oxford: Basil Blackwell, 1980).

52

Chapter Two

person, event, etc. that we wish to employ as a sign. Two important further additions to Peirce’s explanation of his conception of the sign must be kept in mind. One is the fact that “we cannot detect a logical end point” to any sequence of interpretants, such that every interpretive activity is theoretically infinite. The infinite chain of signs is “an endless series of representations”. The other addition concerns the permanence of iconicity in the definition of meaning, which is “nothing but the representation itself conceived as stripped of irrelevant clothing. But this clothing never can be completely stripped off; it is only changed for something more diaphanous”. (CP 1.339) Except in formal sign relations, which Peirce treats separately under the concept of the symbolic sign, iconicity in empirical concepts can never be entirely eliminated, no matter how highly abstract our schematization. Generalisation never amounts to formalisation. The empirical concept and its variants in Vorstellung can be expressed in terms of the Peircean sign as either an iconic or an indexical sign. The former is defined by way of resemblance relations, the latter indirectly by way of cause and effect. A weathercock can be viewed as an indexical sign in the sense that it tells not about itself but functions as an indication of something else, whereby the weather cock and the wind are linked via a causal nexus. Nevertheless, without the employment of iconicity, Peirce’s indexical sign could not function as it does. It is indirectly iconic. Our primary interest here, though, is in the iconic sign, which “is an “affair of suchness only,” whereby “suchness” refers to at least a minimum of similarity between what is represented and its representation. (CP 5.74) “The icon of a centaur represents a centaur “by virtue of its shape”,” no matter “whether there be a centaur or not”. (CP 5.73) What matters is the resemblance relation between various Vorstellungen triggered by the linguistic representamen. Peirce calls a sign an icon if it stands for something by merely resembling it. (EP 1.226) Alternatively put, “an icon is a sign fit to be used as such because it possesses the quality signified”. (EP 2.307) As such, an icon is “fit to be a substitute for anything that it is like”. (EP 2.273) Iconicity, however, is not restricted to the relation of a sign to its object; it also appears as syntactic iconicity when we note a resemblance relation between linguistic syntax and the kind of perceptual protosyntax we are forced to stipulate as its condition. Likewise, quantitative iconicity shows itself when the conceptual complexity of something is matched by the formal complexity of its representamen or signifier. In all this it is crucial to realise that the icon is above all a mental operation, a process in and by Vorstellung without which the link between object, representamen, and interpretant cannot occur. It is in this light that

Iconicity in Kant and Peirce

53

we should read the following remark, reminiscent of Kant. The icon, Peirce writes, “is of the nature of an appearance, and as such, strictly speaking, exists only in consciousness”. As an example, Peirce refers to the “geometrical diagram,” which cannot be found in nature. (CP 4.447) Even in such near formal constructs, both Peirce and Kant insist that we must draw on iconicity. As quoted above, Kant insists on mental iconicity, for “we cannot think a line without drawing it in thought”.32 That this kind of iconic thought cannot be of Fregean purity should be self-evident. Not satisfied with this basic description of the icon, Peirce also aimed at a rough differentiation between degrees of iconic schematization. He does so under the heading of the hypoicon which subsumes the “image” as an icon the sign relation of which is based on imitation and resemblance, from the “diagram” which is based on abstraction and idealisation, and from “metaphor” as combining resemblance relation and displacement, or “parallelism in something else”. We now have three kinds of similarity relations: mirroring, abstraction, and displacement. 33 Given that the indexical sign is likewise characterised by iconicity, albeit indirectly, the only sign without any trace of iconicity is the symbolic sign, the entirely arbitrary stand-in for whatever we wish to signify formally. This allows us to distinguish sharply between the unmotivated or arbitrary linguistic signifier from the motivated signified. I leave the bulk of Peirce’s sign theory to one side and foreground iconicity. I do this because the method of the stripping of iconic content has favoured a description of natural language which violates its very “naturalness,” that is, the degree to which it is relies on imaginability. In this respect, Peirce’s insistence on iconicity is not only at the heart of his semiotics but also crucial to the wholehearted commitment made here to the function of imaginability in language. After all, “every assertion,” as Peirce writes, “must contain an icon or a set of icons, or else must contain signs whose meaning is only explicable by icons”. (CP 1.158) If he is right, then iconicity would be a necessary feature of language, something that must have remained anchored somewhere in natural language. Furthermore, since language cannot be conceived of as separate from a speech community, linguistic iconicity must have its parallels in the Vorstellungswelt, the imagined world, of a culture. Language, as does “the very origin of the conception of reality shows that this conception 32

Kant, CPR B154. Charles Sanders Peirce, Writings of Charles S. Peirce: A Chronological Edition, vol. 2, 1867-71 (Bloomington: Indiana University Press, 1984), EP 2.273. 33

54

Chapter Two

essentially involves the notion of a community”. (CP 5.311) In terms of natural language, the dominant sign system for the communication of reality, and what has been said so far in this chapter, this implies that the iconicity of a Vorstellungswelt must be closely related to linguistic schematization. As this chapter has tried to demonstrate, both Kant and Peirce have made significant and lasting contributions to the refinement of Locke’s attempt in the Essay Concerning Human Understanding at the theorization of what happens when humans cognize and communicate about the world, including themselves, by way of language. They have done so by addressing the question of mental representations at one remove from Locke’s reflection theory of ideas by the schematic and semiotic rather than percept-like nature of Vorstellungen. Kant’s notion of nonsubjectivist schematization and Peirce’s systemic theory of iconicity, and in particular the hypoicon, are important stepping stones in their attempt at de-subjectivizing Locke’s semantic privacy while at the same not giving up on the fundamentals of imaginability as a prerequisite for an intersubjective mentalism. The next chapter will resume the question of mental iconicity as a necessary ingredient of linguistic meaning by a critical overview of the phenomenological semantics of Edmund Husserl.

CHAPTER THREE MEANING AND IMAGINABILITY IN HUSSERL

Husserl’s ideality of meaning Before we are in a position to ask what role, if any, mental projections play in Husserl’s conception of language we must address his preoccupation early in his career with meaning ideality. What is “ideality” in Husserl’s description of linguistic meaning? Unlike Frege’s account, Husserl’s is a complex web of a large number of components as it first appears in the Logical Investigations. I sum these up as Husserl’s “essential, primary distinctions”. 1 First we have expressions (syntactic ordered sequences of signifiers) uttered and experienced as an “articulate sound complex” which is to be distinguished, though not separated, from its associated “mental states”, as well as meaning as a Platonic species locatable in the “timeless realm of ideas”. 2 For Husserl, meanings are “ideal unities” characterized by “identity”. 3 They are starre Einheiten (rigid unities). (LI I §28) Expressions, in contrast, are a “part of speech” with or without “communicative intent”. (LI I §5) If “communicative”, they are “always bound up with” an “indicative relation”. (§2) For simplicity’s sake, I leave aside expressions that are unsinnig or nonsensical and widersinning or self-contradictory, the former being declared by Husserl as not well formed, the latter as possibly well-formed 1 Edmund Husserl, Logical Investigations, trans. J.N. Findlay (London: Routledge and Kegan Paul, 1973); LI I §10. Other works to be discussed in this chapter are his Experience and Judgment: Investigations in a Genealogy of Logic, trans. J.S. Churchill and K. Ameriks (London: Routledge and Kegan Paul, 1973); The Crisis of European Sciences and Transcendental Phenomenology, trans. David Carr (Evanston: Northwestern University Press, 1970); Formal and Transcendental Logic, trans. Dorian Cairns (The Hague: Martinus Nijhoff, 1969). 2 Cf. Kevin Mulligan’s distinction of six kinds of meaning in Husserl in “Meaning Something and Meanings,” in Themes from Early Analytic Philosophy: Essays in Honour of Wolfgang Künne, ed. Benjamin Schnieder and Moritz Schulz (Amsterdam: Rodopi, 2011), 273. 3 Husserl, Logical Investigations, 149.

56

Chapter Three

but lacking in self-consistency, such as “round squares”. When language is used either in communication or in internal soliloquy, meaning is generated in such a way that it functions as “a class of universal objects” (LI I §33) or an ideal species under which the various necessary mental acts performed by speakers and hearers are particulars united to guarantee semantic identity. (e.g., in “quadratic remainder”, §11) Husserl distinguishes his Platonic ideality of meaning from “meaning” in the restricted sense of merely “indicating”. (§5) The ideality of linguistic meaning as species is part of Husserl’s larger frame of inquiry into how “the real world” is “spread out before us”, as well as “all ‘ideal’ worlds” and becomes “intelligible as transcendental correlates”. 4 Husserl’s conviction of linguistic ideality is rooted in his notion of “essence” as “an ideal conceptual totality of possible particulars to which it can be related through a thought that is both eidetic and universal”.5 Now and then in the Logical Investigation, however, Husserl deviates from identifying linguistic meanings with their formal cousins, for example when he writes, “the intentional essence of the act of intuition gets more or less perfectly fitted into the semantic essence of the act of expression”. (LI II, §8) Qualms about his radically abstractive semantics also surface in his qualification that meaning ideality is not so much “normative” as “practical”. (LI I §32) Yet when identity is stipulated via “ideal objects” which, for Husserl, “exist genuinely”, it is illustrated by “the number 2, the quality of red, the principle of contradiction”. (LII §8) Note here how the “fuzzy” ideality of “the quality of red” is embedded in arithmetic and logical expressions. Several kinds of mental acts are distinguished in Husserl’s account of natural language. “What marks off an expression” from mere word sounds is the “meaning-intention” in which the expression “undergoes an essential phenomenal modification”. (LI I §10) This “act of meaning” (Bedeuten) in turn is differentiated from “meaning” itself (Bedeutung). (LI VI §8) The “act-complexes” of “sense-conferring acts” and “sense-giving acts” are further to be set apart from the acts of the “intimation of mental states” both in “speaking and reception”. The “intimating function” consists in the way “inner experiences” are conveyed to the hearer. Husserl uses “intimation” in two senses: a narrow one as “acts which impart sense” and a wider one as “all acts that a hearer may introject into a speaker on the 4

Edmund Husserl, Ideas: General Introduction to Pure Phenomenology, trans. W.R. Boyce Gibson. London: Collier-Macmillan, 1969), Preface to the English Edition, 20. 5 Husserl, Ideas §13.

Meaning and Imaginability in Husserl

57

basis of what he says”. (LI I §7) This is an early instance of his later notion of appresentation or co-presenting, mental acts that fill out by way of imaginability what is not immediately given. Yet “intimation” is not to be confused with “indication”. By “indication” Husserl means strictly the mental act conveyed by an expression “in its relation to the object it stands for or that it is to signify”. (§2) In “communicative speech” all expressions “function as indications”. (§7) On the side of semantic uptake, “the hearer intuitively takes the speaker to be a person who is expressing this or that” or “perceives him as such”. (§7) Intuition here likewise requires acts of appresenting. It is already clear at this point that in the kind of semantic picture Husserl is constructing, linguistic comprehension could not be argued without a firm commitment to Vorstellung and Vorstellbarkeit, or imaginability, or perhaps “intelligibility” in Sokolowski’s sense.6 Nor is this mental activity restricted to the hearer of an expression. At its heart lies reciprocity. “Mutual understanding”, says Husserl, “demands a certain correlation among the mental acts mutually unfolded in intimation and in the receipt of such intimation, but not at all their exact resemblance”. (§7; my emphasis) Here already, a fundamental dilemma reveals itself in Husserl’s account of natural language, a dilemma that is to remain unresolved, though by no means unresolvable, to the end of his career: the competing notions of resemblance relations and semantic exactitude. The principle of the reciprocity of mutual mental acts is resumed in Husserl’s distinction between “meaning-conferring acts” and “meaning-fulfilling acts”, the former being defined as “meaning intentions”, the latter as a necessary receptive condition of comprehension. (§9) As Wittgenstein is to observe half a century later, in language, “an expectation and its fulfilment make contact”.7 But Husserl is much more specific on the kind of interaction he observes between intention and uptake in reconstructive acts in Vorstellung, producing as it does an “intimately fused unity” of “meaning-intention” and “meaning fulfilment”. Because of his emphasis on Vorstellung, Husserl rejects the formulation of an “expression expresses its meaning”. (LI I §9) Expressions by themselves do nothing. Rather, what is required for meaning to occur is the mental act of 6

Robert Sokolowski, Phenomenology of the Human Person (Cambridge: Cambridge University Press, 2008), 167. Cf. also his Introduction to Phenomenology (Cambridge: Cambridge University Press, 2000); as well as “The Logic of Parts and Wholes in Husserl’s Investigations,” in Readings on Edmund Husserl’s Logical Investigations, ed. J.N. Mohanty (The Hague: Martinus Nijhoff, 1977), 94111; and Husserlian Meditations: How Words Present Things (Evanston: Northwestern University Press, 1974). 7 Wittgenstein, PI §445.

58

Chapter Three

“understanding” which “shines through the expression”, lending it “meaning and thereby relation to objects”. (LI I §18) Important outcomes of this part of Husserl’s analysis from the perspective taken in this chapter are his notions of “exclusive directionality” and the “steering of attention”. “The function of a word”, writes Husserl, “is to awaken a sense-conferring act” in order “to point to what is intended”, an “intuitive fulfilment” and “to guide our interest exclusively in this direction”. In the same section he speaks of “the steering of attention” from “name to the thing named”, from “expression” to “what is expressed”. (LI I §10) Linguistic directionality, then, is not merely suggestive and haphazard but compelling. As I shall argue later, the concept of “exclusive directionality”, defined by Husserl also more sharply as “determinate direction to objects”, can be used to tighten the relation between speech community and meaning-fulfilment. (LI I §13) I have done so elsewhere under the notion of “the linguistic linkage compulsion”.8 So far we have dealt with linguistic expression, intimation, indication, meaning-conferring and meaning-fulfilling acts, the steering of attention and exclusive directionality and the ideality of meaning. What is still missing is extensionality, in the Leibnizian sense of extension in contrast to intension, the former pointing beyond the expression to aboutness, reference, and voice, the latter to its definitional coherence or, in Husserlian parlance, perhaps its “noematic nucleus”. Husserl addresses extensionality in two ways. It appears in the matter of a meaning act in contrast to its quality. It also appears in what an expression says of something, by way of “the objective correlate of an expression”, defined as “the objective correlate meant by a meaning” in distinction from any actual object. (LI I §12) Here, exclusive directionality results in our attention being pointed to something outside the mental acts which we cannot but perform in the event of meaning, that is, what an expression is about or, in more recent phrasing, aboutness.9 The point Husserl is stressing here is that “the object never coincides with the meaning”. (§12) Husserl does not, however, distinguish between aboutness and reference, the latter being defined as the directionality of an expression “to certain objects”. As such it appears in the well-trodden examples of “The victor of Jena” which is about, and can be used to refer to, the same object as “the vanquished of Waterloo”, (§12) an illustration that may very well be a Fregean residue in Husserl’s 8

Horst Ruthrof, “Semantics of Imaginability – Vorstellungssemantik: 13 Theses,” Review of Contemporary Philosophy, 10, (2011), 165-183; 168. 9 Andrew Woodfield, “Intentionality,” in Concise Encyclopedia of Philosophy of Language, ed. Peter Lamarque (Oxford: Pergamon, 1997), 55-59; 55.

Meaning and Imaginability in Husserl

59

theorization, which otherwise deviates markedly from Fregean semantics. It is preferable, I think, to run with P.F. Strawson on reference as the “referring use” of expressions10 and the detailed analysis of reference in Varieties of Reference by Gareth Evans.11The distinction between aboutness and reference as a special kind of language use has the explanatory advantage of showing that the former is part of meaning and so semiotically homogeneous as intentional object, while the linking of meaning and objects outside the meaning acts is semiotically heterogeneous. In this sense we can rephrase Husserl’s observation that “statements” can be “different in meaning while referring to the same object” (§12) by saying that statements with different meanings can have the same aboutness and can also be used to refer to the same objectivity (Gegenständlichkeit). This of course prevents us from going along with Husserl’s formulation “to use an expression significantly and to refer expressively to an object (to form a presentation of it) are one and the same.” (§15) If Peter Simons is right in saying that “Husserl pursued philosophy of language not for its own sake but mainly to support his conception of logic” and to “refute psychologism”,12 then this would explain why radical forms of abstraction inform his early description of natural language. What drives Husserl towards the idealization of meaning, backed up by formal illustrations, is his assumption of the necessity of semantic identity as “identical sense”. (LI VI, §4) Where he comes from, as do many of his contemporaries in this quest – mathematics – identity is an essential component of the axiomatic foundations of the sign system. Identity is a sine qua non of stipulated formal signs, although even identity in this sense has not remained uncontested. 13 Husserl’s affinity with formal 10

P.F. Strawson, “On Referring,” Mind 59 (1950), 320-344. Gareth Evans, The Varieties of Reference, ed. John MacDowell (Oxford: Clarendon Press, 1982). 12 Peter Simons, “Meaning and Language,” in The Cambridge Companion to Husserl, ed. Barry Smith and David Woodruff Smith (Cambridge: Cambridge University Press, 1995), 106-137; 106. 13 Guillermo Haddock, “Identity Statements in the Semantics of Sense and Reference” in Husserl or Frege? Meaning, Objectivity and Mathematics, ed. G.E.R. Haddock and C.O. Hill (Chicago: Open Court, 2000), 41-51; 42ff.; cf. also his “To be a Fregean or to Be a Husserlian: That Is the Question for Platonists,” in Husserl or Frege, 199- 220; and “On Husserl’s Distinction Between State of Affairs (Sachverhalt) and Situation of Affairs (Sachlage)”, in Husserl or Frege?, 253-261; cf. also J.N. Mohanty, “Husserl’s Thesis of the Ideality of Meaning,” Readings on Husserl’s Logical Investigations, 76-82. 11

60

Chapter Three

identity and meaning ideality pervades the Logical Investigations, where “pure logic” remains his point of orientation because it “is exclusively concerned with the ideal unities that we here call “meanings”“. (LI I §29) This expectation is transferred to “the essence of meaning” in natural language, “not in the meaning-conferring experience, but in its “content”, the single, self-identical intentional unity set over against the dispersed multiplicity of actual and possible experiences of speakers”. (§30) Either, then, we are dealing with “self-identical” singularity or with “dispersed multiplicity”. What is missing in the early Husserl is the in-between of graded typifications, a notion that neither requires identity nor dissolves into private singularities. In any case, to recognize identity as indispensible for logic and mathematics is one thing; it is quite another to demand it for linguistic meanings. Granted, identity is a powerful lure. Wittgenstein criticized this move in his Philosophical Investigations when he notes that “we are dazzled by the ideal” which “absorbs us” to such a degree that it “’must’ be found in reality … we think we already see it there” and so “we must find … the ideal … in our actual language”.14 To Husserl’s credit, in spite of his flirtation with meaning as a Platonic species and the lure of the formal, he never gave up on realist mentalism as an essential part of a comprehensive theorization of language. (LI II §11) The problem for mentalism, though, as we have seen in Locke’s semantics, is to render it intersubjective so it can be reconciled with the public nature of language. As to meaning identity, Husserl was and still is in good company. From Frege to Montague and further to hyperintensional logic, “identity” both in its numerical and qualitative senses has been part of the staple diet of the philosophy of language. When it is imposed on the description of language we always note these three moves: (1) The collapsing of natural language sense into formal sense; (2) the elimination of imaginability from meaning; and (3) the forgetting of linguistic modality, especially its implicit variety. Unlike Frege, Husserl distinguishes identity from alikeness. Identity in Husserl is conceivable as a characteristic of a species, while “alikeness” is reserved for things, objects, or members. Yet, “identity is wholly indefinable, whereas “alikeness” is”. (LI II, §3; my emphasis) Worse, identity as a result of Platonic idealization proves not only indefinable but indemonstrable. This seems to me to be the main reason why Husserl invariably turns to formal examples when he wants to illustrate the ideality of linguistic meaning. A further, serious complication in Husserl’s analysis arises in that having declared meaning an ideal species he has blocked its 14

Wittgenstein, PI §100ff.

Meaning and Imaginability in Husserl

61

description in terms of degrees of generalisation. Meaning as ideal Platonic species prohibits what Husserl is to theorize in Ideas under the heading of generalisation on the ladder of specification and despecification, leaving formalisation or de-materialisation to take the place of ideality, where identity is axiomatic.15 In actuality, what can be said in a language is not dictated by ideality but rather by two major constraints: the world of Anschauung and its myriad variations in Vorstellung, on the one hand, and, on the other, community sanctioned idiomatic phrasing. If we take the Lebenswelt as our starting point, it is not the ideality of meaning or grammar but much more likely a certain degree of schematization in relation to their existential ground that determines their role in language. As Husserl came to see in his late writings, here too the a priori is the Lebenswelt.

Formal demonstration and semantic vagueness As we have already noted, when identity and ideality are to be demonstrated in the Logical Investigations, Husserl does not hesitate to turn to formalisation. From the beginning in the Logical Investigations, the commitment to “transcendental phenomenological idealism” in relation to language is more easily demonstrated by recourse to formal entities than by natural language examples. In veering “from the real relation of acts to the ideal relation of their objects or contents”, he does so not only with reference to geometrical figures and other formal notions where “something objective and ideal is brought to expression” as an illustration, but also transfers the formal to natural language itself. “The same holds”, Husserl writes, “of the other parts of our statements, even of such as do not have the form of propositions”. (LI §11) Thus has he closed the all-important gap between language as “universal medium” and “calculus”,16 between natural language semantics and his “pure morphology of significations” (reine Formenlehre der Bedeutungen). “The three perpendiculars of a triangle intersect in a point”. Indeed, here we have, as Husserl says, “an identity in the strict sense, one and the same geometrical truth”. (LI I, §11) Yet the formality of geometry cannot be equated with what goes on in natural language; less so the laws of arithmetic which “tell us nothing about what is real”.17 So if Husserl is correct in saying that “temporality is 15

Husserl, Ideas, §13. Martin Kusch, Language as Calculus and Universal Medium (Dordrecht: Kluwer, 1989); cf. also his “Husserl and Heidegger on Meaning,” Synthese 77 (1988), 99-127. 17 Logical Investigations, Prolegomena to Pure Logic, §46. 16

62

Chapter Three

a sufficient mark of reality”, (LI II §8) then Frege’s equivocation of arithmetic and geometrical expressions with those of “evening star” and “morning star” is disallowed. And while there is still a minimum of aboutness in geometry, there is no modality. Consider the following example. “The transnationals are beyond the control of elected governments” spoken by a major shareholder of BP with the implicit modality of “Thank God”, after President Obama was pulled back by the Supreme Court, preventing him from putting an embargo on off-shore oil drilling. Pragmatically, the sentence is fundamentally different from the same sentence being uttered by an environmentalist scraping the oil of a seabird in the Gulf of Florida. Nor does the retreat from utterance meaning to sentence meaning resolve this problem. After all, sentences do not speak themselves. Husserl tries to decide the issue under the headings of “intimation” and “essentially occasional expressions”, to be discussed below. The point I am making here is that thou shalt recognize a theory by its exemplification. It highlights both a theory’s strengths and blind spots. Why the distinction between aboutness (of which reference is an application) and voice or modality in the broad sense is so important in natural language is that there is no logical barrier to be discovered between where the “definite description” of the sentence ends and where its deictic modifications kick in. On this point Husserl’s critique of the Millean distinction between “denotation” and “connotation” is instructive. (LI I §16) For the analysis of natural language the distinction turns out to be no more than one between small and large dictionaries. Here, aboutness and modality are inextricably intertwined. Because of this lack of a demarcation we can say that “in natural language, modality typicality overrides propositional content”. Formal sign systems were invented precisely for the reason to prevent this relation and its ensuing fuzziness. The topic of modality as an essential feature of natural language will be taken up again later under the heading of “Essentially Occasional Sentences”. Suffice it here to say that the illustration of linguistic meaning by formal examples can capture only part of what is going on when we speak. Indeed, Husserl’s own definition of “logical meanings” as an “ideally closed set of general objects” disallows the analogy with formal stand-ins. (LI I §35) This much has been conceded even from within the ranks of analytical philosophy by some prominent philosophers, as for example Russell, Tarski, the later Wittgenstein and Quine.18

18

This point will be taken up once more in Chapter Four.

Meaning and Imaginability in Husserl

63

It may come as a bit of a surprise, then, that we find buried in the ideality and identity driven Logical Investigations that in principle Husserl had such insights a few decades earlier. “Most expressions used in ordinary life”, Husserl writes, “such as ‘tree’, ‘shrub’, ‘animal’, ‘plant’ etc., are vague”, their meaning “oriented towards types, only partially conceived with clearness and definiteness”. He deplores that “the haziness of such expressions” is not conducive to “definite identifications” because they reflect “phenomenal properties” which “shade continuously into one another, producing “fluid” relations and “vague borders”. (LI I §27; my emphasis) A fundamental tension then exists between Husserl’s preoccupation with meaning identity and his observation that the bulk of natural language expressions supports a fundamental vagueness. When he places “type” and “typification” at the centre of his investigation towards the end of his career he is encouraging us to resume his theme of semantic vagueness in order to revise his semantics accordingly.

Speech community: the missing component In Experience and Judgment it is the pre-predicative world experienced via passive synthesis that still supplies us with the rudiments of attribution and predication, the “passive protodoxa”, 19 a perspective that can be argued only by taking the social dimension into account. So I ask, where in the Logical Investigations is what Husserl is later to call “the community of life” which, as speech community, sustains language? 20 While Locke had granted the speech community a prominent place in his empiricist semantics, even if he was not able to forge the vital link between subjective meanings and communal control, Husserl’s semantic ideality made the speech community superfluous. The speech community has no stewardship over Platonic entities. And yet, without a speech community, there is no such thing as a language. The latter presupposes the former, while the former could not be called what it is without a language. Speech community and language evolve as one. The early Husserl escapes both speech community and its role in meaning by recourse to ideality and internal soliloquy. While the commitment to ideality can be found in many accounts of semantics, the notion of 19

Edmund Husserl, Experience and Judgment: Investigations in a Genealogy of Logic, trans. J.S. Churchill and K. Ameriks (London: Routledge and Kegan Paul, 1973), 65. 20 Edmund Husserl, The Crisis of European Sciences and Transcendental Phenomenology, trans. David Carr (Evanston: Northwestern University Press, 1970), §36.

64

Chapter Three

soliloquy is out of sync not only with the majority of natural language theorists but also with the principles of the Lebenswelt. So we ask “How solipsistic is soliloquy?” In Husserl’s solitary internal performance of language, “expressions function meaningfully” without serving “to indicate anything”. (LI I §1) Nor is there any “intimation”, (§8) because “the soliloquizing thinker ‘understands’ his words, and this understanding is simply his act of meaning them”. (§23 n1) If soliloquy were to be viewed as an instance of “private language” use, then Wittgenstein’s wellrehearsed arguments would be devastating to Husserl’s account, with soliloquy being no more than a de facto private use fully determined by communicable rules. Accordingly, we could ask whether we could do all the things in silent soliloquy that we can do publicly and do things that we could not do in public performance. The answer would show that such things as internal role playing are inescapable, including question and answers and especially the performance of modality. For once we perform linguistic irony and other enunciative spins, the fundamentally dialogic (or poly-logical) nature of language comes to the fore. In any case, in a philosophy which is to “satisfy the principle of freedom from presuppositions” the assumption of ideality at the price of omitting the role of the speech community must look odd.21 Once we reach the Crisis and Experience and Judgment where the Lebenswelt has been declared the ultimate a priori, intersubjectivity looms large and the speech community is able to play its appropriate role.

Two kinds of noema, modality and Sachlage Husserl’s definition of formalization as the empty end point on the axis of mental “dematerialization” in contrast to generalisation on the axis of despecification provides a useful tool for inquiring whether ideality and formality are not preferably replaced by degrees of generalisation. 22 Mirroring this distinction, Husserl allows for an argument in favour of the distinction between two kinds of noemata, the formal noema, which remains congruent with Husserl’s demand for meaning identity in the Logical Investigations, and a non-formal noema which anticipates his later notion of “non-essential typifications”. 23 Such a reading of Husserl’s overall endeavour is borne out by the fact that as early as in the Ideas the noema is presented as remaining linked to intuition and so cannot be 21

Logical Investigations II, Introduction §7. Husserl, Ideas, §13. 23 Husserl, Experience and Judgment, §83. 22

Meaning and Imaginability in Husserl

65

regarded as “empty” as Frege’s Sinn. In agreement with Robert Sokolowski (2000) and Lenore Langsdorf (1984), to name just two Husserlian scholars, I think that this amounts to a strong rejection of reading the noema as formal, as it has been influentially argued by Dag Føllesdal.24 Nor does Husserl’s description of the internal structure of the noema alter this fundamental distinction.25 For our purposes, the sharp distinction between two kinds of noema is essential, bearing in mind that only non-formal noemata have the capacity to transfer resemblance relations from the nonverbal aspects of the Lebenswelt into natural language, its dominant sign system. This is so on the proviso that we understand that neither noemata themselves nor their linguistic signifiers are agents; it is the speakers of a language who have been trained to accomplish such a transfer by means of expressions. They do so under the gun, as it were, of the “linguistic linkage compulsion”. That is to say, at the moment of comprehending an expression, Adolf Reinach’s “punctual act of meaning,”26 speakers are not able to substitute for their habitual meaningacts any alternative, private ones. If this is so, as I have argued elsewhere, then meanings must be indirectly public.27 Husserl’s emphasis on mental acts in the performance of linguistic meaning, then, remains indispensable, though it cannot be argued via subjective ideality. In Logical Investigations Husserl draws another influential distinction, spawned as it probably did such notions as Russell’s “ego-centric particulars” and Adolf Reinach’s speech-act phenomenology of 1913, an important precursor of Wittgenstein’s “multiplicity of the tools in language,” 28 as well as its uptake in the speech-act theories of John Langshaw Austin and John R. Searle. I am referring here to Husserl’s definition of “essentially occasional expressions”. As meaning depends on “occasion, speaker and situation” and if it displays “universally operative 24

Dagfinn Føllesdal, “Husserl on Evidence and Justification” in Phenomenology: Critical Concepts in Philosophy, v. 1, ed. Dermott Moran and Lester E. Embree (London: Routledge, 2004); “Noema and Meaning in Husserl”, Phenomenological and Philosophical Research 50 (1990), 263-271; “Husserl’s Notion of Noema”, The Journal of Philosophy 66 (1969), 680-687; cf. also David W. Smith and Ronald McIntyre, Husserl on Intentionality ( Dordrecht: Reidel, 1982). 25 Cf. Rudolf Bernet, “Husserl’s Begriff des Noema” in Husserl-Ausgabe und Husserl-Forschung, ed. Samuel Ijselling (Dordrecht: Kluwer, 1990), 61-80. 26 Adolf Reinach, “Zur Theorie des Negativen Urteils,” in Sämtliche Werke, v.1, ed. Karl Schuhmann and Barry Smith (Munich: Philosophia, 1989), 95-140; 103. 27 Horst Ruthrof, “Metasemantics and Imaginability,” Language Sciences 35 (2013), 20-31; 25. 28 Wittgenstein, PI §23.

66

Chapter Three

indication” such as I, you, etc. (LI I §26) Using his distinction between “act character” and “ideally unified meaning”, Husserl reasons that while the former differs from case to case, the sense of the sentence “should remain identical”. (§31) In Husserl’s own terms, one could agree that its “indicated meaning” does not change, while its “intimated meaning” introduces a principle of modification, part of linguistic modality.29 In this early work Husserl still insists that such modifications do not affect the meaning itself but only “the act of meaning”. (§28) Husserl’s “wesentlich okkasionelle Ausdrücke” have been well rehearsed by Aaron Gurwitsch in “Outlines of a Theory of “Essentially Occasional Expressions”, which I will briefly sum up in order to argue a radical correction, namely that in natural language all expressions are essentially occasional as a result of implicit deixis carried by voice. As Gurwitch sees it, “a theory of such expressions must fulfil two requirements: (1) it must account for their essential ambiguity and (2) it must take the circumstances of their occurrence into account”.30 Gurwitch resolves the first requirement by saying that meaning is often specified by way of perception or its equivalents in Vorstellung. (119) Yet, since “the meaningfulness of essentially occasional expressions” according to Husserl, “does not derive from any specifying act but resides entirely in these expressions themselves as bearers of meaning functions”, perception and other mental representations can only be viewed as “meaning-specifying” but not as “meaning-bestowing”. (120) Yet it is only once the meaning-specifying circumstances of utterance are taken into account that the expressions themselves can be disambiguated. After all, says Gurwitch, “the singular status of the essentially occasional expression derives from the particular organizational aspect with which the perceptual world presents itself”. (121) In other words, modality affects aboutness such that it in natural language the latter cannot be established on its own. To do so would make language into something it is not. My main objection here is that the special status which Husserl bestows on “essentially occasional expressions” rests on the assumption that linguistic 29 Meaning shifts as a result of “particular occasions” are recognized for example by Paul Grice as “fundamental”. As he says in “Meaning”, “I am sympathetic to this more radical criticism, though I am aware that the point is controversial”, in Studies in the Way of Words (Cambridge, Mass.: Harvard University Press, 1989), 213-223; 217. 30 Aaron Gurwitch, “‘Outline of a Theory of Essentially Occasional Expressions,” in Edmund Husserl’s Logical Investigations, ed. N.J. Mohanty, 112-127; 117.

Meaning and Imaginability in Husserl

67

modality occurs only when it is explicitly marked. To be sure, Husserl is not alone in holding this view. In linguistics, it is common practice to deal with such expressions under the notion of “deixis” as marked reference to speaker and speech situation. Only when it is signalled in a sentence are we supposed to be aware of its “enunciative modalities”, to use Foucault’s broader term. 31 Yet there is no language usage whatsoever without speakers and utterance situations, even if they are not specifically foregrounded. And without use, language does not exist as language. When it does exist in a social setting, all language carries, at least for the native speaker, the manner of its utterance as voice. This we could call the fundamental modality of natural language or, in linguistic terms, its implicit deixis. All expressions of natural language, then, must be comprehended in terms of the double construction by the user of their aboutness and voice. One could say that there is no meaning without spin. Husserl himself provides encouragement for this kind of argument when he distinguishes Sachverhalt (state of affairs) from Sachlage (situation of affairs), the former being grasped as what an expression is about, the latter requiring interpretive reconstruction of the kind of situation within which a particular Sachverhalt has been observed.32 An alternative reading is given by Guillermo Haddock who distinguishes “states of affairs” as the “referents of statements” from “situations of affairs as a sort of referential basis of states of affairs”.33 Husserl’s important distinction strikes me as anticipating the kind of work Paul Grice introduced under the heading of “presuppositions”.34 What sets Husserl’s enterprise sharply apart from Grice is that it involves nonverbal imaginability as variation of perception and as a necessary feature of the interpretive process, whereas Grice conceives “presuppositions” within the narrower frame of propositional logic. What Husserl allows us to do is to say that if the presupposed Sachlage changes, so must its Sachverhalt, as well as meaning. As our examples suggest, what is characteristic of essentially occasional expressions is part and parcel of every linguistic expression, suggestive of a relevant Sachlage, whether its modality is marked or not. Having argued this way, however, we now have pushed the distinction Husserl and linguists are making down to a more specific level 31

Michel Foucault, The Archaeology of Knowledge (London: Tavistock, 1986), 55ff. 32 Husserl, Experience and Judgment, §59. 33 Guillermo Haddock, “To be a Fregean or to Be a Husserlian: That Is the Question for Platonists,” in Husserl or Frege, 199- 220; 208. 34 For a reading of Grice’s presuppositions cf. Chapter Nine.

68

Chapter Three

where we must separate explicit from implicit deixis, or marked versus unmarked modality. On a closer reading of the Logical Investigations, though, we discover that in a way Husserl has anticipated my criticisms when he says in the “Foreword” to the second edition, that “in strictness, all empirical predications belong” to “occasional meanings”. Here again Husserl himself provides the means with which to repair his early language semantics.

Resemblance relations and degrees of schematization If semantic ideality and formal identity can only be stipulated but not demonstrated for meanings in natural language, where in Husserl’s oeuvre do we go for a more satisfactory explanation? If language indeed inherited resemblance relations about the world and the way they were communicated in the pre-predicative community as speech modality, our dominant discourses in linguistics and the philosophy of language do not handle these two features well. Let me emphasize here the merit of Husserl’s tightrope act of not letting go of the mental acts he saw as necessary components of language while at the same time avoiding what has been called Locke’s psychologism. We are facing three distinct solutions: (1) to declare meanings a Platonic ideal species or some other kind of universal entity; (2) to make them formal, that is syntactic relations governed axiomatically by definitions; and (3) to socialize them. As I will show, Husserl in the end felt compelled by his own investigations to choose the third option without, however, being able to rewrite his semantics accordingly. For an argument in favour of resemblance relations in language Husserl himself provides a number of footholds, such as certain mental acts which he says we cannot but perform when we use language: the transformation of perception into non-formal noemata, profiles, Anschauung, appresentation, Vorstellung, “differences in vitality”, 35 “analogical representation”, (LI VI §21) retentions and protentions, exclusive directionality, “gradations of fullness”, “richness”, “liveliness”, “vivacity”, “Realitätsgehalt”, (LI VI §23), “combinatory forms” (LI III §21), “abstractive percept” (LI VI §48), “semblance acts”, 36 habitual “circles of resemblance”, “differing resemblances”, (LI II §36) and above all the a priori Lebenswelt and its typifications. As recipes for social action, 35

Edmund Husserl, Phantasy, Image Consciousness, and Memory (1898-1925), trans. John B. Brough, Collected Works, ed. Rudolf Bernet, v. XI (Dordrecht: Springer, 2005), 373. 36 Husserl, Phantasy, Image Consciousness, and Memory, 710.

Meaning and Imaginability in Husserl

69

in Schutz’s sense, typifications do not require the domain of ideality, as do triangles, but can be argued as generalisations at differing degrees of abstraction without ever altogether losing their mental material content. This is uncontroversial with respect to perceptual expressions. Here, “the whole sense of the statement finds fulfilment through our underlying percept”, the whole statement giving “utterance to our percept”. (LI VI §40) At least in such cases, resemblance relations are preserved in linguistic meaning, even if “we cannot suppose that perception is the act in which the sense of a perceptual statement, its expressive intention, is achieved”. (LI VI §4) In another passage in the Logical Investigations, Husserl deviates from semantic ideality, noting that the “generality of words” belongs “to an endless array of possible intuitions”. (LI VI §7) One may wonder why Husserl did not stumble more fruitfully over the discrepancy he had created between the relation of “meaning-intention” and “corresponding intuition” on the one hand and meaning ideality, on the other. To his credit, in an “Addendum” Husserl writes, I cannot here suppress a difficulty connected with the otherwise illuminating notion of the unity of identity or recognition … Shall we say that an emphasis of attention decides the matter? Or ought we not rather grant that there is not here a fully constituted act of identification: the nucleus of this act, the connective union of significant intention and corresponding intuition is really present, but it “represents” no objectifying interpretation. (Auffassung)37

Another promising avenue to explore from the Logical Investigations to Experience and Judgment is Husserl’s concept of “profiles”. In perception, material objects are realized in consciousness as a synthesis of visual, tactile, gustatory, and other heterosemiotic profiles. As Sokolowski sums it up, “a profile is a moment of a thing-synthesis”.38 In this sense “profiles” belong to Husserl’s theorisation of “parts and wholes” where he distinguishes two kinds of parts, “pieces” and “moments”. Profiles are instances of the latter. Though in the Logical Investigations Husserl mentions profiles together with “aspects” of an object, the concept of profile seems to me to emphasize more the outline of things rather than their adumbrational sides. As such, the notion of profile is reminiscent of Kant’s transcendental schema of a dog: “The concept of a dog is a rule according to which my imagination is able to sketch in a general manner 37

Edmund Husserl, Logical Investigations, VI §8; Addendum, 697. Robert Sokolowski, “The Logic of Parts and Wholes in Husserl’s Investigations,“ in Readings on Edmund Husserl’s Logical Investigations, ed. J.N. Mohanty (The Hague: Martinus Nijhoff, 1977), 94-111; 102. 38

70

Chapter Three

the shape of a certain four-legged animal, without being restricted to any singular and specific shape, given to me by experience or any possible picture that I could represent in concreto”. (CPR A141/B180) Certainly, the doctrine of “profiles” in its relation to the notion of schematization offers another way out of Husserl’s early commitment to semantic ideality and formalisation in the sense that as generalised abstractions, that is, despecified perceptual grasp, profiles can be held in Vorstellung as schemata bearing reduced and yet still sufficient resemblance relations for recognition “in a flash”, to use Wittgenstein’s phrase once more. 39 A similar comparison between Husserl and Kant has been drawn by Dieter Lohmar in “Husserl’s Type and Kant’s Schemata” resulting in the conclusion that apart from their different genesis the two notions are “functionally almost identical”.40 If linguistic meanings are to carry social and historical ingredients, as I think they must in contrast to formal signs, then we require not only the notion of schematization in some shape, but also degrees of schematization. And if we follow Lohmar and take schematizations in the sense of “typifications”, then degrees of generalisation or de-specification can be argued within Husserl’s own parameters.41 We can furthermore draw on Husserl’s concept of “grades of evidence” in the Cartesian Mediations42 where he distinguishes “apodictic” from “adequate” and “certain” evidence. Since neither “absolute indubitability” nor “idealization” without lacunae of indeterminacy apply to linguistic meanings, we have to lower our sights to degrees of certainty of evidence according to which speakers of a language do not doubt that a signifier like “envious” directs us towards a certain mental disposition consisting in a certain range of mental materials regulated by its concept. At least in habitual speech, then, evidential certainty could be said to characterize relevant degrees of schematization. In interpretive and problematic meaning constructions, as described by Alfred Schutz in his last manuscript, Husserl’s tripartite division of “evidence” would have to be complemented by a more fluid notion of schematization. The use of natural language in hermetic poetry, 39

Wittgenstein, PI §§139; 191; 318f. Dieter Lohmar, “Husserl’s Type and Kant’s Schemata: Systematic Reasons for their Correlation or Identity,” in The New Husserl: A Critical Reader, ed. Donn Welton (Bloomington: Indiana University Press, 2003), 93-124; 93. 41 Husserl, Ideas, §13. 42 Edmund Husserl, Cartesian Meditations, trans. Dorian Cairns (The Hague: Martin Nijhoff, 1960), §6. 40

Meaning and Imaginability in Husserl

71

as for example in the late poems of Paul Celan, may serve as a case in point. What is missing in both the Kantian and Husserlian accounts is how semantic schematization functions in the relationship between speakers of a language and the speech community. Instead of seeing Husserlian “profiles” as idealized species or as schemata dictated by human biology, they should also be investigated as intersubjectively shared rules monitored by a community, facilitating approximate semantic agreements. Indeed, such rules may be shared to the extent of sameness as a result of “exclusive directionality” in the sharpened sense of “linguistic linkage compulsion” which rigorously curtails meanings in individual semantic events. The linkage of resemblance relations and language is hinted at by Husserl in his observation that “with each circle of resemblance a particular name is associatively linked, so that the inner reflection, together with its characteristic ‘angle’ of treatment, also fixes the general name”. (LI II §36) What is remarkable here is that in one sentence Husserl appears to intimate an entire non-essential approach to natural language capturing both linguistic aboutness (“inner reflection”) and modality (“angle” carried by voice). Followed to its logical conclusion, it would seem that here once more Husserl is pointing us towards a semantics of imaginability.

Nonessential typifications, Vorstellung, and appresentation By the time of the posthumous Experience and Judgment, Husserl has quite dramatically changed his description of the way we grasp the world by means of “typifications”, a change that has profound consequences for the revision of his earlier natural language semantics. Now he distinguishes between “essential” typifications as they function in science and “nonessential typifications” characteristic of understanding in the Lebenswelt.43 Both kinds of typification not only display differing levels of generality but, in agreement with distinctions drawn early in Ideas, are also marked as “formal” and “material” generalities, respectively. (EJ §85) For both, Husserl stipulates that the greater the extension of an expression, the higher the level of generality. What is central to our argument here is that “similarity … has a gradation”, such that “the levels of generality are conditioned by the degrees of likeness of the members of the extension”. (EJ §84) Here, then, we have degrees of resemblance relations built into non-formal typifications. I suggest that typifications understood in this way furnish precisely what is required to fit the bill also of meanings in 43

Husserl, Experience and Judgment, §83.

72

Chapter Three

natural language. Instead of meaning viewed as formal or ideal species, “complete likeness” is now no more than a limiting case, “the limit of similarity”. (EJ §84) Identity is now reserved for “essential” typification, as it should be, while natural language has its semantic anchor in similarity or resemblance relations generalised to appropriate degrees. As Natanson writes with reference to the Lebenswelt, “typification is the generic term for an abstractive process, one whose central accomplishment is the experience of the familiar”.44 To retain a sense of any such experience in language, we are bound to stipulate degrees of schematized resemblance relations as part of signifieds in the form of mental materials and their regulation by concepts. Only in this manner, I suggest, can the “Totalitätstypik” and “Habitualität” der Lebenswelt be rendered linguistic.45 In the Logical Investigations, Husserl raises the question, resumed prominently by Wittgenstein, whether the meaning process “necessarily involves … an act of fancy, or whether, on the other hand, mental imagery lies outside of the essence of an expression, and rather performs a fulfilling role, even if only of a partial, indirect or provisional character”. (LI I §10) “There is a lack of clarity about the role of imaginability (Vorstellbarkeit) in our investigation,” says Wittgenstein, “namely about the extent to which it ensures that a sentence (Satz) makes sense”. And elsewhere, “the language of imagining does not allow us to imagine anything senseless”. But while Wittgenstein relegates the contribution of the imagination to language to the incidental play on the “Vorstellungsklavier,” 46 Husserl rightly retains Vorstellung as an indispensable component of the semantic process, in that it is responsible for “quasi-perceiving”, “quasi-judging”, “quasi-wishing”, and a myriad other mental quasi-acts. 47 Indeed, his very concept of “appresentation” that plays such an important role in the Cartesian Meditations is all about imaginability and its function in the constitution of everyday reality. Not only does “reproductive consciousness” belong to every act of perceiving, as a possibility. “Every memorial as-if” is a sort of “transforming fiction” and “every experience has as its counterpart a phantasy (a re-presentation) corresponding to it”.48

44

Maurice Natanson, Husserl: Philosopher of Infinite Tasks (Evanston: Northwestern University Press, 1973), 140. 45 Edmund Husserl, Experience and Judgment, 36, 52. 46 Wittgenstein, PI §§395; 512; 6. 47 Husserl, Phantasy, Image Consciousness, and Memory, 413. 48 Husserl, Phantasy, Image Consciousness, and Memory, 369; 700f.

Meaning and Imaginability in Husserl

73

If so, it stands to reason that imaginability must play the lead role in our functioning in the Lebenswelt as the machinery for “inactuality modification”. Even in fantasy can I feel “real desire”. As to language, the vast bulk of our linguistic performance does not draw immediately on perception but on reproductive and productive Vorstellung, defined by Husserl as “semblance acts” and “as-if-modification”.49 The point is how to bring Husserl’s insistence on Vorstellung to fruition with respect to meaning in natural language without collapsing it into the private meanings of psychologism. As already suggested, we can do so by describing Vorstellung not in general but in Husserl’s terms of “exclusive directionality”, revised as community sanctioned combination of the sounds of expressions and meaning at the moment of the event, subject to the “linguistic linkage compulsion”. Here, far from being anything private, Vorstellung is functionally defined as an intersubjectively shared and so indirectly public schematization.

Language and Lebenswelt Given Husserl’s starting point in the description of the acts of consciousness we are bound to perform differently in arithmetic, geometry, logic, perception, linguistic meaning, Vorstellung, and other ontological domains, how does he finally arrive at the a priori of the Lebenswelt, and does so in such a forceful manner? How does Husserl transform his firstperson perspective into the third-person perspective of the intersubjective life-world? How does the Lebenswelt arise in Husserl as his overriding ontological commitment? The short answer seems to be that having described acts of consciousness for almost a lifetime, all of which left dangling the Kantian question of the condition of possibility of such individual acts, the Lebenswelt began to loom as a massive and pressing inference of a necessary ground. If our subjective acts are indeed such and such, then only a radical intersubjectivity can answer the question of their genesis, an answer postponed for half a century. Husserl’s long-sustained first-perspective perspective turns into the epistemic and methodological device for the third-person perspective of the “intermonadic community” of the Lebenswelt. 50 In the Crisis Husserl offsets the “formula world” sharply from the Lebenswelt.51 In the former, nature is seen though the lens of formalisation and “numerical magnitudes”, its “meaning” lying in 49

Husserl, ibid., 436; 447f.; 709f. Husserl, Cartesian Meditations, §56. 51 Husserl, The Crisis, §9. 50

74

Chapter Three

idealities; in the latter, intuitive grasp, Gewohnheit (habit) and familiarity determine our “non-essential typifications”. The “formula world” thrives on “prediction to infinity” and confuses its method of measurement with what the world is actually like, forgetting that its tools of predictive inference are in reality derived from “ordinary inductive knowledge”: scientific prediction is the forgotten Vor-meinen (meaning in advance) that has always characterized the Lebenswelt. Although initially, the Lebenswelt was a “scientific critical, not a social-philosophical concept”,52 its social implications were unavoidable. We could sum up the characteristics of the Lebenswelt thus: intuitive “pre-givenness as existing for all in common”; “normality” and “habit” in a “horizon of familiarity”, underestimated by Bourdieu and Habermas, as Dermot Moran has pointed out, experienced as Heimwelt versus any Fremdwelt, characterized by a foreign tongue;53 a certain “naivety (partial, personal, and non-technical) and “taken-forgrantedness … rich modes of mundanity … experiential density … familiarity and [its] fundamental historical and cultural horizons”, familiarity being defined by Natanson as the “monothetic result” of “the capacity of consciousness” of setting aside the history of polythetic acts of perception”;54 and in Husserl’s own words, “the historical communal life” and the Geltung of “its various subjective ways”.55 In Husserl’s holism of the Lebenswelt, language does not simply occupy a clearly demarcated domain. Once language has been invented, it is not possible to separate it from the Lebenswelt. Whatever meanings we are able to produce non-verbally can also be covered by language, even if not at the same level of specificity. Furthermore, we cannot describe the world we live in without recourse to language either as definiens or definiendum. This means that Husserl’s view of the Lebenswelt as the ultimate a priori, the transcendental ground of all human endeavour, which includes science, is always already saturated by natural language. This is not to say, however, that language has fully replaced non-verbal meaning acts. What it does mean is that language as the main vehicle of predication, having transformed S as p into S is p, has inherited and incorporated its nonverbal precursors perception and imaginability in a way concealed by idealisation 52 Klaus Held, “Husserl’s Phenomenology of the Life-World,” trans. Lanei Rodemeyer, in The New Husserl, ed. Donn Welton, (Bloomington: Indiana University Press, 2003), 32-62; 50. 53 Dermot Moran,”Edmund Husserl’s Phenomenology of Habituality and Habitus,” JBSP 42, 2 (2011), 53-77; 67f. 54 Natanson, 1973, 127; 133. 55 Edmund Husserl, Crisis, 152.

Meaning and Imaginability in Husserl

75

and formalisation. This suggests that what Husserl deplores in the Crisis as the non-reflective mathematization of the world, is what prevented him in Logical Investigations from seeing that natural language semantics in some way “contains” and so preserves the Lebenswelt. The Heimwelt has its specific “language practices” 56 reflecting what is specific about any particular Lebenswelt. It is not in propositional generality, but only from within a “thick” conception of the Lebenswelt that such features as Grice’s “presuppositions”, “conversational maxims” and “cooperative principles”, or Austin’s “felicity conditions” can be argued.57 The question to be asked, then, is what language must be like to be so intimately enmeshed in the Lebenswelt. Above all else, such a description must grant centre stage to perception and its variants produced by imaginability. Both are indispensable for Meinen and Vor-meinen, meaning and meaning projection. In this sense, the following definitions attempt to remedy Husserl’s early semantics by drawing on his later insights. (1) Language is a set of instructions issued and monitored by the speech community to speakers for imagining, and acting in, the Lebenswelt and its various ontological sub-domains. (2) Linguistic meaning-acts instantiate the social rules of the Lebenswelt. (3) Meaning is the necessary activation of expressions, meaningless outside a specific Lebenswelt, by non-verbal mental materials (aboutness) regulated by concepts according to “exclusive directionality” (indication and intimation), quality (act-quality; modality), quantity (amount of act-matter), and degree of non-essential typification of resemblance relations, under community guidance (sufficient semiosis). In this sense, meanings are socially sanctioned schematized Vorstellungen in response to signifiers. As such, meanings are indirectly public. (4) Signifiers are arbitrary, while all natural language signifieds are motivated by virtue of carrying resemblance relations. (5) Linguistic concepts appear in at least three different versions: as “hardedge concepts” in formal systems; “soft-edge” concepts in non-formal theorizing; and “soft-core” concepts in the bulk of natural language use. (6) In natural language, meaning ideality and identity are replaced by the fuzzy concept of typified intersubjective meaning agreement guaranteed by the “linguistic linkage compulsion” which determines every meaning as a consequence of pedagogy exerted by the speech community. (7) Formal meanings remain well described by the definitional ideality of syntactic relations and identity conditions and the reduction to zero of mental material aboutness and modality. This crucial difference sets the 56

Dermot Moran, “Edmund Husserl’s Phenomenology of Habituality and Habitus,” 68. 57 Cf. Chapter Nine.

76

Chapter Three

description of formal systems sharply apart from a semantics of imaginability.58

58

A summary to be resumed in detail in Chapter Ten.

CHAPTER FOUR THE PROPOSITIONAL ROUTE: FROM FREGE TO HYPERINTENSIONAL SEMANTICS

Introduction: Frege’s opening gambit “The more closely we examine actual language, the greater becomes the conflict between it and our requirement [that is] the crystalline purity of logic”, writes Wittgenstein in the Philosophical Investigations. 1 Wittgenstein’s observation is pertinent in the context of the question of what sort of presuppositions have informed analytical language philosophy since Frege to writings in hyperintensional semantics. But why should the conflict between our object of inquiry and the analytical method introduced by Gottlob Frege widen as a result of addressing natural language as carefully as we possibly can? The answer to this question is by no means obvious. One way of attempting a partial answer is to explore how and to what extent Frege caters for the kinds of ingredients of language that our commitment to imaginability demands. We can do this by recalling the opening of his famous paper “On Sense and Reference” of 1892 where he distinguishes the tautological statement “The evening star is the evening star”, an uninformative expression of the self-identity of one and the same sense, from the expression “The evening star is the morning star”, which is informative in that it differs in its cognitive value by providing two different senses.2 However, while the evening star and the morning star differ as to their senses, they share the same referent, the planet Venus. Frege then generalises sense as determined by “the mode of presentation of that which is designated” and reference as a proper name or individual designation. Thus the morning star and the evening star have 1

Ludwig Wittgenstein, PI §107. Gottlob Frege, “On Sense and Reference”, in Translations from the Philosophical Writings of Gottlob Frege, ed. Peter Geach and Max Black (Oxford: Blackwell, 1970), 56-78. 2

78

Chapter Four

different descriptions but identical reference. The possible combinations Frege introduces also permit us to distinguish between true and false statements, depending on what sort of references we associate with the chosen senses. Thus reference is identified with the truth-value of a statement. All this looks persuasive enough, until we take Wittgenstein’s opening observation seriously, for then we will discover some serious flaws in Frege’s opening gambit. First of all, cave exempla! For philosophical examples can hide as much as they reveal, and Frege’s choice of planetary relations is acutely deceptive. Indeed, had he chosen an actual and typically more intricate sentence of natural language as his starting point, analytical language philosophy may very well have taken a very different direction. What his example hides is that it allows him to embark on two strategies which more complex, “natural” sentences would not permit. His first strategic move is to equate the three senses of Venus with the formal senses of the sides of a triangle. He does so without justification, even though this equivocation has farreaching consequences, as we shall see. Above all, it has since invited generations of logicians to represent natural language in formal terms, a pursuit which language allows but which, I submit, changes its character into something it is not. I will have more to say on this topic later in the chapter. An obvious problem of formalisation is that it covers up the important difference in comprehension and performance of these two distinct sign systems. While the senses of “morning star” and “evening star” require the user to be acquainted with typical mornings and evenings, no such requirement affects the arithmetic and geometric relations of the sides a, b, and c, which are strictly governed by axioms and definitions. As such, all formal systems are ruled externally by syntax. In order to perform the formal equation of x = yĶ all that matters is the syntactic relationship between x and y. In natural language, in contrast, syntax is only one important ingredient. In a nutshell, formal signs systems are a priori stipulations; any systemic reconstruction of natural language by contrast is of necessity a posteriori. At the very least then, Frege’s reduction of natural language terms to merely syntactic relations should have been carefully justified. The second strategic move, related to the first, is Frege’s banishment of Vorstellung, or iconic mental projections, from natural language sense as well as reference, on the grounds of subjective variability. Since that moment imaginability has been shunned by the majority of analytical language philosophers. Had Frege chosen a sentence such as “Do you

The Propositional Route: From Frege to Hyperintensional Semantics

79

recall the party when Janice Howard wore that gorgeous blue dress, drank a bit too much, and wouldn’t stop talking about her unreliable husband?” neither formalisation nor the elimination of mental states could have been that neatly demonstrated. Frege’s choice of example strongly favours an externalist approach, while such expressions as “gorgeous blue dress”, “drank a bit too much”, “unreliable husband”, and in particular the indirect speech act of nudging the addressee to “recall’ a colourful incident from the past, require the activation of a mental iconic scenario without which the meaning process could not occur. So, what a linguistic expression is about, or aboutness, in Frege looks artificially reduced to external, scientifically secured facts, whereas in the bulk of natural language sentences aboutness requires a good deal of mental labour, inaccessible to externalist explanation and so requiring an analysis informed by some kind of intersubjective mentalism. Furthermore, aboutness is only one side of the linguistic meaning process, of which reference is a special application. The other side of meaning, which is equally important yet altogether missing in Frege’s account, is the manner of speaking or implicit deixis, carried by voice. The tone of voice in which the “Do you recall …” is spoken makes all the difference to the kind of reception in the act of comprehension the sentence receives, that is, in the meaning event. The all-important feature for instance of “intention-reading” theorized by Michael Tomasello3 is denied us in Frege’s account, as is the option of the reversal of propositional content by tone. Later analytic attempts to capture speech intention by way of “utterance meaning” and “speech acts” as adjunct features of language suggest that natural language as an object of inquiry was initially too narrowly conceived. 4 Yet without the intricacies of implicit deixis, natural language would not be what it actually is, a process demanding interpretive mental projections, grounded in imaginability. The elegance, persuasive force, and historical significance of Frege’s magisterial opening of modern language philosophy are of course not in question. What is, is whether as a result of his chosen interpretive frame, language as an object of inquiry was perhaps altered into something it is not. While Frege partially revives Aristotle’s resemblance relations (homoiomata) and elaborates the Aristotelian path towards truth relations in language via his axiomatic theory of propositional logic, his equation of the sense of linguistic expressions with “pure”, that is, definitionally ruled 3 4

See Chapter Seven. See Chapter Nine for a discussion of pragmatics.

80

Chapter Four

“thought”, looks like a radical distortion of what actually occurs in and by means of language. Wittgenstein’s “requirement” of the “crystalline purity of logic” in Frege’s case was an import into natural language from his earlier Begriffsschrift (1879), his Logical Notation or Concept Script, a “formula language of pure thought”. No wonder then that the Fregean analysis produced the kind of conflict Wittgenstein is alluding to, a conflict which shows up in a number of areas of the propositional approach to natural language. In the following, I list these under the heading of “post-Fregean fallacies”.

Post-Fregean Fallacies By way of qualification, I want to emphasize that many of these fallacies are fallacious only when they are applied to natural language. Compositionality, for example, also known as Frege’s principle, “the meaning of a compound expression as a function of the meanings of its parts” works perfectly well in formal systems and sometimes but not always in natural language. Fallacies here, I want to show, are a consequence of the application of propositional and formal assumptions to natural language expressions and in particular of an underestimation of the centrality of mental, iconic projections as a necessary ingredient of the meaning process. In other words, the fallacies to be discussed are meant to show that the kind of interpretive labour involved in natural language is in significant respects different from the kind of labour required for processing arbitrarily stipulated propositional and formal systems.

The truth trap Donald Davidson’s is arguably the most influential proposal for a truthconditional semantics to date. This is why I take it as my stalking horse in pursuit of the question of how such a theory stacks up against imaginability as the stipulated metasemantic deep ground of natural language. For Davidson, understanding a language is knowing “what it is for a sentence – any sentence – to be true”, which is the same, he says, as knowing “the semantic concept of truth for a language”. 5 The natural language lexicon and its syntactic, combinatory rules can be regarded as if they were a finite formal grammar. In parallel with this grammar we can construe a formal semantics in a metalanguage M, such that the lexical 5 Donald Davidson, “Truth and Meaning,” in Semantics: A Reader, ed. S. Davis and B.S. Gillon (Oxford: Oxford University Press, 2004), 222-233; 226.

The Propositional Route: From Frege to Hyperintensional Semantics

81

semantics, together with its rules for semantic combinations, determine the manner in which sentences have meaning. To achieve this, Davidson first replaces the intensional “means p” in “s means p” by the intensional “means that” with the result of “s means that p”, whereby “p” is replaced by a sentence. In a second transformation he substitutes for the “obscure”, intensional “means that” a sentential connective for the replacement sentence and provides “the description that replaces s” with its own [extensional] predicate”. As a result, Davidson arrives at Tarski’s Convention T, (T) s is T iff p, but with a difference. As a consequence of his turn from intensional to extensional entities, Davidson’s theory of meaning is now what Quine called a theory of reference, as Davidson himself acknowledges. 6 In a footnote Davidson draws our attention to Quine’s earlier support of the semantic determination of a “word” by the “extent to which the truth or falsehood of its contexts is determined”.7 The single most important source of Davidson’s approach to meaning is Tarski’s Convention-T and its well-worn formula “‘snow is white’ is true, iff snow is white”. However, a crucial point in Tarski’s account, consciously set aside by Davidson, is that the natural language terms, “snow”, the copula “is”, and “white” are to be taken strictly only as standins for formal terms. Yet Tarski was adamant that none of them could be identified with formal entities. After all, he writes, “it is only the semantics of formalized languages which can be constructed by exact methods”, while as far as natural language is concerned “not only does the definition of truth seem impossible, but even the consistent use of this concept in conformity with the laws of logic”.8 If we disregard this vital difference we are guilty, Tarski warns, of destroying what makes natural language “natural”.9 What makes language “natural”, I suggest is its core ingredient of imaginability in both aboutness and implicit deixis. No matter how sophisticated our truth-conditional scheme, truth supervenes on imaginability. But instead of bluntly charging Davidson with having granted truth an undeserved priority over imaginability, I want to ask the following questions. (1) Are truth-conditions not just another languagegame which we can play with language but which does not bestow an essential function on them in how natural language expressions mean? (2) 6

Ibid.; Willard van Orman Quine, From a Logical Point of View: LogicoPhilosophical Essays (New York: Harper and Row, 1963), 130. 7 Davidson, “Truth and Meaning,” 233, n.8. 8 Alfred Tarski, Logic, Semantics, Metamathematics [1936], trans. J.H. Woodger (Oxford: Clarendon Press, 1956), 403; 153. 9 Tarski, Logic, Semantics, Metamathematics, 267.

82

Chapter Four

Are natural language sentences truth bearers in the first instance in the way propositions are? (3) Are truth and its conditions typically retrievable in social discourse? (4) Does the logical form of a sentence capture all there is to its semantic structure? (5) Can the rules for the retrieving process be appropriately formulated via a formal metalanguage M by which a specific object language L can be fully covered? (6) Can such a meta-theory be arrived at by the principle of radical interpretation? (7) Is the principle of charity a realistic basis on which to reconstruct belief systems? (8) Can the principle of radical interpretation be made realistic by the principle of charity? (9) Is it a safe bet to move from belief ascription to semantics? (10) Does radical interpretation yield both, a theory of truth and a theory of meaning? (11) Can all sentences which do not yield easily to truth be replaced by sentences that do? (12) Are sentences for which the replacement strategy fails not essential for a general description of how natural language works? (13) Is Tarski’s “naturalness” of language in no way undermined by the proposed truth program? If one were to try and answer such questions, perhaps the following considerations could be taken into account. (1) A sentence such as “I was murdered” immediately yields its meaning, while the question under what conditions the sentence would be true is a language game ill suited to get at its meaning. (2) Even when sentences are truth bearers, meaning does not appear to depend on that feature. (3) In a sentence like “Do you think John was right when he told us that Mrs MacDowell was being facetious when she complimented Sarah that evening?” truth is irretrievable, as are its conditions. We cannot be certain what possible conditions to choose from. A variety of such conditions would qualify. (4) Can we be confident that the logical form of a sentence, in Tarski’s Convention-T or in Davidson’s version, caters for the non-propositional resemblance relations it conveys? Can it capture them beyond a reductive indication? (5) Since Davidson’s metalanguage M is derived from the object language which he has conceived in a certain way, the results of that conception will be duplicated at a more abstract level. If the basic conception should prove erroneous, so will M. (6) Ascribing mental states to native speakers looks like a Husserlian appresentation, a co-presentation of the observer’s assumed mental parallel states, a process that can go awfully wrong unless the speaker is “at home” in the culture. (7) Davidson’s principle of charity is an optimistic assumption in the face of the irrationality and often undecidable features of human belief systems. (8) If so, how can radical interpretation be rendered realistic by the principle of charity? (9) Given the many undecidable cases testifying to the unreliability of beliefs, taking

The Propositional Route: From Frege to Hyperintensional Semantics

83

the reconstruction of a belief system from the outside of culture as the basis for semantic inferences may very well be in conflict with what we think is true about reality. (10) In light of concerns (6) to (9) the assumption that radical interpretation can generate both a theory of truth and of meaning looks massive. (11) Whether sentences that fail to yield to truth-conditions can be replaced salve veritate by sentences that do, could be answered only by an analysis of a very large corpus of examples. Given endless possible variations on such sentences as “I was murdered”, all of which are immediately comprehended by speakers of the language, the substitutability thesis would need shoring up. (12) We cannot assume that fictional sentences are not essential for natural language, since they illustrate a fundmental feature of natural language: its self-reflective, experimental and rule breaking character. (13) Lastly, it would seem that truth-conditional considerations of the Davidsonian variety impose restrictions on natural language of a kind that seriously undermines its “natural” character. While semantic holism is attractive, Davidson’s version appears constricted by its propositional character. It would seem that in the face of Tarski’s “naturalness” of language, Davidson has, as he himself fears, taken a far too “optimistic and programmatic view of the possibilities for a formal characterization of a truth predicate for a natural language”.10Lastly, I ask whether it is likely that the heterosemiotic nature of imaginability in language can be captured by Davidson’s homosemiotic and monosemiotic procedure. Reassuringly, some such doubts can be found in his own writing. In his posthumously published Truth and Predication Davidson once more addresses his adaptation of Tarski’s Convention-T to natural language while also granting us a glimpse of a larger horizon within which linguistic meaning may have to be understood. 11After defending the position taken in “Truth and Meaning”, Davidson makes a startling declaration, a view he had already indicated in the Dewey lectures. He writes, “the conceptual underpinning of interpretation is a theory of truth; truth thus rests in the end on belief and, even more ultimately, on the affective attitudes”. 12 I italicize the final phrase because of the all-important place it occupies in the Vorstellungswelt. Is Davidson at the end of his last work perhaps conceding the subordination of truth to Vorstellung and so to imaginability? Has some sort of “psychologizing” suddenly been legitimated? Just as 10

Davidson, “Truth and Meaning,” 232. Donald Davidson, Truth and Predication (Cambridge, Mass.: Harvard University Press, 2005). 12 Davidson, Truth and Predication, 75. 11

84

Chapter Four

Aristotle’s pathemata relate to homoiomata, Kant’s Anschauungen rest on “Affektionen”13, and Wittgenstein concedes “Affektion der Seele”, while denying their connection with meaning,14 could such “attitudes” perhaps be the result of “receptivity”, our “capacity of being affected in a certain manner with Vorstellungen”?15 I conclude by asking whether it is possible after all that Davidson’s “affective attitudes” offer a promising path to meaning and so furnish an answer to his rhetorical question “How on earth do we ever manage to understand what speakers say?” 16 If so, then Davidsonian truth relations would be parasitic on imaginability. Vorstellbarkeit, then, looks larger than propositionally reduced belief and meaning looks larger than truth.

Meaning as observable If truth supervenes on meaning rather than the other way around, what about the claim, argued famously by Wittgenstein, that meaning is use and as such observable?17 After all, a great deal goes on in the minds of native speakers of a language in linguistic exchanges, some of which is incidental, to be sure, but much of which is necessary for successful meaning events to occur. “See you at the party” … “See you then” are sentences that can only be partially comprehended by uninformed bystanders. Far from this being an exception, natural language use typically relies rather heavily on unspoken but relevant aboutness. Behaviourist stimulus-response approaches fail in the face of this obstacle, as does formally grounded externalism. But even the most sophisticated notions of “meaning as use” in the wake of Wittgenstein’s Philosophical Investigations fall foul of the necessary condition of mental realisations, such as memory, attentionsharing and “intention-reading” 18 as part of the process of meaning exchange. In order to be able to use a colour term (sepia, vermilion) I have to be trained to master both a “paradigm” of colours19 and the appropriate 13

Kant, CPR A 68/B 93. Wittgenstein, PI § 676. 15 Kant, CPR A 494/B 522; my emphasis. 16 Donald Davidson, “The Perils and Pleasures of Interpretation”, in The Oxford Handbook of the Philosophy of Language, ed. E. LePore and B.C. Smith (Oxford: Clarendon Press, 2006), 1056-1068; 1059. 17 Ludwig Wittgenstein, Philosophical Investigations, trans. G.E.M. Anscombe, P.M.S. Hacker and Joachim Schulte (Oxford: Blackwell, 2009). 18 Cf. Michael Tomasello’s contribution to this topic in Chapter Seven. 19 Wittgenstein, PI §51; further references to the PI in this section are provided in the text. For a detailed treatment of “meaning as use” see Chapter Five. 14

The Propositional Route: From Frege to Hyperintensional Semantics

85

“lexicon”. Once such mastery has been achieved I do not have to pull out the colour scheme from my pocket each time I hear the term, but am able to link the appropriate Vorstellung (of sepia instead of vermilion) with the word sound. In this way “it is in language that an expectation and its fulfilment make contact”. (PI §445) Nor is the mental image a mere “private sensation”; (PI §274) it is socially controlled. To test this claim, try to image your shoulder when you hear the signifier “kneecap”. We cannot escape the “rigidity” of the link between word and Vorstellung which the speech community has dictated to us via sufficient semiosis. For Wittgenstein, the involvement of Vorstellung in meaning is incidental; it is no more than accompanying language by playing on the “piano of the imagination” (Vorstellungsklavier). (PI §6) I will address the case of Wittgenstein in detail in Chapter Five, restricting my argument here to the claim that if we accept the definition of natural language as “a set of social instructions for imagining, and acting in, a world,” observables furnish insufficient conditions for meaningful linguistic exchanges.

The sentential fallacy By “sentential fallacy” I mean the assumption that mental states are well characterized as sentences.20 This is not to suggest that we do not form sentences in the mind. Of course we do. Yet sentences in the mind form only a miniscule part of the field of the mental processes summed up under the concept of imaginability. The fallacy could also be levelled for instance against Fodor’s “language of thought” (LOT) according to which the internal representations of the world in the mind are sentence-like. It is fallacious in that it falls foul of Gilbert Ryle’s “ghost in the machine” criticism which says that any explanation of language via a language in the mind of necessity leads to an infinite regress of ever further languages being required to provide the basis for each precursor. 21 So LOT is begging the question in a big way. Language cannot function as an anchor for itself. This is why imaginability, as transformation of perception into mental iconic projections, enjoys a significant advantage. Since it is rooted in perception, made up of tactile, olfactory, gustatory, thermal, visual, gravitational and other readings, imaginability can serve as the heterosemiotic anchor of language. Now linguistic infinite regress is avoided. It is at the level of perception and imaginability that the buck of 20

Michael Devitt and Kim Sterelny, Language and Reality (Oxford: Basil Blackwell, 1990), 116. 21 Gilbert Ryle, The Concept of Mind (London: Hutchinson, 1949), 32f.

86

Chapter Four

meaning can be arrested for practical purposes. Importantly, this leaves intact Kant’s arguments in favour of the possibility of an infinite chain of conceptual modifications and the open-ended analysis of empirical concepts, as well as Peirce’s resumption of this Kantian theme in his notion of infinite semiosis, popularized by Derrida.22

Syntactocentrism A major hurdle to a rich description of language is “syntactocentrism,” so aptly termed by the linguist Ray Jackendoff.23 It can be found in both the post-Fregean accent on formal relations, of which the chess-game analogy is a ubiquitous example, and some branches of linguistics. The strength of research on syntax lies in explicating the ordering principles we can abstract from living speech, as well semantic changes resulting from syntactic variation. But syntax cannot tell us how meanings come about; in natural language, the recognition of syntax presupposes meaning. Nor does syntax run quite as freely as its formalisations suggest. This criticism applies as much to Carnap’s semantics defined as an alignment of two kinds of syntax, 24 as it does to the syntactic picture painted by such linguists as Ferdinand de Saussure and the early Noam Chomsky. 25 However, we cannot deny Wilhelm von Humboldt’s claim that we can generate an infinite number of sentences from a limited stock of terms. The criticism of syntactocentrism, however, throws serious doubt on the idea of conditionals as a linguistic innovation. After all, it is not plausible that before the event of language hominids lived in chaos. “If-then” patterns must have shaped their world according to principles of human physiognomy and respect for objective constraints to secure survival. If so, conditionals, in a primitive form, have been inherited from pre-linguistic

22

Kant, CPR A727ff.; Peirce, CP 1.339; Jacques Derrida, e.g. in “Sending: On Representation,” Social Research 49, 2 (1982), 294-326; 299. 23 Ray Jackendoff, Foundations of Language: Brain, Meaning, Grammar, Evolution (Oxford: Oxford University Press, 2002), 7ff.; cf. also Anna Wierzbiecka, “Syntax vs. Semantics,” Theoretical Linguistics 5, 1 (1978), 115-133. 24 Rudolf Carnap, Introduction to Semantics and the Formalisation of Logic (Cambridge, Mass.: Harvard University Press, 1975); Meaning and Necessity (Chicago: University of Chicago Press, 1967). 25 Saussure, Cours, 141-192; Noam Chomsky, Syntactic Structures (The Hague: Mouton, 1957); Aspects of the Theory of Syntax (Cambridge, Mass.: MIT Press, 1965).

The Propositional Route: From Frege to Hyperintensional Semantics

87

behaviour, a protosyntax already incorporated in, rather than excluded from, any kind of “proto-language”.26 We can formulate the “syntactic stranglehold” in two versions, a weak and a strong form. In the weak version, language is regarded as a kind of chess; in its strong form, syntax rules language such that semantics is a function of syntax. The study of natural language in both its linguistic and philosophical guises has been seriously affected by the simile of language being like chess.27 To quote from Saussure’s Cours, “Of all comparisons that might be imagined, the most fruitful is the one that might be drawn between the functioning of language and a game of chess. In both instances we are confronted with a system of values and their observable modifications. A game of chess is like an artificial realisation of what language offers in a natural form”. Yes, one could say, language is like chess but only at the level of syntax. And yes, the “language” of chess, as Sellars observes, “by virtue of its special vocabulary, has a certain autonomy with respect to the everyday language in which it becomes embedded”.28 But what is fundamental to language are its aboutness and voice. Chess is not about anything. This is why chess, contrary to language, can be perfectly well played by means of matrix mathematics; it is in principle homosemiotic and as such pure syntax. Nor can we alter a chess move by an ironic stance, an option crucial for the description of natural language.

Compositionality Frege’s principle of compositionality formulated simply as “the meaning of a compound expression is a function of the meanings of its parts” can be observed to apply to a large number of sentences in natural languages. However, when so applied, compositionality fails to deliver the goods in the way it does in formal systems. This is so for two reasons. One is that compositionality is insensitive to the modal problem in natural language, 26

See e.g., Derek Bickerton, Language and Species (Chicago: University of Chicago Press, 1990), 130-163. 27 Wittgenstein PI §§ 33, 108, 190, 563; Gilbert Ryle, “The Theory of Meaning,” in British Philosophy in the Mid-Century, ed. C.A. Mace (London: Allen and Unwin, 1957), 239-264; Wilfrid Sellars, Science, Perception and Reality (Atascadero, California: Ridgeview, 1991), 344ff.; Ferdinand de Saussure, Cours de linguistique générale, ed. Charles Bally, Albert Séchehaye, and Albert Riedlinger (Paris: Payot, 2005), 43, 125, 153. 28 Saussure, Cours, 125; Sellars, Science, Perception, and Reality, 344.

88

Chapter Four

the manner of its speaking or implicit deixis. Only in natural language can we reverse the propositional content of a sentence by an ironic illocution. But this relation is by no means exhausted by merely two opposing meaning options. Rather, the polarities of straight and oppositional readings function as distant polarities allowing for, at least theoretically, an infinite number of inflections and meaning shifts in between. This is where Michael Tomasello’s all-important mental activity of “intentionreading” bites.29 The exclusion of natural language modality, as the overall manner of speaking, as we find it in much of the philosophy of language, turns out to be a self-serving procedure that makes language into something that it is not. A second stumbling block for Fregean compositionality is the prominence in language of idiomatic expressions. We would be hard pressed to show the contribution that each component makes to meaning in sentences like “Don’t come the raw prawn” or “He throws his money around like a man with no hands”, suggestive of a miserly person. In an expression like “I could eat the crutch out of a lowflying duck”, the law of compositionality provides us with a false meaning or, according to Davidson’s account of metaphor, a “literal” meaning that is false while nudging us to seeing something else via its utterance meaning.30 Even if we succeeded in rewriting these sentences in a way more conducive to the mechanism of compositionality, all we would have shown is that we have to strain to make the concept of compositionality work in natural language. As such, compositionality in its propositional form reveals itself as an alien import. What is imported into natural language here is the requirement for calculus propositions that their meanings are fully determined by their syntax and the meaning of their individual parts. Furthermore, in natural language, expressions often mean different things under different circumstances, which once more also demonstrates that the reduction of natural language to semantics without retaining at least a minimum of pragmatic imaginability is misleading. To comprehend compositionality we must first get to base one of the meanings of single terms, which in their various combinations give rise to overall meanings not always predictable by individual lexical components, as in a large number of idiomatic expressions. Tu long zhi ji (The art of killing dragons) means as much as doing something without any practical value. Yet how do we get to base one? And how do we get to base two in idiomatic speech? Semantics in its many forms does not answer such 29

See Chapter Seven. Donald Davidson, “What Metaphors Mean”, in On Metaphor, ed. Sheldon Sacks (Chicago: The University of Chicago Press, 1981), 29-45.

30

The Propositional Route: From Frege to Hyperintensional Semantics

89

questions. We must turn to metasemantics, that is, reconstruct the conditions that would allow us to transform verbal cues into Vorstellung. In other words, once again imaginability is at the root of the matter. It is in and by means of imaginability, in both its initially defined senses, that we are able to project imaginary scenarios. Where does this capacity come from? We have been trained to do so. We imagine situations of the kind allowed for by language or, vice versa, we find combinations of words that would best allow the projection of the sort of scenario we wish to express linguistically. Importantly, those scenarios, indispensible for the comprehension of language, are themselves not linguistic, let alone logical relations. Above all, it is pedagogy that trains us what to imagine in idiomatic expression, as in ordinary speech.

Recursivity Just as does compositionality, recursivity in natural language often fails in the face of idioms. And as Jackendoff rightly notes, to regard idiomatic phrases as a “marginal” feature of natural language is a fundamental and inexcusable error. 31 In formal systems, recursivity is a necessary consequence of the system’s axiomatic principles, definitions and syntactic rules. While recursivity also partially works in natural language, as grammatical paradigms demonstrate, it cannot be regarded as the same kind of rule in language. Recursive infinity is a theoretical option unexplorable in cultural practice. In any case, the mastery of a language is judged by idiomatic competence at least as much as by expertise in recursive rules. “She made herself a cup of tea” and “I envy you your fine garden” are syntactic patterns with a high recursive potential, as are hundreds of other sentences in English. But the logical ideal of recursivity runs into difficulties in cases such as, “You are telling me,” “He is talking through his hat and lying through his teeth,” “A sop to Cerberus,” “Born on the wrong side of the blanket,” and thousands of other expressions in English. Since “recursive functions” are always “calculable functions,” idioms throw a big spanner in the works of recursivity, and in this respect English is certainly no exception. It is fallacious to claim recursivity to be generally and broadly applicable to natural language. Contrary to formal systems, in natural language recursivity covers only part of the total. And partially applicable theories cannot claim general validity. 31

Ray Jackendoff, Jackendoff, Ray (2002) Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford: Oxford University Press, 2002); Languages of the Mind: Essays on Mental Representation (Cambridge, Mass.: MIT Press, 1992).

90

Chapter Four

Predicability Another import from propositional logic is what can be predicated about something. As such, predicability, (Keil 1979; 1981) is assumed to be dictated by syntax. The assumption, for instance, that “the tree was sincere” or “colourless green ideas sleep furiously” are not predicable, while “the apple tastes sour” is. This is misleading. First, both allegedly nonpredicable sentences could function perfectly well in their own possible worlds, say cartoon stories with personified trees or puppets called “colourless green ideas” who snore and toss violently in their beds. Predicability now is a function of the sort of world we imagine rather than of syntax. Second, if it is Leibnizian “compossibility” rather than grammaticality that dictates predicability, then we must turn our gaze to the socio-perceptual frame of language, its pre-linguistic precursors and the possibility of a perceptual protosyntax. Much of what can be predicated about subjects and what cannot is predetermined by the phenomenal world prior to the emergence of language. Language, rather than inventing all predicability, has inherited much of it from its perceptual and gestural forerunners. This would be acceptable to a Fregean but hard to agree by Saussureans for whom the world is wholly absorbed in language and who must regard non-linguistic grasp as “nebulous”. Again, in natural language as opposed to formally regulated sign systems, what can be said about persons, things, and events depends on the heterosemiotic frame within which language is embedded rather than on the internal rules of the system. The perceptual world, idiomatic expression and pragmatics rather than an idealised semantics have the last word in this respect.

Meaning identity Applying the notion of semantic identity to meaning events in natural language is fallacious. Meaning identity is a formal requirement which is neither necessary nor available in natural language. Meanings are always approximations, a sufficient condition for linguistic exchange. When in doubt, the speech participants typically negotiate meaning events within the frame of sufficient semiosis rather than according to identity or truth conditions. About the sentence, “A thing is identical with itself”, Wittgenstein famously quipped “there is no finer example of a useless proposition, which yet is connected with a certain play of the imagination”.32 Likewise, 32

Wittgenstein, PI §216.

The Propositional Route: From Frege to Hyperintensional Semantics

91

in his paper on “Vagueness” (1923) Russell makes this important concession. “The meaning of a word is not absolutely definite”, he writes, it always has “a greater or less degree of vagueness”. Comparing linguistic meaning with an “area” and a “target”, he says that “it may have a bull’s eye, but the outlying parts of the target are still more or less within the meaning, in gradually diminishing degree”. For Russell, meaning always has “a doubtful region”.33 This characterizes well the concept of semantic aboutness as essentially fuzzy and invites a theorization of natural language concepts as “soft-core” compared to the “hard-edge” concepts of formal sign systems and the “soft-edge” concepts of philosophy.34 When Quine contrasts his insistence on identity in analyticity with “domestic meaning” in ordinary language, he concurs in principle with Russell’s concession. For “sentences are cognitively equivalent,” Quine writes, “if putting one for the other does not affect” their “empirical content”.35 That there is nevertheless a certain “logic” even in “the vaguest sentence” is not in doubt.36 Linguistic meaning, then, is better viewed as horizonal than as definitional. And what goes for aboutness also goes for modality in both its explicit and implicit forms. In natural language, where meanings are events unfolding in time (however fast), “identity” of meaning has to be replaced by such notions as “rough similarity,” “family resemblance,” “likeness,” and similar approximations.

Sorites puzzles A fallacy which is the result of a language-game misapplied to natural language is the “sorites paradox”.37 In Schiffer’s formulation of the puzzle, we a have a premise, “A person with $50 million is rich,” followed by a second premise, “You can’t remove someone from the ranks of the rich by taking 1 Cent away from her fortune,” and the conclusion “therefore a person with only 37 Cents is rich”. This, Schiffer says, “constitutes a paradox because it appears to be valid, each of its premises appears to be true (at least when considered on its own), and its conclusion certainly appears to be false”. Schiffer then goes on to show how various 33

Bertrand Russell, “Vagueness”, Australian Journal of Psychology and Philosophy 1 (1923), 84-92. 34 See Chapter Ten. 35 Quine, In Pursuit of Truth, 53f. 36 Wittgenstein, PI §98. 37 Stephen Schiffer, “Vagueness,” in The Blackwell Guide to the Philosophy of Language, ed. Michael Devitt and Richard Hanley (Oxford: Blackwell, 2006), 225-243.

92

Chapter Four

contemporary philosophers have struggled unsuccessfuly with the puzzle before offering his own intriguing solution along the lines of “indeterminate borderline propositions”. 38 But even Schiffer does not seem to be aware of the basic fallacy being committed here. The sorites puzzle does not exist in natural language. It comes into existence only as the result of the language-game of logical identity and hard-edge concepts being played in the arena of soft-core concepts. Playing in the wrong playground, like adults sitting on toddlers’ swings, produces a paradox the solution of which feeds an industry. In natural language, speakers understand when someone is called “bald” and whether it is a joke or not; they know when to apply terms such as “fast” versus “slow” in a children’s” miniature yacht race or in a formula one event, constructing meanings differently under different circumstances. If language is viewed as “a set of social instructions for imagining, and acting in, a world,” its speakers know how to use it and the sorites paradox evaporates.

Untranslatability and ostension For all his behaviourist ontological commitments, Quine looks like an idealist logician when he enters the jungle, having stripped human interaction of all its essential characteristics. “Indeterminacy of translation,” “untranslatability” and “indeterminacy of reference” never really occur, which they would iff the world were reduced to Quine’s strictly analytical principles. Of course, Quine’s purpose is to demonstrate the impossibility of meaning conceived in a strictly analytical sense, and he certainly succeeds in this respect. 39 Quine’s jungle is no more than an imagined laboratory that fails to reflect human reality. The reason why the translation manuals “Jungle” and “English” would in reality yield adequate results after a very short time is that language is never on its own. “Gavagai” will soon mean to Quine’s linguist what it means to the native hunter because they share the same body; they see, hear, and smell in very similar ways, they eat the gavagai while they drink the native brew. Time is of the essence here, of course, as an essential feature in natural language. Soon the linguist will hunt the gavagai without confusing it with its imaginable time slice. Natural language is always embedded in a multitude of heterosemiotic relations securing meaning for its members, including newcomers. Having debunked the very notion of meaning in its strict, 38

Schiffer, “Vagueness,” 225; 236-242. Willard van Orman Quine, Pursuit of Truth (Cambridge, Mass.: Harvard University Press, 1993), 47-52. 39

The Propositional Route: From Frege to Hyperintensional Semantics

93

analytical sense, Quine introduces the lesser standard of “domestic meaning” on condition of “giving up on propositions as languagetranscendent sentence meanings”. 40 Quine’s “gavagai” episode also functions as a demonstration of the difficulties the concept of “ostension” faces if conceived in an analytical framework. It is instructive here to recall Wittgenstein’s critique of ostension early in the Philosophical Investigations. On Wittgenstein’s account, ostension fails as a language teaching device because it cannot deliver the meanings of a substantial number of verbal expressions. In particular, ostensive pointing finds it hard to distinguish between function words, “this” and “that,” colours and shapes, and so Wittgenstein asks: “how is that done?” 41 Amongst the difficulties he bridles at is what “pointing” itself refers to: “we cannot specify any one bodily action which we call pointing”. (PI §36) He also rightly objects to the idea of “private ostensive definitions” because “the institution of their use is lacking”. (PI §380) And while we can accept Wittgenstein’s earlier warning that we must not “confound the meaning of a name with the bearer of the name,” (PI §40) we can still insist that “a mental activity corresponds to these words”. (PI §36; translation corrected) This correspondence which ostension exploits is the linkage between a word and a Vorstellung. So why is Wittgenstein’s rejection of ostension as a way to meaning a fallacy? One way of putting the answer is “because he has limited the discussion to ostension in isolation”. Change the ostensive situation to one in which the main principles are repetition, variation and placing items in consistent series, ostension becomes a very persuasive pathway to meaning. To get from natural kind terms to function words, for example, one must first establish the former in a simple series. Once that series is mastered one can begin to vary it such that quantifiers and other operators become the focus of attention. Likewise, the distinctions between shapes, colours, “this” and “that”. As in Quine’s jungle, time and communicative ingenuity are a key requirement. With the imperative addition of mental activity as a necessary condition of meaning understood as a linkage process between word sounds and conceptualized iconicity, ostension can now be regarded as a special manner of demonstrating “meaning as use”.

40

Quine, Pursuit of Truth, 53. Wittgenstein, PI §38; for the remainder of this section, see the text for references to PI.

41

94

Chapter Four

Homosemiotic and monosemiotic assumptions Two further obstacles to a rich view of language are the related assumptions of its alleged homosemiotic and monosemiotic character, as in the sentence “Language is a symbolic system”. In homosemiotic systems we emphasize that all signs are of the same kind, as in calculus; in monosemiotic systems that we are dealing only with one kind of sign, as for instance in olfactory readings as a response of the organism to airborne molecules. If we regard natural language as homosemiotic we deny the possibility of semiotic heterogeneity. This denial is widespread in philosophical semantics. In the naturalism of Devitt and Sterelny, it leads to the alignment of language and world without a tertium comparationis.42 Much the same can be said of some truth-conditional theories.43 In contrast, I have tried to show that language is both polysemiotic and heterosemiotic. An intriguing case of the monosemiotic fallacy is Searle’s objection against AI semantics known as the “Chinese room argument” in which he rightly attacks the confusion of syntactic and semantic mastery of a language. Where his argument goes wrong is in the assumption that language exists semantically on its own. What is missing is the pragmatic frame of language without it cannot have a semantics. So the question Searle is leaving unaddressed is whether AI could construct a pragmatic frame for a language program by linking up its syntax with a series of machines that are in touch with the world, as for example by x-ray and infrared vision, tactile capabilities, chemical analysis, and other devices. For such a polysemiotic and heterosemiotic system Searle’s critique must be rephrased. 44 Certainly, language, by drawing on a diverse range of semiotic signs as a necessary feature of its meaning potential, especially iconicity and indirect iconicity (indexicality) in its motivated signifieds and symbolicity for its arbitrary signifiers, is fundamentally heterosemiotic.

The applied-calculus fallacy Rejecting the treatment of natural language as a formal system is one thing; it is quite another to show that language does not work like an applied calculus. For it is tempting to see significant parallels between applying a measuring tape to objects and the use of linguistic signifiers. Moreover, given Frege’s initial conflation of formal and natural language sense, it is 42

Michael Devitt and Kim Sterelny, Language and Reality, 15. E.g., Ruth Kempson, Presuppositions and the Delimitation of Semantics (Cambridge: Cambridge University Press, 1975), 31. 44 John R. Searle, Minds, Brains and Science (Harmondsworth: Penguin, 1984). 43

The Propositional Route: From Frege to Hyperintensional Semantics

95

perhaps not surprising that language has been called an “interpreted formal sign system”. 45 To be sure, regarding natural language as a kind of “interpreted” calculus does have the advantage of acknowledging at least a limited form of aboutness. In this sense, measuring a table with an inch tape is an application of a formal system to a portion of the aboutness of the phenomenal world. The tape can be regarded as formal in the sense that its subdivisions are definitionally controlled and so can be translated into another similar formal system, such as a metric tape. And the analogy would indeed be valid if natural languages were first designed as coherent and self-sufficient systems definitionally secured in the way our measuring tapes are. In short, the essential characteristic here is the a priori constitution of an axiomatically determined and definitionally secured sign system and its practical application. In this important respect natural language does not qualify. The fallacy lies in a confusion of a priori and a posteriori constituted systems. Natural language has grown in relation to and as an economising reflection of our socio-perceptual world and its imaginable modifications. As such, language takes the relation between its arbitrarily chosen signifiers not from a formal system of axioms, definitions and syntactic rules, but from the translation of nonverbal realisations of materials, events, and feelings. In language as an a posteriori constituted sign system first comes speech as social practice, and only then follows its systemic formulation in lexicons and grammar books. An all-important consequence of this is that the structure of formal systems does not reflect a non-linguistic order but merely measures it according to its own terms, while natural language not only “measures” but also, in its very make-up, mirrors a human perceptual order. This order can be expressed as an intersemiotic integration of a multi-modal, nonverbal set of heterogeneous readings, such as olfactory, tactile, gustatory, gravitational, aural, emotional, visual and other readings of the world. In this sense, all natural languages relate language and world. The main reason why Alfred Tarski warned his fellow logicians in 1936 not to apply Convention-T to natural language was that its terms are anything but formal.46 Yet there is a further complication. At a yet more primitive level this interpretive integration performed by the human organism draws fundamentally on radiation, molecular states, and other physical and bio45

Günther Grewendorf, Fritz Hamm and Wolfgang Sternefeld, SprachlichesWissen: Eine Einführung in moderne Theorien der grammatischen Beschreibung (Frankfurt: Suhrkamp, 1987), 377. 46 Tarski, Logic, Semantics, Metamathematics, 267.

96

Chapter Four

chemical inputs. The phenomenal world, the world as perceived by humans, is the result of the intersemiotic unification of a number of heterosemiotic systems, the combined result of interpretations of universal deep constraints. Language has inherited this semiotic heterogeneity. So far then the flaw of the interpreted calculus analogy lies in the failure to recognize that the aboutness of language is incommensurate with formal principles. Even more obvious is the flaw of attempting to apply calculus relations to the speech modalities of natural language carried by voice.

The formal fallacy I conclude this gallery of fallacious rogues of reasoning about language with its most serious version, the “formal fallacy”, or the confusion of natural language sense with formal sense. As we noted at the beginning of this chapter, it was employed as a deliberate strategy by Gottlob Frege in the opening pages of “On Sense and Reference”, a move that has been continuously repeated in the philosophy of language to this day. While this move highlights the distinction between sense and reference, it also results in a truncated conception of natural language, truncated in the sense that while formal systems are fully defined by syntactic relations, syntax in language is only one specific component. So reduced, language as our object of inquiry now lacks aboutness as broad referential background and above all speech modality as the all-important manner of speaking conveyed by voice. Yet for all natural languages to function as languages, both components are trained into the minds of native speakers by their speech communities and held in memory as indispensable ingredients for the constitution of meanings as indirectly public. Unlike the execution of formal systems, the performance of natural language is always a mental activation of motivated, that is, iconic signifieds on the cue of arbitrary signifiers. A fundamental difference, as we saw, between formal systems and language is their genesis, the former being stipulated a priori, the latter only admitting an a posteriori version of inquiry. Not even Leibniz’s zureichender Grund (sufficient reason) is able to wipe out that crucial difference. A serious consequence of treating natural language as if it were formal, or as yielding to formalisation without damage, is the necessary elimination of imaginability as one of its fundamental constituents. While mental iconic “cleansing” started with Frege, it remains an unavoidable by-product of all formalisation. In euphemistic analytical language, the removal of intersubjective mentalism from language since Brentano is referred to as the liberation of language philosophy from psychology. This liberation can be found, for example, in Quine’s separation of the theory of

The Propositional Route: From Frege to Hyperintensional Semantics

97

meaning as intensional (strictly definitional), comprising meaning, synonymy, significance, analyticity, and entailment, from the theory of reference as extensional, consisting in naming, truth, denotation, and extension, a sequestration that works in the domain of analyticity but not, as Quine himself concedes, for natural language. 47 The development of formal accounts of language, however, reaches its pinnacle in the post-Fregean tradition of a purely intensional semantics, from Montague Grammar to its more recent hyperintensional variety.

From Montague Grammar to hyperintensional semantics We cannot, of course, deny that even the most idiosyncratic idiom can ultimately be represented in formal terms. But when it is so represented, the initial methodological advantage of formal economy will have evaporated. In this last section of the chapter I want to sketch briefly some important steps in the formalisation of the Fregean heritage, from Montague Grammar to the most recent development in formal attempts at capturing natural language by way of hyperintensional semantics of the kind proposed by Marie Duzi et al. (2010) in which meaning is decomposed into a syntactic set of procedures which, however, to be meaningful, turn out to rely on retranslation into a natural language, a requirement which begs the very question of the full formalisation of meaning as meaning.48 If Frege had well described one kind of thought, in his terminology “pure thought”, what one could call “propositional thought”, he had to discard quite another kind of thought, “iconic thought”, that is, thought that carries resemblance relations and is tied to perception and its variation in and by Vorstellung, in short, the kind of thought we associate with imaginability. Frege’s “pure thought” is at the heart of Montague Grammar, so named after Richard Montague, whose fundamental assumption was that there is no theoretical difference between the syntax and semantics relations in natural language and those of a formal language.49 This, he thought, justified the description of natural language syntactically, semantically, and pragmatically under the umbrella 47

Quine, Pursuit of Truth, 52f. Marie Duzi, Bjorn Jespersen and Pavel Materna, Procedural Semantics for Hyperintensional Logic: Foundations and Applications of Transparent Intensional Logic, (Dordrecht, Springer, 2010); in the wake of M. Creswell, “Hyperintensional Logic,” Studia Logica 34, 1 (1975), 25-38. 49 Richard Montague, “‘English as a Formal Language” in Formal Philosophy, ed. R.H. Thomason (New Haven: Yale University Press, 1974), 188-221; “Universal Grammar,” Theoria 36 (1970), 373-398. 48

98

Chapter Four

of applied mathematics. In other words, Montague Grammar made a virtue of committing what I have called the applied calculus fallacy. In Montague’s theorization, quantification, compositionality and recursivity strictly determine syntactic combinations, while meaning is secured by the necessary and sufficient conditions of the truth of sentences, whereby truth is a function of the correspondence between meaning and how the world is. Montague grammar can be seen as a radicalization of Frege’s collapse of natural language sense into formal sense as pure thought, as well as a refinement of Frege’s principle of compositionality. By way of mathematical logic, Montague opened a path to an entire tradition of looking at natural languages as formal systems. Sentences are broken up into their various component sets which are aligned with one another by a combination of syntactic and semantic rules. Thus a sentence such as “Philosophers do not have gills” is formalized as two component sets, of philosophers and a negated having gills. Syntactic rule (1) covers the negation of a predicate. This is combined with a semantic rule (1) interpreting the first syntactic rule. In this way we arrive at a “complement formation”. Syntactic rule (2) combines the noun phrase with the predicate such that with the help of the semantic rule (2) we achieve the inclusion of one set (philosophers) in another set (not having gills). At the heart of Montague’s enterprise are the reinterpretation with the help of quantification rules of such semantic topics as identity, synonymy, entailment, intension and extension, and various kinds of ambiguity. Accordingly, synonymy is defined as sentences entailing one another. For questions of identity in relation to nouns associated with value, Montague invented this solution. In a reasoning chain of the form (1) “The temperature is 25 degrees Celsius”. (2) “The temperature is rising”. (3) “Therefore 25 degrees Celsius is rising”. According to Montague, the mistake here is one of conflating value readings and property readings, resulting from an ambiguity in the concept of “temperature”. Disambiguation is accomplished by distinguishing values and properties. From the perspective of imaginability, the ambiguity is the result of treating the shorthand expression (1) as an identity statement, while in ordinary comprehension the expression is understood as “The temperature reading is 25 degrees Celsius”, in which case the ambiguity does not arise. Another procedure of Montague grammar is the distinction drawn between “wide-scope” and “narrow-scope” reading. Thus the sentence “Every girl catches a ball” is regarded as ambiguous, allowing for “the same ball” and “different balls”. Yet competent speakers of English do not comprehend the sentence as ambiguous. On the cue of the sound sequence “Every girl

The Propositional Route: From Frege to Hyperintensional Semantics

99

catches a ball’ they typically imagine a situation in which every girl catches a ball. To evoke the mental scenario giving meaning to “the same ball” would require precisely this phrasing. This, not unlike the sorites puzzles, looks like a case of distinction dictated by logic rather than one inherent in language in its standard use. Following Montague grammar, a sentence of the form “Mary wants an angel” can be regarded as ambiguous since “want” allows for both existence and non-existence. The explanation of the problem would then be a case of “scope ambiguity”. Once we have disambiguated the two kinds of “want” the problem vanishes. Alternatively, we could regard the ambiguity of the sentence to be a result of the de re and de dicto distinction, for the disambiguation of which Montague introduces a clarifying quantifier fixing “an angel” existentially to secure the de re reading. Without a quantification rule of the kind of “there is an angel” the sentence remains de dicto. Importantly, to draw such distinctions as finely as Montague grammar invites us to rely on our ability to imagine distinct situations, one in which an angel is mentally projected as actual and one in which the projection is a mere wish. Our attempts at coming to grips with daily confrontations between suicidal religiosity and secular sobriety demonstrate that such mental scenarios are anything but far-fetched. Whereas in Montague grammar the Fregean externality remains intact, we still can imagine what the signifiers “man” and “woman” embedded in their formal syntax are pointing to, hyperintensional semantics is an attempt at capturing everything within its intensional orbit, including a formally abbreviated ontology. Drawing on the work of Frege, Cresswell50 and especially Tichy, 51 recent proponents of hyperintensional semantics such as Pavel Materna and Marie Duzi conceive of natural language as a “set of grammatically well-formed expressions”, each “semantically selfcontained” with its “syntax and semantics”. Its “syntactic string of characters is a way of how to encode meaning”. In turn, “the meaning of a NL expression” is defined as “an abstract procedure encoded by the syntactic part of the expression. If this procedure is executed in a specific world-time, the product is the entity (if any) beyond the language denoted by the expression”. 52 In a paper entitled “Parmenides Principle (The Analysis of Aboutness)” Materna and Duzi argue that since logical analysis can lead us to intensions, but not to reference as traditionally 50

Max Creswell, “Hyperintensional Logic,” Studia Logica 34, 1 (1975), 25-38. Pavel Tichy, “The Analysis of Natural Language,” From the Logical Point of View, 3 (1994), 42-80. 52 Marie Duzi 2011; personal correspondence. 51

100

Chapter Four

understood, we need an alternative description of reference such that “the reference of an expression in the world W at the time T is the value of the respective intension in W at T”.53 The move to so intensionalize reference has resulted in an elaborate hyperintensional semantics in which the world, traditionally left outside the parameters of formal semantics, is spatially and temporally encoded as intensional ontology. Thus, somewhat ironically, has hyperintensional logic repeated the incorporation of reference within a metalanguage, a move for which Saussure was reprimanded by analytical philosophers.54 Hyperintensional semantics would be difficult to reconcile with a semantics based on the principle of imaginability, where language is viewed as “a set of social instructions for imagining, and acting in, a world”, while meaning is “a process of activating arbitrary, empty linguistic signifiers by motivated signifieds, consisting of mental iconic materials regulated by concepts according to directionality, quantity, kind, and degree of schematization, under the community control of the linguistic linkage compulsion”. Nor would the difference between the two approaches be resolved by an argument in favour of a division of labour between formal semantics and pragmatics. On the assumption that imaginability is indeed the deep ground from which natural language draws its infinite resources, hyperintensional semantics looks very much like a formal game testable within its own logical parameters rather than in terms of the salience of imaginable resemblance relations. It is one thing for a semantic theory to be radically intensional, transparent and logically coherent,55 quite another to be able to capture richly what is actually going on in live language. Natural language meaning is presupposed in hyperintensional semantics, and with it what has been eliminated: imaginability.

53

Pavel Materna and Marie Duzi, “Parmenides Principle (The Analysis of Aboutness),” Philosophia 32 (2005), 155-180; 161. 54 Devitt and Sterelny, Language and Reality, 215f. 55 Emma Borg resumes Grice’s “what is said by a sentence” by drawing an even sharper distinction between sentence meaning and speaker meaning, perhaps on the assumption that a sentence can speak itself and so mean on its own; in Minimal Semantics (Oxford: Clarendon Press, 2004). However, as argued here throughout, without voice no meaning and so no natural language. Borg’s as other minimalist approaches describe language as if it were a formal system.

CHAPTER FIVE A CRITIQUE OF MEANING AS USE

Introduction There is hardly a more intriguing case of barring imaginability from linguistic meaning than the one presented by Ludwig Wittgenstein in his Philosophical Investigations. And yet imaginability kept haunting his later thinking on how natural language works. That much is evident in his observation that “there is a lack of clarity about the role of imaginability (Vorstellbarkeit) in our investigation … namely about the extent to which it ensures that a sentence (Satz) makes sense.1 The doubts intimated here temper the standard accounts of what has been referred to as Wittgenstein’s meaning scepticism, according to which we should give up looking for meaning in language and instead embrace the concept of meaning as “use”. Drawing on our earlier definitions throughout the book, in this chapter I want to challenge the Wittgensteinean version of externalist semantics by suggesting that many of the very sentences employed in the Philosophical Investigations do not satisfy the outward criteria stipulated by its author. In particular, Wittgenstein’s rule to look at how an expression is used to grasp its meaning runs into difficulties when he asks the reader to imagine something. A request such as “imagine a picture representing a boxer in a particular fighting stance …” (PI §22n) renders Wittgenstein’s insistence on observables problematic. I do not know what to look for here in gauging meaning as “use”. The mental 1

Ludwig Wittgenstein, Philosophical Investigations, trans. G.E.M. Anscombe, P.M.S. Hacker and Joachim Schulte (Oxford: Blackwell, 2009), §395; Where required I have returned to the German original, as for instance here by rendering the standard English “proposition” as “sentence”, which in my view more accurately reflects what Wittgenstein is indicating in the context of this paragraph. On the problematic of translation of Wittgenstein’s German writings cf. Dinda Gorlée, Wittgenstein in Translation: Exploring Semiotic Signatures (Berlin: de Gruyter Mouton, 2012); see also my review of the book “Wittgenstein and the Complexities of Semio-Translation,” Nordic Wittgenstein Review 2, 1 (2013), 186195.

102

Chapter Five

events required for imagining escape Wittgenstein’s rule that “an “inner process” stands in need of outward criteria”. (PI §580) What is fascinating about such cases is that they are ubiquitous in the Philosophical Investigations. One of the themes in this chapter therefore is an exploration of the tension that exists in Wittgenstein’s masterwork between the ubiquity of the use of imaginability and his refusal to grant it a place in the definition of linguistic meaning. If meaning is use in Wittgenstein’s public sense and Vorstellung, understood as iconic mental projection, looms as large as it does in the Philosophical Investigations and other manuscripts, then the use he makes of Vorstellung is an important language-game with a special relationship to meaning. For instance, the imaginability game is employed in such propositional attitude phrases as “Ich koennte mir leicht vorstellen …” (I could easily imagine), “Wir koennen uns nicht denken …” (we cannot think of), “We can imagine … ,” “suppose …” and dozens of similar formulations. How important a role Vorstellung actually plays in the PI becomes clear when we consider what would happen if we were to eliminate the author’s use of Vorstellung and all its derivates from this text. The PI would hardly be recognizable. And without paying attention to Vorstellung we may very well be led, as Wittgenstein was himself in the Tractatus, “to think that if anyone utters as sentence and means or understands it he is operating a calculus according to definite rules”. (PI §81) That this cannot be the case follows from the observation that natural language is not a symbolic system but more accurately described as a heterosemiotic hybrid, a combination of symbolic and iconic components, a hypothesis to which I commit myself throughout this book. Here, I will develop my argument in six steps. First I address Wittgenstein’s notion of Abrichtung, the way society trains its members to use language according to community norms. A second step briefly addresses the central notion of “use”. Third, the chapter explores what Wittgenstein has to say about Vorstellung. Fourth, I ask how private our mental projections really are when they are associated with linguistic signifiers. I do so from the perspective of Wittgenstein’s “private language” argument. Fifth, I propose to replace the notion of “mental image” by the more general notions of “mental schema” and degree and manner of schematization. Lastly, I will discuss the kind of “assertability conditions” that govern the language game of Vorstellung and the aboutness of language. The chapter concludes with a proposal for resolving the tension between Wittgenstein’s definition of meaning as public “use” and his own ubiquitous use of imaginability.

A Critique of Meaning as Use

103

Abrichtung Language pedagogy, I would say, amounts to making individual speakers competent in the task of linking the arbitrary signifiers of linguistic expressions with appropriate mental, nonverbal scenarios functioning as motivated signifieds. This is why when we hear or read a sentence such as “The engagement between Mary and Johnny is off” we cannot but imagine, in a schematic manner, the kind of social situation the sentence is about. To repeat the book’s central mantra, “if I can imagine what you are talking about and the way you do so, there is meaning; if not, there is not”. The guidance the speech community so exerts upon its members can be regarded as a complex network of constraints, which I sum up under the term sufficient semiosis.2 Wittgenstein, however, does not accommodate the communal regulation of Vorstellung in this way. Instead he deals with the training of the speakers of a language under the concept of Abrichtung, a notion which deserves special attention. To grasp the meaning of Abrichtung we follow Wittgenstein’s rule of looking at how he uses the term. Early in the Philosophical Investigations he tells us that when we teach children their mother tongue this process is not an “explanation” but rather “ein Abrichten,” a strict form of training, (§5) consisting to an important extent in a combination of ostension, directing a child’s attention to what is being pointed out, while at the same time uttering a word. (§6) An example of Abrichtung is to teach the pupil the technique of reading charts, (§86) another, the teaching of algebra. (§§146; 151f.; 189; 320) Wittgenstein also invokes such imaginary scenarios as using “human beings, or creatures of some other kind” being “abgerichtet” (strictly trained) as reading machines”. A person so trained, says Wittgenstein, has not primarily changed his mental disposition, but his behaviour. (§157f.) According to Wittgenstein, the result of such training is not well explained by reference to any individual performance. Bringing into play the formula “y = x2,” Wittgenstein demonstrates that anyone trained in the correct application of the formula will come up with identical results. Comparing different kinds of formulae and diverse kinds of uses will reveal the result of “different kinds of Abrichtung”. Likewise, Wittgenstein suggests, the meaning of natural language terms, as in the use of the word “determine”. (§189) Generalizing, he asks what sort of relation exists between the expression of a rule, a signpost as it were, and my actions. The answer he 2

Horst Ruthrof, Semantics and the Body: Meaning from Frege to the Postmodern, 48f.; The Body in Language, 140-150. For a summary statement here, see Chapter 10.

104

Chapter Five

gives is that “I have been trained to react in a specific way to this sign, and this is how I now react”. Once we have acquired the “grammar of the expression ‘to follow a rule’, such reactions turn into “Gepflogenheiten (Gebräuche, Institutionen)”, that is, habits (customs, institutions). (§198) That he conceives of the phrase “to follow a rule” in a highly strict manner is made clear a little later when he compares it with obeying “an order”. We have no choice in the matter because we have been abgerichtet to react to it in a “bestimmter Weise,” a phrase usually rendered as “in a particular way”. What gets lost in translation here is that the German “bestimmt” also suggests the sense of “determined,” an emphasis that shores up Wittgenstein’s obvious intention to make the notion of rulefollowing more rigorous a requirement than the broad concept of “following” suggests. (§206) The rhetorical question what would happen if someone were to respond one way, another person in a different way, is answered a few paragraphs later where he insists that the rule “always tells us the same, and we do ’what it says’. To a person that has been so abgerichtet we could say: ‘see I always do the same thing: I …’”. (§223) To be sure, Wittgenstein’s analogies between natural language and algebra are problematic. While the sense of determined compulsion applies to both it does so in two quite different senses. In the formal signs, rule-following is restricted to syntax which not only regulates permissible sequences but also the definition of sense. This is not the case in natural language the meanings of which are not primarily regulated by syntax but by nonverbal aboutness. In other words, Wittgenstein’s algebraic analogy, like his frequent analogies with chess, is misleading to the extent that it conflates a formal, intensional sign system with an iconically extensional system, natural language. Nonetheless, in habitual speech semantic uptake appears to occur roughly at the same speed as needed for equivalent algebraic performance. 3 To that extent, then, his analogy holds, even if we are dealing with somewhat different kinds of Abrichtung. Different Abrichtungen, says Wittgenstein, lead to different types of “techniques,” such as for “Wunschäusserungen” (expressions of wishes), (§441) or in the disciplines of sport and in chemistry. (§630) Lastly, in the “Philosophy of Psychology” Wittgenstein considers the case of persons abgerichtet in such a way that on seeing something red they would utter a certain sound and on seeing something yellow another one, and so on. Such training would however not amount to a description of objects by 3

Cf. the discussion of shutter-speed comprehension in Chapter Eight.

A Critique of Meaning as Use

105

their colour, even if this could assist us in producing a description. What is missing is “a representation of a distribution in a space (for example, of time)”. (PI PPF 70) I am inclined here to add that usually, there is neither a real space nor a real passing of time, nor a real distribution within such spaces. In typical linguistic situations we merely imagine such circumstances. And we do so at astonishingly high speeds. So successful is linguistic Abrichtung that we are able to comprehend small units of meaning in an incredible short span of time, with an upper limit of 250ms. As Wittgenstein likes to refer to this phenomenon, we understand an expression “at a stroke” or “in a flash”. (PI §§139; 191; 318; 319; 339) That the themes of Abrichtung and high-speed comprehension are linked in the Philosophical Investigations is of course not surprising. There can be little doubt that Wittgenstein sees a strong causal nexus between the two. Only if we are abgerichtet are we able to “understand the meaning of a word when we hear it or say it “mit einem Schlage” (at a stroke; in a flash, PI §191). But now Wittgenstein draws an important distinction between “what we grasp in this fashion” and “the “use” which is “extended in time”. (§138) He asks “how can what is present to us in an instant, what comes before our mind in an instant, fit a use?” (§139) In the subsequent paragraphs Wittgenstein dissociates the two. We should ask, however, why it is not possible that nothing comes to the native speaker’s mind on the cue of a word sound. Wittgenstein does not address this crucial question. Yet, the idea of semantic uptake “at a stroke” in relation to his notion of “use” continues to dog his investigation. Can’t the use – in a certain sense - be grasped at a stroke? And in what sense can’t it? – It is indeed as if we could “grasp it at a stroke” in a much more direct sense. (§191)

Unfortunately we lack a “Vorbild (example, model) for this übermässige Tatsache” (excessive fact). Because all we have is “the result of the crossing of images,” we are “seduced into using an Überausdruck (a metaterm; a super-expression) … a philosophical superlative”. (§§191f.) It is telling, I think, that in spite of such difficulties, Wittgenstein refuses to let go of what he termed “this excessive fact”. A few paragraphs later he reiterates, “it’s as if we could grasp the whole use of a word at a stroke”. And at the end of the same paragraph he poses the question “where is the connection effected between the sense of the words “Let’s play a game of chess” and “all the rules of the game?” His answer refers to three things: the index of rules; pedagogy; and practice. (§197) There is little doubt, then, that the concept of Abrichtung in the Philosophical Investigations

106

Chapter Five

serves the purpose of drawing our attention to the relentless character of the pedagogy by which the speech community enforces the rules that naturalize language as meaningful for native speakers. In turn, native speakers are to be conceived as persons abgerichtet to strictly follow those rules. This relation has been addressed in the literature under the heading of the normativity of use.4

Use At the centre of Wittgenstein’s theorisation of natural language as a public phenomenon we find the definition of meaning as “use,” (PI §30, 43, 138) as for example in his statement that the “Gebrauchsarten (kinds of use) of ‘understanding’ make up its meaning”. (§532) Given the parlous state of natural language semantics it should not be surprising that this externalist, pragmatic notion of meaning has become so very influential. But in order to render meaning observable, Wittgenstein strips it of any possible mental, private component, instead focusing on the way the speakers of a language engage publicly in a variety of overlapping Sprachspiele or languagegames. By this move he is able to equate meaning with use. “The meaning of a word is its use in the language,” at least in most cases, as he qualifies. (PI §43) As powerful an intervention in the longstanding debate about an adequate natural language semantics and pragmatics this has been, it has one major shortcoming. While it describes well what we see and so satisfies scientific demands for evidence, it covers up features of natural language that are not so readily identified, yet likewise play a necessary role in the meaning process. This is a suspicion which underlies the arguments developed here. If we stay with Wittgenstein’s position on meaning as “use,” the tentative proposal for meaning as the community guided activation of signifiers by 4

See Saul Kripke, Wittgenstein on Rules and Private Language (Oxford, Basil Blackwell, 1982). Cf. also Paul Boghossian, “The Normativity of Content,” Philosophical Issues 13 (2003), 31-45; Anandi Hattiangadi, Oughts and Thoughts: Rule-Following and the Normativity of Content (Oxford: Oxford University Press, 2007) James Connelly, “Meaning Is Normative: A Response to Hattiangadi,” Acta Analytica, 27, 1 (2012), 55-71; Kathrin Glüer and Asa Wikforss, “Against Content Normativity,” Mind, 118 (2009) 31-70; A. Wikforss, “Semantic Normativity,” Philosophical Studies 102 (2001), 203-226; Alan Millar, “Thoughts, Oughts and the Conceptual Primacy of Belief,” Analysis 68 (2008), 234-238; Daniel Whiting, “On Epistemic Conceptions of Meaning: Use, Meaning and Normativity,” European Journal of Philosophy, 17, 3 (2009), 416-434.

A Critique of Meaning as Use

107

nonverbal mental materials does not go through. We must then either abandon such a scheme or modify the Wittgensteinian picture to let mental acts in through the backdoor, but not without redefining their character. To choose the latter course is to significantly broaden Wittgenstein’s notion of “use”. I will attempt to do so by employing his own strategy which he mounts against the possibility of a private language. But first, a brief look at the motivation behind his arguments in the Philosophical Investigations. In the Tractatus Logico-Philosophicus,5 which he strongly qualified in his later work, Wittgenstein had regarded propositions as Bilder for possible states of affairs. (TLP 4.01) The states of affairs so depicted are the sense of a proposition. This postulates a relation between reality and picture described as “pictorial form” (TLP 2.151) and “logical form” (TLP 2.18). Furthermore, language as the “totality of propositions” (TLP 4.001) presents pictures which can be either true or false. (TLP 2.21) Here Wittgenstein neatly combined Aristotle’s double description of language as resemblance relations, homoiomata, and propositions. As Michael Beany observes, in the Philosophical Investigations Wittgenstein has performed a turn from the “simples” of the Tractatus to “samples”. 6 In the Philosophical Investigations the simples that are supposed to make up language are rejected in favour of Sprachspiele out of which every natural language is composed. Propositions are no longer regarded as the definitionally ruled common denominator of language (PI §65) and replaced by looser forms of relations, “family resemblances” (PI §65) and “language-games” (PI §7, 22, 24, 47, 49, 53, 65, 77, etc.) This results in a radical complication of the relation between language and what it is about, its aboutness. What is retained, however, is the public nature of meaning. Just as the operations of logic and the senses of their propositions are open to inspection, so too are language-games publicly displayed, much as any other public communal performance. Wittgenstein now makes two critical interventions. One is to stand the Augustinian “naming theory” of language on its head by saying that names are the result rather than a condition of language-games. (PI §49) The other is a Frege-style reduction, the elimination of imaginability as a central component of linguistic meaning.

5 Ludwig Wittgenstein, Tractatus Logico-Philosophicus (London: Routledge and Kegan, 1966). 6 Michael Beaney, “Wittgenstein on Language: From Simples to Samples,” in The Oxford Handbook of Philosophy of Language, ed. Ernest Lepore and Barry C. Smith (Oxford: Clarendon, 2006), 40-59.

108

Chapter Five

Vorstellung In spite of certain moments of hesitation, at the core of the Philosophical Investigations we find the conviction that “it is not the purpose of the words to evoke Vorstellungen. (It may, of course, be discovered that that helps to attain the actual purpose)”. (PI §6) Here, Wittgenstein refers back to §2 in which he characterises a “primitive” language in which Vorstellung appears to do important work, but only covers part of what we now call a language. As his arguments unfold, it is logical grammar, rules, the community of speakers, Abrichtung and justification that are stressed as more important ingredients. After all, he writes, “nothing is more wrong-headed than calling meaning a mental activity”. The question is whether what Wittgenstein calls the Vorstellungsklavier, the keyboard of the imagination, merely produces incidental tones in the meaning symphony of natural language, whether such tones are a constant side effect, or whether the melodies so created are an indispensable and necessary feature of linguistic meaning. As Wittgenstein phrases it, “uttering a word is very much like striking a key on the piano of our imagination”. (PI §6) If one of the first two answers proves to be correct, then the definition of meaning as observable use stands confirmed. However, if the last answer is more appropriate, then both meaning and use have to be redefined to include the very mental acts Wittgenstein set out to eliminate. Or, rather, we may have to agree with his elimination of Vorstellung as mental “image” as too specific a candidate for meaning, only to replace it by something else at a more general level. In either case, what we must retain is Wittgenstein’s insight that meaning is best described as a process rather than a property. Any attempt at redefining Wittgenstein’s “use” faces the hurdle of his rejection of Vorstellung as a mainstay of natural language. Another difficulty is the bugbear of the translation into English of the term Vorstellung itself, confounded by its ubiquitous use throughout the Philosophical Investigations. “Vorstellung is not an image, but an image may correspond to it,” depending on the kind of Vorstellung we are talking about, one might add. (PI §300) Or again later, “Vorstellung must be more like its object than an image”. (PI §389) When Wittgenstein writes of the Vorstellungsbild, “image” will do. But more often than not the translation of Vorstellung as “image” muddies the theoretical waters, especially when it comes to gustatory, thermal, olfactory, auditory, gravitational and other non-visual, iconic schemata. This is why I think it is preferable to leave the original term untranslated and allow readers to gradually get the

A Critique of Meaning as Use

109

author’s intended meaning by getting accustomed to his “use” of the term in context, and perhaps even by means of appropriate Vorstellungen. A case in point is Wittgenstein’s phrase “motorische Vorstellung” which is rendered meaningful by entertaining mental, quasi-visual and kinetic schemata. In order to foreground public use as the assertability condition for meaning, Wittgenstein is trying to ween us off the habit of associating language with imaginability. This is why he speaks of Vorstellung as “foggily drawn scenes” which cannot “explain intention”, pointing out that “we communicate with other people without knowing if they have this experience too”.7 In a strict sense of “knowing,” Wittgenstein is of course right. What he wants to get away from, as did Frege and the early Husserl, is the psychological contamination of the process of meaning. “’I meant this by that word’ is a statement which is differently used from one about an affection of the soul” (Affektion der Seele; Aristotle’s pathemata). (PI §676) However, can we not have a much more robust conception of Vorstellung? Within a speech community, do we not take it on trust that its members have more or less the same Vorstellungen as we do? Moreover, Wittgenstein’s objection could, with much the same justification, be made to seeing, hearing, smelling, feeling, or tasting. Sometimes, we ask “Can you smell what I smell?” on the assumption that the other person shares roughly the same olfactory apparatus and interpretations, which can be verbally confirmed. May not the same go for linguistic aboutness and speech intention in Vorstellung? In his efforts of demoting imaginability as an important ingredient of language, Wittgenstein also speaks of “image-mongery” (Vorstellerei), which “is of no interest to us”. (PI §390) He cautions against a primitive view of language and is adamant that the process involved in Vorstellung tends to be a distortion (PI §368) inclined to introduce absurd scenarios, such as imagining a parrot whom God has given the gift of inner speech. To this Wittgenstein objects by saying that first I have to imagine a god, then an unlikely situation. Vorstellung is thus heaped on top of Vorstellung. (PI §346) One could of course intervene here by pointing out that there is no difference between Vorstellung and language in terms of their fictional potential. Indeed, in both perceptual and fictional usage, language and Vorstellung appear to run parallel with one another. What Wittgenstein is reluctant to concede is any form of interiorisation. Yet there also hovers a 7

Wittgenstein, Philosophical Investigations, Part II, vi.

110

Chapter Five

more general query above the Philosophical Investigation and its demotion of Vorstellung in that the discursive strategies employed interact in such a manner that Wittgenstein’s critique cannot be taken as radically as it usually has. Not only is the text a “perpetual dialectic” between the author, his “imaginary interlocutor,” and the reader,8 the reader must also deal throughout with a multiple and complicated imaginary scenario. We must follow the author’s instructions to “imagine,” “suppose,” “think of” a variety of mental scenes, and construct two personae, the author and the author’s less insightful alter ego, and their thoughts, quite apart from our own interpretive negotiations. In all this, Vorstellung proves an indispensable tool; so much so that we could say that for every Sprachspiel there is a corresponding Vorstellungsspiel. Not everything that Wittgenstein wrote in the Tractatus is wiped out in the later writings. Rather, the Tractatus is now seen as a highly specialized and so limited language-game of logic that cannot be taken as a basis on which to describe the more complicated relations characteristic of natural language. For example, the idea that “If I know an object I also know all its possible occurrences in states of affairs” is compatible with Wittgenstein’s tendency in the Philosophical Investigations to invite the reader to “imagine” this or that scenario as a reliable guide to understanding.9 Except that now it is an imaginative process rather than propositions that does the work. So frequently is this device employed that without it Wittgenstein’s very philosophical style of inquiry could not get off the ground. Certainly and importantly, Vorstellungsspiele play as much a role in the book as do Sprachspiele. Phrases such as “We can imagine …,” ““It is easy to imagine …,” “Think how …,” “Now imagine …,” or “suppose …” ask the reader to conjure up and participate in imaginary variations of possible scenarios, an exercise that is at the very heart of his philosophising. He not only takes it for granted that we can entertain possible occurrences in our Vorstellung, he takes such advantage of this capacity that one may wonder whether this is merely a methodological device he found more convenient than to point to actual perceptual situations, or whether Vorstellung is perhaps more deeply implicated in language than his rejection of mental representations as components of meaning suggests. Yet perhaps the gulf between these two alternatives is not as wide as one might think. If so, we may ask how the reader typically fulfils Wittgenstein’s instructions in replicating the 8 9

Kripke, Wittgenstein on Rules and Private Language, 3. Wittgenstein, Tractatus, 2.0123.

A Critique of Meaning as Use

111

iconicity of his Vorstellungsspiel. Should we take this process as a special case of “following a rule”? Certainly, we cannot help but play with variations of possible perceptual and quasi-perceptual scenarios such that our mental replication of the colour sepia in our as well as in other people’s Vorstellungsklavier is conceived as reasonably compatible. (PI §6) We can do this because we are in a position to “free” our initial experience and subsequent cases from their singularity. So how do we read the Philosophical Investigations in terms of its author’s foregrounding of Vorstellung as a methodological device for analysis and dismissal as a legitimate component of linguistic meaning? There may very well be a contradiction at the heart of the Philosophical Investigations. On the one hand, Vorstellungen are dismissed as private and not belonging to language, on the other hand, they appear indispensable in the process of communication, in which language plays an essential role. Vorstellungspiele, for Wittgenstein, occupy a central place in our mental life and, just like Sprachspiele, are “part of an activity, or a Lebensform”. (PI §23) The tight relation which I am arguing here exists between Vorstellung and language seems to be assumed by Wittgenstein in this passage. “When we forget which colour this is the name of, it loses its meaning for us; that is, we are no longer able to play a particular language-game with it. And the situation then is comparable with that in which we have lost a paradigm which was an instrument of our language”. (PI §57) What we have lost here is the Vorstellung which we have been trained via communal pedagogy to activate in response to the word sound “red”. On this important point I think Baker and Hacker’s analysis is unsatisfactory when they argue that Wittgenstein merely uses this as a possible objection to his position, while rejecting the relationship between simples of language and simples of Vorstellung. 10 Would the situation change if we were to replace the simples and samples by two parallel systems, those of language and the Vorstellungswelt? Surely, if there were no semantic relation between the two, any association between “red” and our sharable nonverbal, mental grasp of the colour would be spurious or at best incidental. This strikes me as most unlikely. If a speaker lost her Vorstellung of the colour in connection with the signifier she could be retrained by the community. If the community as a whole dropped the shared mental schematization of “red,” its meaning would indeed be gone, as is the case with lots of forgotten combinations of 10

G.P. Baker and M.S. Hacker, An Analytical Commentary on Wittgenstein’s Philosophical Investigations, v.1 (Oxford: Basil Blackwell, 1983), 117.

112

Chapter Five

signifiers and their activation by communal, conceptual regulations of iconicity, which we call “motivated signifieds”. So, either Wittgenstein is wrong or Baker and Hacker’s interpretation is. Certainly, the kind of systemic relation between Vorstellung and language that Wittgenstein urges the reader to perform cannot be empty. If it is, that is, if our Vorstellung fails us, then meaning fails. And since meaning in the Philosophical Investigation is the use of a particular language-game, if Vorstellung fails, we are at the same time incapacitated to play the language-game. For playing a language-game to be meaningful cannot be like playing chess, which lacks both aboutness and modality, in the form of voice. Language-games are always about something and change when their intonation, that is, our providing them with a certain tone, changes. Now the merely incidental connection between mental act and word has been seriously compromised. The two main reasons why Wittgenstein did not pursue the significance of imaginability in this respect seem to me his behaviourist convictions and a residual, unacknowledged Fregean assumption of the private nature of mental events. It is to the second of these likely motives that the next section addresses itself, with a double emphasis on the alleged privacy of mental representations and the question whether Wittgenstein has perhaps eliminated such phenomena too readily.

How private are mental scenarios? If we compare the German originals of Wittgenstein’s “Vorstellung” and “Bild” and the English texts of the Philosophical Investigations, we find that Vorstellung and vorstellen are invariably translated as “image” and “imagine,” a practice which undermines his point that “a Vorstellung is not an image, but an image may correspond to it”. (PI §301; my emphasis) This suggests that his objection to regarding mental acts as components of linguistic meaning is actually stronger than it appears in the translation. Not only are the specifics of mental images discarded, but also the more general level of mental activity of imaginability of which images are a part. Vorstellungen, as we remember, also cover olfactory, thermal, gustatory, and other non-visual mental projections. It seems to me important that in order to make any headway on the question of Vorstellung, we need to ask how private Vorstellungen really are. I try to offer an answer by leaning on Kripke in invoking some of Wittgenstein’s own arguments in his critique of the possibility of a “private language”. Roughly following Kripke’s summary, the “private language” argument looks as follows. We can accept that our language expresses concepts

A Critique of Meaning as Use

113

which, once acquired, we are justified in applying. Yet, Wittgenstein insists, an “inner process” as in conceptualisation “stands in need of outward criteria” (PI §580). We can find such outward criteria in “circumstances, observable in the behaviour of an individual, which, when present, would lead others to agree with his avowals”.11 This is why we must look at circumstances under which speakers employ a languagegame in such a way that they can be said to have followed the relevant rules. In Kripke’s account, Wittgenstein’s theory of natural language is “one of assertability conditions”.12 The “private language” claim then fails on the grounds that following a rule privately violates the very concept of a rule as something any person could follow. And since a language is by definition rule governed and so communicable, such a “private” language could be private only de facto, but not in principle. What then about a person who uses “private” words “to refer to what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language”? (PI §243) Since here too, as Baker and Hacker show, “linguistic competence (understanding a language) is mastery of a rule-governed technique,” to speak of a “private language” is a contradiction in terms. 13 Or, in P.F. Strawson’s words, “the idea of a language of any kind used only by one person was an absurdity”. 14 The proof of the pudding is provided in Bill Bryson’s account of “Boontling, or harpin boont”, which started as an attempt at creating a private language, but got out of hand and for forty years survived as a local dialect in Boonville, north of San Francisco.15 Yet all this leaves us with a dilemma as we resume our discussion of mental images. On the one hand, mental projections are said to be private; on the other, they appear to play a big part in social interaction. This dilemma could be resolved if one were able to show that mental images are likewise subject to rules, even if such rules may prove a little more obstinate to identify than those of language. In an imaginary dialogue between P. and Q., John Wisdom has P. open the conversation by saying “When you say ‘How blue the sky is’ you know what you mean but I

11

Kripke, Wittgenstein on Rules and Private Language, 100. Ibid., 110. 13 Baker and Hacker, An Analytical Commentary, 267. 14 P.F. Strawson, “Pain and Persons,” in The Private Language Argument, ed. O.R. Jones (London: Macmillan, 1971), 27-32; 28. 15 Bill Bryson, Mother Tongue: The English Language (London: Penguin Books, 2008), 232f. 12

114

Chapter Five

don’t know and can’t know what you mean and nor can anyone else”.16 To any ordinary English speaker this sort of statement looks nothing short of ludicrous. Of course, we roughly know what P. means in the sense of understanding what he is saying. Even a colour blind person has become attuned to taking on trust that whenever someone utters this sort of sentence, she understands what the words amount to pragmatically and accepts the statement on condition that the way she sees the sky is consistent with the typical way people with normal vision say they see it. In other words, the sentence “How blue the sky is” is regarded by members of a speech community as standard verbal communication which is made meaningful by way of the appropriate response in and by imaginability. What is being said is imaginable. We can imagine what is being talked about. We can also imagine if the manner of speaking contradicts aboutness, for instance when the sentence is being uttered during a rain storm. This suggests that Vorstellung is not as private as it is assumed to be in the literature and very much for the same reasons that Wittgenstein invokes for his position on the necessarily public nature of language. In this respect, the most pervasive of nonverbal, mental grammars is the televisual screen on which prefabricated combinations of language and mental imagines appear as a never ending digital stream. As a result of such pedagogy, the viewer must make an effort not to reproduce mentally the televisual images of blue helmeted soldiers on the verbal cue of “United Nations peacekeepers”. Such is the power of the repetition of nonverbal typifications. In addition to cultural constraints on Vorstellung, there are biological constraints starting with the pre-given ratio of about 1 million visual input cells compared to 100 billion neurons turning the incoming information into human perception and its imaginable modifications. Roughly the same ratio applies to the perceptual realisations initiated by our other four senses. This suggests that on the side of consciousness we can assume a similar ratio between a single Wahrnehmungswelt, a single world of perception and, potentially at least, an infinite number of Vorstellungswelten, of imaginable worlds. Nor is it likely that this asymmetry is restricted to a mere quantitative difference. Since we are able to vary perceptions infinitely, at least in principle, in Vorstellung, the resulting variations are likely to be different also in kind. After all, Wittgenstein’s Vorstellungsklavier is able to replicate as well as 16

John Wisdom, “Wittgenstein on ‘Private Language’,” in Wittgenstein: Philosophy and Language, ed. A. Ambrose and M. Lazerowitz (London: George Allen and Unwin, 1973), 26-36; 27.

A Critique of Meaning as Use

115

dramatically distort perceptual input. None of these variations, though, as we have suggested, should be assumed to be any less structured than language.17 To this extent, then, mental acts are not as free or private as we have been led to believe but are both culturally and neurologically constrained. If we accept this correction, then Wittgenstein’s notion of following a rule applies also to Vorstellung. This paves the road for the claim that the combinations between words and their mental projections can be understood as fairly well established patterns which are neurally transmitted, culturally modified via pedagogy and reinforced by semiotic practice to form a pivotal part of our Vorstellungswelt. The same, cum grano salis, holds for nonverbal signification. As David Crystal notes, “the communicative value of tactile activities is usually fairly clear within a culture,” that is, also in our Vorstellungswelt.18 This does not mean that members of a speech community have all been trained into producing such combinations in identical fashion. Nor is strict synonymy a necessary condition for communication, nonverbal or verbal. And yet communication occurs, because both competence and performance are embedded in sufficient semiosis. We understand the language and our fellow language users sufficiently to guarantee successful linguistic exchange. Roughly the same observations can be made about the competence and performance in the domain of Vorstellung and mental images. Each member of a cultural Vorstellungswelt shares, produces, and communicates with the help of mental projections because they are similar enough; they suffice. This makes mental acts intersubjective rather than private. Imaginability has been de-privatised.

From image to schematization It is now time to return to Wittgenstein’s insistence that “a Vorstellung is not an image, but an image may correspond to it”. (PI §301) Wittgenstein here employs the ordinary use of the term Vorstellung indicating a level of generality of which mental images may be specific instantiations. Olfactory memories, for instance, are thus subsumed under Vorstellung, though not as images. A further step on the ladder of the generalization of mental acts is to build bits of Kant’s schematism into our argument. As we 17

I will resume this point in Chapter Eight from the perspective of brain research, theories of consciousness, and language evolution. 18 David Crystal, How Language Works (London: Penguin Books, 2008), 6.

116

Chapter Five

saw in Chapter Two, Kant improved on Locke’s private ideas as the content of linguistic meaning by introducing the concept of mental schemata and the monogram, which he stipulated as a necessary condition of our ability to conceive of mental scenarios. This basic level of abstraction must, however, not be confused with formalisation of the kind that produces formulae such as x = yĶ. While the latter lacks aboutness and voice, as ingredients of linguistic meaning Kant’s non-formal schemata retain both. Otherwise, a cartoon, for example, could not be read as funny, offensive, or vicious. Graded, schematized Vorstellung, I suggest, is the most likely way by which iconic, mental material reappears as the quasi-perceptual content regulated by concepts in natural language. This accounts, for example, for the difference between the generality of “I said baking, not deep frying” and the relative specificity of “baking a cake” and “baking a potato”.19 How do we do this? We have been taught to associate habitually two very different kinds of things: (1) imaginable schemata of the experience of different kinds of baking and (2) their relevant linguistic expressions. Now the crucial question is this. How does meaning as use in Wittgenstein’s sense handle this situation when we are not involved in an actual situation in a kitchen but merely report or discuss such scenarios? How, in other words, does the idea of meaning as a process of public rule following handle talk about merely imagined scenarios? According to Wittgenstein, the meaning of “bake” in all three instances is their use, a behaviourist externalisation and abbreviation of a complex process. However, in order to understand the three situations “in a flash”, in the absence of any external evidence, it is the Vorstellungswelt rather than the immediate Wahrnehmungswelt that we must rely on. This is where semantic externalisation runs aground. In the bulk of language use we rely on mental reconstructions and approximations in and by way of Vorstellung, as indeed Wittgenstein does in his dialogue with his readers. This places a big question mark on the reduction of meaning to “use” in Wittgenstein’s public sense. Certainly, if it worked, the externalisation of meaning would permit public scrutiny of language use by way of justification. On the other hand, Vorstellung escapes the public eye. So Wittgenstein asks, “how can I justify that I form this Vorstellung in response to these words?” (PI §382) We cannot accept 19

James Pustejovsky, “The Generative Lexicon,” In Semantics: A Reader, ed. S. Davies and B.S. Gillon (Oxford: Oxford University Press, 2004), 369-393.

A Critique of Meaning as Use

117

the description of Vorstellung as “testimony,” because it is really only a description of what someone is “inclined to say”. And “how does one point to a Vorstellung?” (PI §386) Nor do we have any criterion for judging whether two Vorstellungen could be called the same. “What is the criterion for the identity (Gleichheit) of two Vorstellungen?” (PI §377) After all, he argues, the word “same” is justified only if someone else can also regard it as justified. (PI §378) Yet it is not only Vorstellung that is thus placed under scrutiny. In keeping with these criticisms, Wittgenstein also takes observation sentences to task. “To the private transition from what is seen to the word, I could not apply any rules” for “the institution of their use is lacking”. (PI §380) The only way out for both Vorstellung and perception, then, is the assumption of a culturally shared mental world. Is Wittgenstein then guilty of a performative paradox when on the one hand he insists on the public definition of meaning as use and, on the other, consistently employs imaginable scenarios? I suggest that the paradox can be resolved by arguing that there is indeed an institution that regulates not only the system of language but also our Vorstellungswelt. For it would be odd to assume that the speech community were in charge only of verbal communication and had nothing to do with the internalised culture of which all native speakers partake of necessity. And if the speech community is as crucial in this context as Wittgenstein rightly maintains, is Saul Kripke’s emphasis on “assertability conditions” a good way of specifying the notion of meaning as use? To explore this question is the purpose of the last section of this chapter.

Assertability, aboutness and voice Kripke’s explanation why he regards Wittgenstein’s theory as “one of assertability conditions” is that the speech community “can assert of any individual that he follows a rule if he passes the test for rule following applied to any member of the community”.20 What kind of test then is the community applying in the following case? “Lying is a language-game that needs to be learned like any other,” says Wittgenstein. (PI §249) How do we decide whether an utterance is the use of the language-game of lying, understood as a deliberate misrepresentation of what is the case? Take a sentence such as “Jane promised him to be kind”. This looks straightforward enough. If she had not intended to be kind but said she would be, it was a lie. But what sort of conditions would allow us to check whether the appropriate rules had been followed? Typically, perception 20

Kripke, Wittgenstein on Rules and Private Language, 110.

118

Chapter Five

does not provide the required “assertability conditions”. For if we have to wait for evidence as to whether Jane’s behaviour should be called “kind” or “unkind”, we have to wait a while, which cannot possibly be the way language works. After all we understand the sentence “in a flash”. Or, if we emphasize the notion of “conditions”, then imaginative processes must kick in immediately as a hypothetical play of likely mental scenarios. Without some such assumption, meaning as “use” remains stuck in its syntactic straightjacket, without content or, to use Putnam’s metaphor, with no ropes to anchor the balloon. This suggests that Wittgenstein’s notion of meaning as use is the public application of intensionally defined signifieds within the constraints of language-games. One may be tempted to read the Philosophical Investigations in this way by the author’s frequent alignment of natural language with examples from arithmetic, geometry, and chess, a move that would result in a vicious circularity that severs the vital connection between Wahrnehmungswelt and Vorstellungswelt on the one hand, and language on the other. (PI §31, 33, 197,205,337) To get to the aboutness of language we must be able to connect the syntactically ordered, empty signifiers, their intensional relations, with a domain outside of language, which requires an extensional connection. For unlike formal systems, language is not a syntactically defined, self-sustained sign practice. Wittgenstein himself supplies the crucial clue. “Suppose,” he writes, ”I said ‘a b c d’ and meant: the weather is fine … What is supposed to be the criterion for my having that experience?” (PI §509) Simply, the reason why “a b c d” do not grant us the experience of “the weather is fine” is that the community has not sanctioned this particular combination of sound sequence and mental, iconic aboutness. Furthermore, the letters “a b c d” by themselves have already been customised to be employed in different circumstances. This means that meaning as use must somehow cater not only for the public display of linguistic competence in justifiable performance but also for the way the speech community has pre-arranged what counts as the appropriate mentally recoupable aboutness of linguistic expressions. We often use the interrogative “What are you talking about?” We ask this question when we are not sure what a linguistic expression was about. Or, what the person speaking to us was trying to point to by means of language; or what I was supposed to imagine as a result of her words. In such cases, we expect a rephrasing for clarification. Yet it is not the rephrasing itself that is the purpose of our indirect speech act. Rephrasing

A Critique of Meaning as Use

119

is only the necessary intermediary step for the actual purpose of the original statement to be understood. What we are after here is the aboutness of language. The other side of linguistic meaning is what we have characterized by voice, the manner of speaking, or speech modality. Both aboutness and voice are fundamental components of meaning as use and yet both require imaginative reconstruction on the part of the listener or reader. On the problematic of modality, Wittgenstein introduces this nonverbal example: “I see a picture which represents a smiling face. What do I do if I take the smile now as a kind one, now as malicious? … Thus I might supply the picture with the Vorstellung that the smiler was smiling down on a child at play, or again on the suffering of an enemy”. (PI §539) What Wittgenstein has done is to provide alternative enunciative modalities. Applied to language, deictic variations of propositional content are a vital ingredient of the process of meaning constitution, especially when we are dealing with contexts requiring the construction of implicit deixis, just as in the nonverbal semiosis imagined here by Wittgenstein. Indeed, in natural language, deictic modification is fundamental because it is potentially always able to override propositional content.21 If so, and since both aboutness and modality, or voice, require mental reconstructions, none of Wittgenstein’s interpretive moves could have taken place without Vorstellung. What then are the “assertability conditions” of Wittgenstein’s use of the Vorstellungsspiel? What gives him the confidence that when he writes “Imagine a picture representing a boxer in a particular stance” that the reader will be able to apply the kind of community sanctioned rules in her Vorstellung that would guarantee Wittgenstein’s pedagogic intention to be carried out? (PI §22n) He must assume that all his readers will be able to imagine the boxer in typical enough fashion to produce sufficiently similar Vorstellungen. But how could this be asserted? Only on the assumption that imaginability, no matter at what level and in what manner of schematization, is rule governed and therefore neither private nor merely incidental. And since Vorstellung is neither objectively public nor private, our best predicate is that it is intersubjectively mental. This allows us to distinguish the objectively public features of language, such as the signifiers in a dictionary or in a public speech, from such intersubjectively shared and so indirectly public instantiations of the connections we have

21 Horst Ruthrof, Pandora and Occam: On the Limits of Language and Literature (Bloomington: Indiana University Press, 1992), 4f.

120

Chapter Five

been trained to perform between language and Vorstellungswelt as meanings.

Conclusion By way of conclusion, I offer a tentative solution to Wittgenstein’s dilemma in a number of steps. (1) Natural language is not like chess; only its syntax is. (2) As a semiotic hybrid, language combines the symbolicity of arbitrary signifiers with the iconicity of motivated signifieds which receive their content from mentally projected aboutness and voice, or modality in the broad sense. (3) Natural language requires imaginability for meaning since the bulk of use is about things and events in absentia. (4) The association of the sounds of linguistic expressions with their appropriate meaning endowing mental projections is not private but rule governed and monitored by the speech community. In light of these remarks and in agreement with the book’s imaginability thesis, I propose the following redefinition of Wittgenstein’s description of meaning as “use”. Linguistic meaning as use is the activation, under community rules, of arbitrary, empty signifiers, combined into language-games, by nonverbal, iconic mental materials in and as Vorstellung, regulated by concepts in terms of directionality, kind, quantity, and degree as well as manner of schematization.22

22

For a detailed exploration of conceptual regulation of mental iconic materials in the motivated signified, see Chapter Ten.

CHAPTER SIX THE LINGUISTIC ROUTE: SAUSSURE AND VYGOTSKY

Introduction In this chapter I want to pay attention to two seminal contributions made to our understanding of language and linguistic meaning by two theorists: the father of modern linguistics, Ferdinand de Saussure, and the psychologist of language acquisition, Lev Vygotsky. As in the previous chapters, my analysis will be informed by what I have termed the imaginability thesis, according to which language can be viewed as “a set of social instructions for imagining, and acting, in a world”. Without imaginability, understood both as a human capacity and a feature of motivated signifieds, I have argued, we would not be able to render the arbitrary sounds of linguistic expressions meaningful. I have further insisted that the habitual association between signifiers and signifieds is best explained by way of an intersubjective mentalism, governed by the rules of the speech community. In light of this position, the chapter asks how the mainstays of Saussure’s linguistics and Vygotsky’s developmental psychology square with these definitional parameters. In revisiting Saussure’s Cours de linguistique générale in the first section of the chapter, I will avoid the usual overview and instead concentrate on two specific areas which, from the position taken here, look problematic: his assumption that without language we would live in a foggy and so chaotic world and his radical arbitrariness thesis of the linguistic sign.1 The focus of the second section will be the question of how Vygotsky, in his Collected Papers, argues the transformation of meaning from nonverbal thought to the conceptual thought of the mature native speaker.2 1

Ferdinand de Saussure, Cours de linguistique générale, ed. Charles Bally, Albert Séchehaye, and Albert Riedlinger (Paris: Payot, 2005). 2 Lev S. Vygotsky, Thought and Language (Cambridge: Mass., MIT Press, 1991); The Collected Works of L.S. Vygotsky, v.1, Problems of General Psychology, ed. R.W. Rieber and A.S. Carton (New York and London: Plenum Press, 1987); The

122

Chapter Six

Linguists who believe that they have resolved the problems identified in Ferdinand de Saussure’s Cours de linguistique général (2005; [1916]) and philosophers who think that Saussure is irrelevant to the concerns of the philosophy of language are both mistaken. Reading Saussure’s foundational linguistics from the perspective of a broadly conceived philosophy of language means in the first instance to address some of the most basic principles involved in the study of natural language. The fact that this text was compiled posthumously from Saussure’s notes is hardly an obstacle to such an undertaking since none of his surviving manuscripts contradicts those leading principles. 3 In particular, the point that Saussure had to distance himself from historical linguistics in order to pursue the core characteristics of language in general remains beyond doubt.4 Certainly, the Cours as it has come down to us clearly reflects three important steps in Saussure’s thinking as documented in his public lectures: language as system instantiated in the mind (1907); the notion of the linguistic sign as part of semiologie (1908-9); and the internal mechanism of the linguistic sign (1910-11).5 In quoting Saussure, I will stay close to the original text, providing English paraphrases where necessary. Historically, the Saussurean projects stands very much on its own. At its heart we find, among other seminal insights, the bare bones of a theory of meaning, crystallised in his definition of the linguistic sign as an arbitrary relation between the publically available lexicon of a natural language on the one hand and, on the other, the Vorstellungen which we typically entertain in the process of understanding. I will substitute throughout the Collected Works of L.S. Vygotsky, v. 2, The Fundamentals of Defectology, ed. R.W. Rieber and A.S. Carton (New York and London: Plenum Press, 1987). 3 Ferdinand de Saussure, Linguistik und Semiologie. Notizen aus dem Nachlass. Texte, Briefe und Dokumentation, ed. Johannes Fehr (Frankfurt: Suhrkamp, 1997); Robert Godel, “Notes inédites de F. de Saussure”, Cahiers Ferdinand de Saussure 12 (1954), 49-71; “Nouveaux documents saussuriens”, Cahiers Ferdinand de Saussure 16, (1958-9), 23-32; Simon Bouquet and Rudolf Engler, Ecrits de linguistique général (Paris: Gallimand, 2002); Rudolf Engler, “The Making of the Cours de linguistique général,” in The Cambridge Companion to Saussure (Cambridge: Cambridge University Press, 2004), 47-58. 4 E.g., Bouquet and Engler, Ecrits, Introduction. 5 Eisuke Komatsu and George Wolf, Premier Cours de linguistique général (1907) d’après les cahiers d’Albert Riedlinger (Oxford: Pergamon,1996); Eisuke Komatsu and George Wolf, Deuxième Cours de linguistique général (1908-1909) (Oxford: Pergamon, 1997); Eisuke Komatsu and Roy Harris, Troisième Cours de linguistique général (1910-1911), d’après les cahiers d’Emile Constantin (Oxford: Pergamon, 1993).

The Linguistic Route: Saussure and Vygotsky

123

term Vorstellung for Saussure’s two terms idée 6 and concept 7 on the grounds that in the Cours they are deployed indiscriminately and that Vorstellung defined as “mental transformation of perception” seems to me to capture best the text’s “intention”. Saussure places the relation between linguistic expression (word, phrase, sentence) and Vorstellung (idée; concept) within a general ‘sémiologie,” a theory of signs in which language occupies a special niche. From this perspective, language can be compared with nonverbal systems at the level of the sign viewed as a combination of a signifiant, or signifier, and a signifié, or signified. For natural language, this semiological unity Saussure calls the “linguistic sign,” characterised as an “arbitrary” combination of its components. If it turns out that Saussure is correct in his assumption that both signifier and signified are indeed arbitrary or unmotivated, then his conclusion that the linguistic sign as a whole must be arbitrary goes through. If, however, it can be demonstrated that one of the two components of the linguistic sign is motivated, then his radical thesis of arbitrariness collapses. I will further argue that Saussure’s assumptions of the chaotic nature of idée, concept or Vorstellung and the arbitrariness of the signified are mistaken and, as a result, the linguistic sign is in need of re-definition.

Radical arbitrariness At the centre of Saussure’s systemic conception of language as langue, distinguished from langage as language in action and parole as individual speech event, we find his definition of the linguistic sign, which differs markedly from the theorisation of language in the Fregean tradition. Unlike Frege, Saussure does not allocate the notion of the sign to linguistic expressions. Rather, the linguistic sign is made up of two components, a sound image (l’image acoustique) and a Vorstellung (concept, idée). Only when these two components come together has a linguistic sign been realised. As Saussure insists, une suite de sons n’est linguistique que si elle est le support d’une idée (144). At the same time and as a result of the incorporation of this Vorstellung within the linguistic sign, l’idée now reappears as a “qualité de la substance phonetique” (144f.) These two elements must however not be confused with a name and a thing. (98) They are both part of a mental process, an “association psychique” (28) together constituting “une entité psychique”. (99) That is, the linguistic 6

Saussure, Cours, e.g., 33, 47, 104, 144, 155, 166. In order to stay close to the original phrasing all page numbers in the text refer to the 2005 edition of the Cours. 7 Saussure, Cours, e.g., 28f., 31, 98f., 144f., 158f.

124

Chapter Six

sign is instantiated in the mind. Saussure conceives of this process as a “circuit”. (31) Once this circuit is completed the linguistic sign is severed from the origins of one its components, the “faits de conscience” (28), “idées” (33, 47, 144, 155, 166), “concepts” (28f., 31, 98f., 144f., 158f.) or Vorstellungen, understood here as “mental transformations of perception”. To avoid the charge of mentalism, the independence of the linguistic sign from the origins of and in Vorstellung is summed up in the conviction that natural language, viewed as langue and just like its formal cousins, “est une forme et non une substance”. (169; 157) Importantly, the linguistic relation between sound image and Vorstellung turns out to be only a special case within a much broader theoretical frame, Saussure’s sémiologie, a general theory of signs in which language occupies centre stage. “On peut donc concevoir une science qui étudie la vie des signes au sein de la vie sociale”. (33) Since the study of language, Saussure predicts, is a branch of such a new science, the laws which semiology is going to discover will apply to language. In anticipation of such laws, then, Saussure formulates the principles which govern the specifics of the linguistic sign by generalising its two components. The sound image is covered by the notion of the signifier (signifiant), the Vorstellung by that of the signified (signifié). (97ff.) It would appear that by invoking Vorstellung (idée, concept) as the content of the signified and so as an essential component of the linguistic sign Saussure revived an already discredited Lockean notion. 8 After all, Frege had eliminated Vorstellung from his description of Sinn and Bedeutung in 1892 on the grounds that its subjective ingredients disqualified it from playing a role in any logical account of natural language semantics. 9 And could one not argue that since Saussure’s central aim was to put the study of language on a proper scientific footing he undermined his own goal by giving mental processes such prominence in his theory? There is a deep rift in the theorisation of language between approaches shunning psychological questions and a resurgence of theories in which mental processes once more play an indispensable role. On the one hand we have truth-conditional theories of a Fregean persuasion 10 or truth-

8

Cf. Chapter One. Cf. Chapter Four. 10 David Wiggins, “Meaning, Truth-Conditions, Propositions: Frege’s Doctrine of Sense Retrieved, Resumed and Redeployed in Light of Certain Recent Criticisms,” Dialectica 46 (1992), 61-90. Cf. Chapter Four. 9

The Linguistic Route: Saussure and Vygotsky

125

conditional arguments in the wake of Tarski’s “Convention-T”11 and the tradition of meaning scepticism best represented by the description of meaning as use”; 12 on the other hand there is a burgeoning literature insisting that without consideration of mental facts we can capture only part of what is going on in language. 13 Suffice it to say here that the philosophy of language is not in a strong position today to discredit Saussure’s psychological emphasis. Nor are the real problems of the Saussurean account to be sought in his description of the linguistic sign as mentally instantiated. Saussure’s flaws lie elsewhere. What does require careful scrutiny is the widely accepted idea of the radical arbitrariness of the linguistic sign. “Le signe linguistique est arbitraire”. (100) Such is the stark claim of the arbitrariness of the linguistic sign, an assertion so readily accepted both in linguistics and the philosophy of language that Saussure could confidently write “le principe de l”arbitraire du signe n”est contesté par personne”. (100) Saussure’s contention is radical in the sense that we are dealing here with four different kinds of arbitrariness: (1) of the signifier; (2) the signified; (3) the bond between signifier and signified; and (4) the sign as a whole. Leaving to one side a number of concessions Saussure makes to qualify his arbitrariness thesis, which however do not fundamentally alter his central description of the linguistic sign, I make arbitrariness and his characterization of Vorstellung my central concerns. In what is arguably the most thorough defense to date of Saussure’s conception of the systematicity of language, or langue, Paul Thibault’s Rereading Saussure: The Dynamics of Signs in Social Life (1997), the author places the accent rightly on the “intrinsic systemic” rather than any extrinsic criteria for the definition of the linguistic sign. 14 Accordingly, what constitutes meaning is the differential relation between, for example, “making love” and “giving a lecture” within a “cultural system”. No need, then, to involve a “separate and pre-existing world”; all we require is a value system of “contextualizing relations”. We may well ask, “values 11

Donald Davidson, “Truth and Meaning,” in Semantics: A Reader, ed. S. Davis and B. S. Gillon (Oxford: Oxford University Press, 2004), 222-233; cf. Chapter Four. 12 Ludwig Wittgenstein, Philosophical Investigations, trans. G.E.M. Anscombe, P.M.S. Hacker and Joachim Schulte (Oxford: Blackwell, 2009). Cf. Chapter Five. 13 See the discussion in Chapter Seven of cognitive linguistics and especially such writers as Hurford 2007; Jackendoff 2002; Corballis 2002; Johnson and Lakoff 1999; Gallese and Lakoff 2005; Fauconnier and Turner 2002. 14 Paul Thibault, Re-reading Saussure: The Dynamics of Signs in Social Life (London and New York: Routledge, 1997), 164fff.

126

Chapter Six

about what?” Or are these values akin to those of empty place-holders like x or y in a formal, definitional context? Not quite. As Thibault concedes, the intrinsic linguistic system of values is used, after all, for “categorizing the phenomena of the world”. A threatening naturalism, so carefully avoided by Saussure as Thibault, is sneaking in through the back door, only to be denied once more by the assertion that “the semiotic relation between language and what exists outside it has no naturalistic basis”. And it is Saussure’s insistence on a radical form of arbitrariness that “designates the fact that the categories belonging to the system are functionally organized on the basis of relations of similarity and difference among the categorical terms that constitute the system”. The missing link here is imaginability, which facilitates aboutness modified by the reconstruction of speech modalities and so bridges the gap between the linguistic system and its nonverbal outside. For neither “making love” nor “giving a lecture” are linguistic facts. While such activities typically involve speech; as acts they are anything but linguistic, they are heterosemiotic social events. Without this extrinsic frame, the differential relations between the linguistic expressions reduce to a difference between two arbitrary strings of sounds. One serious consequence of Saussure’s theorization in terms of langue is that it undermines the very core of natural language: social meaning. It would seem, then, that only if we restore Saussure’s signified to the status of the motivated component of the linguistic sign and so seriously qualify the centre piece of his theory, are we able to argue meaning in a way that complies with what actually occurs in language as speech.

The nebulous world of Vorstellung in Saussure A commitment crucial for Saussure’s radical arbitrariness is his belief that “nôtre pensée n’est qu’une masse amorphe et indistinct”. Without language, we could not distinguish two Vorstellungen (idées) clearly and consistently. In fact, on its own, our thinking is nothing but a nebulous cloud, “une nébuleuse,” in which nothing is delimited of necessity. (155) This passage, while uncontroversial at the beginning of the twentieth century when evolutionary biology had scarcely become a respectable science, stands in sharp contrast to what we are now learning about our animal relatives and the lives of hominids several millions of years prior to the invention of language. Tigers, known for their precision hunting, could not have survived if they lived in Saussure’s “royaume flottant”; nor could our pre-linguistic forebears, who for millennia produced precision instruments out of carefully selected materials. What is not in dispute here

The Linguistic Route: Saussure and Vygotsky

127

is the weaker thesis that language constitutes a significant refinement of nonverbal distinctions, a point to be taken up later. Suffice it to say here that in languaging modern humans linguistic concepts massively override the sort of primitive concepts which permit the tiger’s hunting skills. But this addresses only one side of the motivated signified, its conceptual regulation. What this study is foregrounding are the mental iconic materials that are so regulated.15 Given Saussure’s conviction that our thought processes are “chaotic” by nature unless they are organised by language and his endeavour to challenge historical linguistics by a synchronic, scientifically generalised approach, it is not surprising that he debunked the study of the evolution of language.16 (156) The question of the origin of language, says Saussure, should not be posed at all. (105) After all, in the study of language there is nothing to be found that looks like “natural data”. (116) But it is one thing to say that language is a refinement of Vorstellungen in the sense that it transforms raw Vorstellungen into a series of fine-grain linguistic concepts and quite another to say that without language everything is nebulous. (cf. 26) There is now a growing body of studies aiming to review this matter.17 While the jury is still out on what precisely we mean by linguistic refinement, it looks as if Saussure’s strong assumption of the vagueness of Vorstellungen may be no more than a prejudice. Another objection to Saussure’s foggy world of Vorstellung can be raised from measurements of the human ability to make thousands of olfactory distinctions in the face of the stark poverty of olfactory differentiations in the lexica of our natural languages. 18 Not everything we can clearly distinguish is equally differentiated by language, which is precisely the obverse of Saussure’s claim. Now the boot is on the other foot; it is language that is vague rather than perception and its mental derivatives in Vorstellung.

Verbal and nonverbal semiosis When a new term is created in the world of techno-logos, such as “buzzer,” or in the class-room, such as “rank,” what comes first, language or a new 15

See Chapter Ten for a summary of redefinitions of the central terms of a semantics of imaginability. 16 See Chapter Eight for a brief discussion of evolutionary linguistics. 17 E.g. James R. Hurford, The Origin of Meaning (Oxford: Oxford University Press, 2007). 18 Diane Ackerman , A Natural History of the Senses (New York: Vintage Books, 1991).

128

Chapter Six

experience in need of a linguistic expression? Once the new lexicon is in place both appear to come as a package. Speech communities introduce neologisms to express new nuances of experience and we grasp their conceptualisation when we hear or read the novel term. We are socialized into meaning as “use”. Barring the red herring of onomatopoeia, the terms chosen by culture in such situations are indeed arbitrary in Saussure’s sense. But are the Vorstellungen coded by the new terms equally arbitrary? And more specifically, are they not only arbitrary but also “chaotic” with respect to our Wahrnehmungswelt and its imaginative variation, our Vorstellungswelt? If we accept that Vorstellung is neurally tied to perception, as cognitive science has assured us for some time that it is, then Vorstellung likewise would be systemic but not as a result of linguistic relations. Isn’t it much more likely that both perception and its mental modifications in Vorstellung are always already systemic? After all, both stand in a systemic rather than chaotic relation with the world. Another angle from which we can view Saussure’s nebulous world of nonverbal idées is realist semantic externalism of the kind introduced by Frege about a decade before Saussure’s lectures. Here perception provides the anchor for deciding not only meaning but also a notion of meaning tied to the truth of observation. From Frege’s link between the truth values of true/false and the notion of meaning to Moritz Schlick’s verification procedure and the Tarskian truth-conditional arguments mounted by Donald Davidson, there is the fundamental assumption that perception itself is a consistently reliable measure by which to judge the sentences of natural language.19 What separates Saussure’s linguistics sharply from the analytical tradition is that the former cannot accept the demarcation the latter draws between perception and language. According to Saussure and his followers, perception (as well as Vorstellung) is always already linguistically framed. Once we have acquired language we can no longer step outside its orbit and act as if we were non-linguistic beings, a claim that deprives structuralist linguistics of reference in Frege’s sense. In Saussure, reference is intralinguistic. An alternative to both Saussure’s and Frege’s position would be to declare reference intersemiotic.20 Nothing, not even reference, then can escape the formal orbit of the Saussurean linguistic sign and so the idealised circularity of langue as a whole. Yet it is not enough to reject the Saussurean account on the 19 20

For a discussion of some of the issues involved here, see Chapter Four. Cf. Ruthrof, The Body in Language, 56f.

The Linguistic Route: Saussure and Vygotsky

129

grounds that it “omits reference”. 21 In structural linguistics reference is still present, albeit as an internalised relation amongst linguistic signs. What requires emphasis is the absence of independent nonverbal phenomena. Yet in Saussure’s closed-circuit verbal world there are no such things as “non-linguistic facts”. Before and outside of language, gravity is no more than a “nébuleuse”. But having named it for example diqiu yinli (earth globe pulling force), as the Chinese do, the mist has lifted. The critique of Saussure’s fog of Vorstellung also finds support in recent evolutionary biology. If it is the case that human intrinsic intentionality is the result of a very long evolutionary history of the cellular division of labour, the question of the systematicity of perception and its modifications in Vorstellung can be laid to rest for good. Cells responsible for perceptual input are said to have been increasingly matched, from amoebae to humans, by cells monitoring such input for the benefit of the survival of the organism, to the point where the monitoring capacity of neurons far outstrips the input capacity of cells responsible for perception. This scientific narrative provides a strong case for the view that both perception and its mental variants have been systemically tied to the objectivities of the world long before language was invented.22 One tentative and yet important conclusion we can draw from all of this is that it is very likely that Vorstellung is by no means random or arbitrary, but rather a systemic reworking of perception.23 If this is so, and there is no evidence to my mind to suggest otherwise, then we must review the Saussurean order. Instead of saying that both signifiers and the Vorstellungen we associate with them are arbitrary, we should say that while verbal signifiers are arbitrary, Vorstellungen and therefore signifieds are not. And if the thesis goes through that the latter are systemically related to perception as transformations, then the Saussurean relation must be reversed. It is not language that orders Vorstellungen, at least not in the first instance. Rather, the opposite appears to be the case. Perception and Vorstellung impose order on the arbitrary signifiers of language. In other words, while Vorstellung is both motivated and systemic, language now turns out to be fundamentally characterised by derivative systematicity or 21 Michael Devitt and Kim Sterelny, Language and Reality: An Introduction to the Philosophy of Language (Oxford: basil Blackwell, 1991), 213. 22 Cf., for instance, the seminal paper by Tecumseh Fitch, “Nano-Intentionality: A Defense of Intrinsic Intentionality,” Biology and Philosophy 23 (2008), 157-177; although Fitch does not extend his analysis to the domain of mental projections. 23 As argued in Chapter Five from the angle of Wittgenstein’s “private language” argument.

130

Chapter Six

derived motivation. At least, this seems to hold for perception sentences and their relatives. Having said this, there is of course a sense in which language calls the shots, as it were. Much of our institutional world appears to be created by language rather than by perception and its mental variants, a perspective sympathetic to Wittgenstein’s vista of language as “refinement”.24 What we have described here speculatively is of course only part of the story, though it remains a foundational relation between words and Vorstellung. As language evolved, it increasingly sharpened its inherited nonverbal distinctions. So much so that significant portions of our lexica and grammars would appear to support the priority Saussure grants the verbal. However far from conceding that the derivative systematicity or derived motivation of language is a thing of the distant past, it should be regarded as the continuing rock bottom of verbal signification. This I take to be the gist of Putnam’s metaphor of the balloon of language tied forever, even if only sporadically noticeable, to the ground of our Wahrnehmungswelt. One might add to Putnam’s observation that though perception provides the main stays that prevent the balloon from drifting, it is Vorstellung in schematised form that provides language with meaning every step of the way. So we could modify Putnam by adding to his “widely scattered and very thin (but all-important) ropes” that tether language to the actual world something like “and millions of fine threads tying language to Vorstellung”.25 This, I suggest, is reflected not only in the lexicon of natural languages, but also in their grammatical structures. As to the lexicon, changing community feelings of justice, for example, are reflected in, as much as they are created by, a forever more specific legal vocabulary. The rights of individuals in general are specified to distinguish the rights of gays, lesbians, mothers, children, and the unborn. All forms of the splitting of concepts constitute refinements produced with the help of language. But such refinement is not restricted to the lexicon; in a certain sense it applies also to grammar. I think that Derek Bickerton is wrong in his emphasis on the revolutionary moment when language as we know it was born as a

24

Ludwig Wittgenstein, “Cause and Effect: Intuitive Awareness,” Philosophia 6 (1976), 3f. 25 Hilary Putnam, Mind, Language and Reality: Philosophical Papers, vol. 2 (London: Cambridge University Press, 1979), 4f.

The Linguistic Route: Saussure and Vygotsky

131

result of the invention of syntax.26 I suggest on the contrary that syntax in its elaborate forms, such as conditional structures, evolved on top of a perceptual protosyntax which structurally reflected the way the world was realised by hominids. For it is most likely that after millennia of perceptual and gestural pre-linguistic hominid social existence, language would have inherited and then refined such relations and their modifications in Vorstellung, which had served our ancestor so well for so long. There is no logical barrier to thinking that thought sequences like if, then or only if x, then y were not already available as a subterranean logic of perceptual reasoning. If language can indeed be regarded as a refinement of perceptual and gestural being, then Saussure’s radical dissociation of the iconicity of pre-linguistic grasp from the linguistic signified looks unlikely. Given such speculative assumptions, what precisely is being “refined” here? We could say that language specifies primitive, perceptual concepts, which we share with non-languaging animals, by an ongoing series of conceptual subdivisions of resemblance relations, or iconicity. This rests on the definitional choice of declaring concepts mental regulatory social mechanisms delimiting perceptual iconicity and its neural transformations. In the visual domain, this process can be readily demonstrated. On the other hand, as the verbal gyrations on wine labels testify, in the domain of olfactory and gustatory representations language is hardly a match for the nonverbal nuances we are able to register. From this basis one could attempt a neural rehabilitation of Kant’s definition of the concept as “a rule of the synthesis of perception” and a “predicate of possible judgment,” while at the same time avoid the unbridgeable chasm that separates Saussure’s linguistic sign from whatever came before. 27 Speculative as such thoughts must remain, they find summary support in Ludwig Wittgenstein’s remark “Language – I want to say – is a refinement, im Anfang war die Tat”.28

26

Derek Bickerton, Roots of Language (Ann Arbor: Karoma, 1981); Language and Species (Chicago: University of Chicago Press, 1990); Language and Human Behaviour (Seattle: University of Washington Press, 1995); cf. the recent Evolutionary Linguistics (Cambridge: Cambridge University Press, 2012) by April McMahon and Robert McMahon. 27 Immanuel Kant, Critique of Pure Reason, (KrV A722/B750) and (KrV A69/B94), respectively. 28 Ludwig Wittgenstein, “Cause and Effect: Intuitive Awareness,” Philosophia 6 (1976), 3f.

132

Chapter Six

Wittgenstein’s bon mot suggests that natural language has grown in relation to and as an economising reflection of our socio-perceptual world and its modifications in Vorstellung. As such, language takes the relation between its arbitrarily chosen signifiers not from a formal system of axioms, definitions and syntactic rules, but from the translation of nonverbal realisations of materials, events, and feelings. Language is therefore an a posteriori constituted sign system. First comes speech as social practice, and then comes its systemic formulation in lexicons and grammar books. This difference is driven home when we apply the formal notion of recursivity, an exercise which sooner or later runs into the wall of idiomatic expressions. Saussure’s langue can be no more than an a posteriori construction on top of language, an abstractive, derivative system of explanation rather than an axiomatic foundation. An allimportant consequence of this is that the structure of formal systems does not reflect a non-linguistic order but merely measures it according to its own terms, while natural language not only “measures” but also, in its very make-up, mirrors a human perceptual order. This order can be expressed as an intersemiotic integration of a multi-modal, nonverbal set of heterogeneous signs, such as olfactory, tactile, gustatory, gravitational, aural, emotional, visual and other readings of the world, a point to be resumed a little later. At a yet more primitive level this interpretive integration performed by the human organism draws fundamentally on physico-chemical and bio-chemical processes, including those of our own physiognomy. In this sense, all natural languages are deeply motivated; they are most certainly not arbitrary. To opt for a “chaotic” and “foggy” world of perception and its infinite variations in Vorstellung in order to be able to argue the arbitrariness of the linguistic sign goes against the grain of the basic principles of even the most modest kinds of naturalism. Language, it would seem, has inherited its base motivation and resemblance relations from perception and its mental variants. And yet, this kind of motivation cannot be aligned with a simple correspondence between language and world. The main reason why Alfred Tarski warned his fellow logicians in 1936 not to apply Convention-T to natural language was that its terms are anything but formal.29 But there is a further complication. The very description of natural language highlights a problem of philosophical inquiry: the shakiness of the correspondence theory of truth. “P” iff p suffers from the need to express nonverbal 29

Alfred Tarski, Logic, Semantics, Metamathematics (Oxford: Clarendon Press, 1956 [1936]), 267; see the brief discussion of Convention-T and its application to natural language by Donald Davidson in Chapter Four.

The Linguistic Route: Saussure and Vygotsky

133

phenomena by verbal signs. A hard-headed semiotician, for instance, could not in good conscience allow the unjustified translation between heterosemiotic systems to go unchallenged. Nor are we dealing here with only two systems, that of predication on the one hand and the phenomenal world on the other. The phenomenal world itself is the result of the intersemiotic unification of a number of heterosemiotic systems, the combined result of interpretations of radiation, molecular states, and other physical and bio-chemical inputs. This makes correspondence theories of truth look simplistic. In a way, analytical philosophy and structuralist linguistics are committing similar errors. While the former is insensitive to the fact that their comparison rests on the untheorised transformation of non-linguistic facts into language, the latter has made a virtue of this equivocation by alleging that there is nothing but language. A more appropriate explication would be to say that in natural language, a culture has chosen a verbal signifier, say “tree,” which evokes a generalised and schematised Vorstellung of a tree, derived from perceptual inputs and their mental transformations under social control. The fact that we can translate “tree” by the terms “shu”, “arbre” or “Baum” is often taken as the result of the self-evident possibility of relating different languages to one another as comparable systems. Yet this is not the way it works. Rather, we can translate between Chinese and French strings of verbal signifiers because we share a basic tertium comparationis: our compatible, even if culturally differentiated, nonverbal worlds of perception and Vorstellung. Put more simply, we are able to produce translations because we share culturally inscribed bodies and so generate comparable worlds by means and out of the same stuff. Importantly, this common baseline is primarily realised by way of nonverbal, iconic signs. This is why the position I am advocating here owes a fundamental debt to Charles Sanders Peirce. One of the crucial differences between the semiotics of Charles Sanders Peirce and Ferdinand de Saussure’s sémiologie is their scope. Peirce embraces everything that is an object of interpretation. Accordingly, Peirce’s sign is “something that stands for something in some respect or capacity. It addresses somebody, that is, creates in the mind of that person an equivalent sign, or perhaps a more developed sign ….”30 Whatever we feel we need to decipher so that we can allocate it a place in our perceptual and conceptual world, is part of human semiosis. In contrast, Saussure’s sémiologie is restricted to social, conventional signs, such as symbolic rites, military signals, sign language, 30

Peirce, Collected Papers, (CP 2.228; [1897])

134

Chapter Six

(33) pantomime, scales as a symbol of justice, and forms of greeting. (100f.) The common practice, then, of referring to both theories as semiotics is misleading; it wipes out a significant distinction. I have elsewhere marked this difference by distinguishing “ROSS,” or read-only sign systems, from “COSS,” or communicative sign systems.31 Nor is it self-evident why Saussure would restrict his sémiologie to COSS, since the interpretation of natural phenomena, such as thunder or the scent of flowers, is likewise a social activity and so well within the purview of his domain of “la vie sociale”. (33) Another notable difference between the philosophical semiotics of Peirce and Saussurean sémiologie is that in the former language is one among a vast number of significatory practices, while in the latter one cannot quite avoid the impression of a kind of linguistic imperialism. At the very least, there is a tension in the Cours between sémiologie as a meta-science in which all nonverbal communicative signs, including linguistic ones, are ordered and explicated “par les lois de cette science” (35) and one of its members, language, which will act as its “patron général”. (101) This tension has had methodological consequences, as fostering such research questions as “What kind of language is cinema?” While this is not quite as absurd as asking “What kind of carrot is a potato?” it has led to an entire industry of film criticism discovering Saussurean linguistic relations in visual semiosis.32 Both these differences have a strong bearing on the topic of Saussure as a philosopher of language. For his constraints on the range of sémiologie seriously diminish the most important component of the linguistic sign as a candidate of investigation: idée, concept, or Vorstellung. In contrast, the Peircean approach allows for olfactory, gustatory, kinetic, thermal, gravitational, aural, tactile, and visual readings as heterosemiotic processes by which we, as well as pre-linguistic beings, read the nonverbal world. Furthermore, we have already established the non-congruence of verbal and nonverbal distinctions of which humans are capable and so are in a position to extend the nonverbal scenario to social practice. At the heart of such a post-Peircean picture is the emphasis on resemblance relations or iconicity by which we find our way about in the world.33 If this is so, then Saussure’s radical arbitrariness of the linguistic sign proves an unmanageable hurdle to the theorisation of the relation between nonverbal readings of the world and language. I choose the opposite emphasis. If it 31

Ruthrof, Semantics and the Body, 37. Christian Metz, Language et cinema (Paris: Larousse, 1971). 33 Peirce, CP 1.158. 32

The Linguistic Route: Saussure and Vygotsky

135

can be shown that iconic relations play a significant role in the linguistic sign by providing the signified with content under conceptual regulation, then language must be declared to be parasitic on nonverbal signs. 34 Moreover, if nonverbal signs are motivated and if language cannot function without nonverbal semiosis, then the linguistic sign cannot be arbitrary in the way Saussure has proposed. The radical arbitrariness of structural linguistics fails, a failure resulting from a linguistic anti-realism which would deny the existence of nonverbal facts by the very definition of the linguistic sign. Neither can Saussurean linguistics accommodate cognitive investigations committed to a neurally motivated explanation of perception, Vorstellung and language, nor is the Saussurean project compatible with Peircean semiotics in which iconicity, or resemblance relations, ties language to the world. According to Saussure, what systematicity we do find in language can only be of its own making. To make matters worse, some of Saussure’s post-structuralist followers have abandoned the signified altogether, leaving the signifier to carry semantic load, which is a contradiction in terms even in a generous reading of the Cours. It would seem that extra-linguistic facts, then, must be retrieved via nonverbal semiosis. Thus, perception and imaginability re-enter the picture, not as nebulous clouds but rather as a systemically coherent world into which language gears. One way by which this can be coherently argued can be found in the writings of Lev Vygotsky.

Vygotsky: from non-verbal to conceptual thought Unlike the Fregean tradition up to hyperintensional semantics (Frege 1970 [1892]; Montague 1970; 1974; Duzi et al. 2010), which eliminated imaginability from sense as a remnant of a misleading psychologism, and in contrast to the structuralist weakening of the signified, Lev Vygotsky takes the opposite route by demonstrating the indispensable role that Vorstellung plays in language learning.35 And unlike Wittgenstein, who accounts for Vorstellung only to liken its function to incidental tunes produced on the keyboard of the imagination, Vygotsky shows how verbal 34

Cf. Horst Ruthrof, Semantic of the Body; The Body in Language; “Metasemantics and Imaginability,” Language Sciences 35 (2013), 20-31. 35 Lev S. Vygotsky, Thought and Language (Cambridge, Mass.: MIT Press, 1991); The Collected Works of L.S. Vygotsky, v.1, Problems of General Psychology, ed. R.W. Rieber and A.S. Carton (New York and London: Plenum Press, 1987a); The Collected Works of L.S. Vygotsky, v. 2, The Fundamentals of Defectology, ed. R.W. Rieber and A.S. Carton (New York and London: Plenum Press, 1987b). All references in the text refer to these three editions.

136

Chapter Six

meaning emerges out of imaginability, in phylogenesis as in ontogenesis. And unlike the early Chomsky, whose syntactic focus can be accused of begging the question by taking meaning for granted, in Vygotsky grammar supervenes on meaning. At the centre of his language acquisition program is the umbrella notion of “thought”. “Thought” straddles a broad spectrum of human mental activity from preverbal thinking in hominids and children to full conceptuality in contemporary society. As such, “thought” has a trajectory from nonverbal, mental realizations to verbal, semantic competence. In light of such a spectrum I ask, “What sort of theory of linguistic meaning can we derive from Vygotsky’s genetic method of analysis?” In lieu of an answer to this question I want to test this hypothesis. Vygotsky’s theory of language acquisition allows for, though does not pursue, a solution to Locke’s paradox, that is, how to reconcile the hidden mental event of linguistic meaning with the public character of natural language. Such a reconciliation can be best achieved, as I have argued, by a semantics of imaginability. To demonstrate this claim, I offer a critical, though necessarily selective, overview of Vygotsky’s position, emphasizing imaginability and resemblance relations, Aristotle’s homoiomata. I suggest that despite a number of grey areas in this respect, Vygotsky appears to have made a strong case for resemblance relations surviving in linguistic meaning at least in some minimal form. Otherwise his “concept” would have to be taken as abstracted to ideality or formal emptiness, a reading that would undermine Vygotsky’s entire program. I will not address here the ongoing debates in the wake of Vygotsky’s writings in psychology, 36 in education, 37 on Vygotsky’s “historical-dialectical monism,”38 or in activity theory and the theorization of cultural practices.39 36 Wolf-Michael Roth and Yew-Jin Lee, “Vygotsky’s Neglected Legacy,” Review of Educational Research 77 (2007), 186-232; Dorothy Robbins, Vygotsky’s Psychology-Philosophy (New York: Plenum, 2001); James V. Wertsch, Vygotsky and the Social Formation of Mind (Cambridge, mass.: Harvard University Press, 1985); Alex Kozulin, Vygotsky’s Psychology: A Biography of Ideas (Cambridge, Mass.: Harvard University Press, 1990). 37 Vasily V. Davydov, “The Influence of L.S. Vygotsky on Educational Theory, Research, and Practice”, trans. S.T. Kerr, Educational Researcher 24, 3 (1995), 12-21. 38 Charlotte H. Liu and Robert Matthews, “Vygotsky’s Philosophy: Constructivism and Its Criticisms Examined,” International Education Journal 6, 3 (2005), 386399. 39 A. Sannino, H. Daniels, and K. Guttierrez, Learning and Expanding with Activity Theory (Cambridge: Cambridge University Press, 2009).

The Linguistic Route: Saussure and Vygotsky

137

Instead, my focus will be strictly on the question of the transformation of nonverbal into verbal thought.

The social basis of language For Vygotsky, our “higher psychological functions,” such as language competence, are fundamentally social. Their phylogenesis and ontogenesis are social in origin. Humans did not develop language as part of the evolution of their “bio-type, but through the historical development as a social creature”. Collective human life is at the heart of Vygotsky’s perspective on language, allowing for both “working together” and “individuation”. (1987b: 192) In this picture, the dualism of nature versus nurture is eliminated by absorbing the former in the latter, such that inherited innateness is perfectly compatible with what is “superorganic, conditioned, that is, social”. (1987b: 154f.) Human culture is human nature. Far from being a “natural form of behaviour,” human thought is “determined by a historical-cultural process”. As such, the relation between thinking and speaking is “not a prerequisite for, but rather a product of, the historical development of human consciousness”. (1991: 210). Language is above all “the social means of thought”. (1991: 94) Importantly, Vygotsky’s deeply anchored social gradualism distinguishes his work sharply from such other research paradigms as structuralism and analytical accounts. So adamant is Vygotsky as to the social, evolutionary character of language that he distances himself also from Piaget’s notion of “socialized” speech, “as though it had been something else before becoming social”. (1991: 35) Instead, Vygotsky prefers the term communicative. When the child’s communicative competence is impaired, as in the handicapped child, the inadequacy is explained by Vygotsky as an “incongruence between his psychological structure and the structure of cultural forms”. (1987b: 47) The child’s “intrapsychological” categories find it difficult to match their social, equivalent, “interpsychological” structures. (1987b: 11) This is so because a significant part of the child’s evolving intrapsychological structure is “the dependence of thought on affect”. (1987b: 233) Hence Vygotsky regards the unity of affect and intellect as essential for the development of the “dynamic reasoning system” and a “cornerstone” in developmental research. (1987b: 238f.) At the same time, he qualifies these observations by making the point that the “dependence of thought on feeling” is not a one-sided affair. Nor, he says, are we dealing here with “an object”; rather, it is “a process”. (1987b: 240) In the gradually inward

138

Chapter Six

movement, emotion “escapes peripheral control”.40 Closely related to the relation of affect and thought is Vygotsky’s treatment of the will as one of the three fundamental problems of research, the other two being “the problem of thinking” and “the problem of the imagination”. (1987a: 349; my emphasis) The will is “will to meaning” in the sense that when a child’s associative mental scenarios are disrupted he or she is likely to change the imaginary world by “producing a meaningful rather than meaningless action” and so accommodates and interpretively masters the disturbance. (1987a: 356) Will as the will to meaning fails at both extremes of the spectrum of “maniacal excitement” over “galloping ideas” on the one hand and, on the other, “obsessive ideas” from which the individual cannot free herself. (1987a: 312) The water shed on the way to full linguistic maturity is whether thought remains “a slave to the passions, their servant” or whether thought develops to be “their master”. (1987b: 239)

The nonverbal ingredients of thought When Vygotsky speaks of the heterogeneity of the “complex formation” that is verbal thinking, this heterogeneity is to be sought partly in the character of language as a temporally unfolding sequence of expressions and partly in the sources from which language draws its “meaningful aspect”. (1987a: 320) Even if Vygotsky does not spell this out in full, the sources of “the concreteness of thought” (1987b: 232) appear to involve such heterosemiotic readings of the world as olfactory, gustatory, tactile, visual and other nonverbal interpretations. This is most obvious in what Vygotsky calls the “pre-verbal growth” of the “intellect”. (1987a: 117) As language develops, it “involves the filling of what we say with meaning, the extraction of meaning from what we see, hear, and read”. (1987a: 320; my emphasis) If sight, touch, smell and taste, as well as other perceptual modalities fill the arbitrary and so meaningless sounds of linguistic expressions, then nonverbal thought is pivotal to our description of how natural language works semantically and pragmatically. Following Vygotsky, we can neither say that thought can be identified with speech, as so many of our semantic theories do, nor that they are absolutely different from one another.

40

Jerome Bruner, “Prologue to the English Edition” in The Collected Works of L.S. Vygotsky (1987a), 15.

The Linguistic Route: Saussure and Vygotsky

139

The presence of perception in language is highlighted by Vygotsky when he writes, “we can no longer separate the perception of the object as such from its meaning or sense”. Linguistic meaning results from a merging of nonverbal and verbal mental states. “It is here,” he says, “that the connection between perception and speech, the connection between perception and the word, arises”. (1987a: 299f.) If this is so, then the combination of perception and speech necessitates the transformation of the nonverbal into the verbal rather than the abandonment of the former. In other words, the nonverbal survives in the verbal. The question is in what form it can be argued to do so. Since perception itself cannot fulfil the role of meaning in language, we need to look at its mental transformation in and as Vorstellung, understood here as “mental modification of perception”. Though the “first form of intellectual activity” in the child is “active, practical thinking,” this presupposes the very possibility of its generalizing transformation into an imagined world. Although “the child prefers the real apple to the imaginary one,” it is the imagined apple that will play the crucial role in linguistic meaning. (1987a: 63) For Vygotsky, the imagination is “a psychological system” anchored in the “unification of several functions in unique relationships”. (1987a: 348) In particular, the “imagination is a necessary, integral aspect of realistic thinking,” so much so that “they two act as a unity”. (1987a: 349) In this unity, it makes sense to think that language functions as an economizing resemblance matrix gradually imposed over nonverbal thought. Furthermore, the imagination is essential even for ordinary perception. “No accurate cognition of reality is possible,” observes Vygotsky, “without a certain element of imagination, a certain flight from the immediate, concrete, solitary impressions in which this reality is presented in the elementary act of consciousness”. (1987a: 349) This commits Vygotsky to a feed-back relation not only between perception, Vorstellung and conceptualization, Kant’s dialectic between receptivity and spontaneity, (CPR A50/B74) but also to the necessary interaction between imaginability and language. The former is an inextricable component of the latter, while language lends precision to Vorstellung. Vygotsky’s reasoning in favour of this interaction has considerable force: “A more profound penetration of reality,” he writes, “demands that consciousness attain a freer relationship to the elements of that reality, that consciousness depart from the external and apparent aspect of reality that is given directly in perception”. (1987a: 349) What is paramount in the transformation of perception into imaginability and its so modified role in language is what Vygotsky calls the “perseveration of representation,” the tendency to follow the

140

Chapter Six

momentum of a once chosen sequence of Vorstellungen. (1987a: 311f.) This tendency has the double effect of continuity and directionality. It is the latter that language has inherited from its semiotic precursor systems, and it is the goal directedness of our mental representations that make them productive rather than merely reproductive. As Vygotsky says, they are “goal determined”. (1987a: 124) Pointing preceded speaking by a long shot; it reappears in language as its “indicative” function.

The heterogeneity of the verbal If “verbal thinking is a complex formation that is heterogeneous in nature,” (1987a: 320) and nonverbal thought emerges from a variety of heterosemiotic systems, then it should not come as a surprise that, as Vygotsky notes in “The Genetic Roots of Thought and Speech,” that “progress in thought and progress in speech are not parallel”. Rather, both in phylogeny and ontogeny “their two growth curves cross and recross”. (1991: 68) Gradually, and as a result of this asymmetrical dialectic between the nonverbal and the verbal, “thought does not merely find expression in speech; it finds its reality and form”. (219; my emphasis) Both “reality” and “form” emphasize the public nature of language in the shape of linguistic signifiers. Yet “form” also points us to the idea of “conceptual regulation” which, though mental, also partakes of “publicness,” but in a different sense, a point to be taken up again towards the end of the chapter. Highlighting the all-important heterogeneity of the verbal, Vygotsky argues that in language, “the semantic and phonetic developmental processes are essentially one, precisely because of their opposite directions”. (1991: 220) This seemingly paradoxical statement alerts us to what one could call the double helix of the interaction between the movement of word absorbing nonverbal materials and the movement from nonverbal “meaningful complex” to words that takes place as one integrated process. In light of this dynamic complexity in the process of language acquisition, Vygotsky feels encouraged to write that “words cannot be put on by thought like a ready-made garment”. (1991: 219) Of course, in habitual speech, this genetic complexity is replaced by an almost automated and almost instantaneous process, such that the word indeed appears to be a “ready-made garment” for thought. As I have maintained, it is the “linguistic linkage compulsion” that makes it so. Vygotsky’s argument in favour of the heterogeneity of the verbal is supported also by obvious differences between imaginability and grammar. Psychologically, they are not homogeneous, “there is no direct

The Linguistic Route: Saussure and Vygotsky

141

correspondence between grammatical form and the sense unity it expresses”. (1987a: 322) In spite of this, Vygotsky never loses sight of the parallels between thought and speech. “Not only subject and predicate, but their grammatical gender, case, tense, degree, etc., have their psychological doubles”. (1991: 221) Verbal heterogeneity, together with its parallels in nonverbal thought, makes Vygotsky’s theory an attractive explanatory option. Although thought and speech develop “along separate lines,” they merge “at a certain point”. (1991: 93) The same would probably hold true for phylogeny, except that there the time scale involved must be vastly different. Notions such as the “psychological double” of signifiers, “the independent grammar of thought,” and the “syntax of word meanings” open the door not only to a semantics of imaginability, but also to the tempting idea of a nonverbal protosyntax as a precursor to rudimentary linguistic sequencing in the evolution of natural language. This line of thinking finds support in Vygotsky’s remark that “the seeds of future reflection – an understanding of justification, proof, and so on – are already contained in the most primitive of children’s quarrels”. (1987b: 197) Both ontogenetically and phylogenetically we could speak of protosyntactic precursors of verbal reasoning. If gender, case, tense, subject, and predicate are “mirrored” psychologically, they must exist in both verbal and nonverbal forms, even if we take the view that we cannot separate our nonverbal realizations from their verbal regulation. Certainly, such “mirroring,” if that is what is happening, cannot start from the verbal; there must already be a preverbal experience of something that can be formed into a subject, a predicate, and so on. Otherwise, ostensive teaching would not work. The child will recognize something as a candidate of a linguistic subject in a simple sentence only if she has already grasped that something as an individuated object. Verbal sequences supervene on perceptual order. The syntax of language, then, can be viewed as a refinement of a nonverbal protosyntax.

Concept formation from syncretism to external speech Vygotsky’s experiments have shown “how the use of the word acts as a means of forming the concept, how from syncretic images and connections, complexive thinking, and potential concepts, there arises that unique signifying structure that we may call a concept in the true sense of the word”. (1987a: 166) According to Vygotsky, concept formation in the child proceeds in three stages. Stage one is characterized by the “syncretic image,” a roughly fused heap of objects which are gradually ordered in terms of a “kinship” by “impressions” and spatial “distributions” until they

142

Chapter Six

are reduced to “a single meaning”. (1987a: 135) At an early stage of the child’s development, “word meaning is an incompletely defined, unformed, syncretic coupling of separate objects” fused into a mental singularity. Syncretism in this sense is characterized by “insufficiency of objective connections” which are gradually corroborated in social “practice”. As such, the child’s “syncretic merging of images” corresponds to the adult speaker’s full conceptual control. (1987a: 134) Stage two of concept formation opens with the creation of “family names,” an empirical “complex collection,” “chained complex,” and a “familial unification of things,” typically resolved in the end by a “pseudoconcept”. (1987a: 136ff.) The function of the pseudoconcept is the linking of “complexive and conceptual thinking”. (1987a: 145) This is possible because the complex already “contains the kernel of the future concept that is germinating within it”. (1987a: 146) The nonverbal core of the unified complex, one could say, reappears transformed in conceptual schematization.41 Yet as a mere “shadow of the concept,” the pseudoconcept can do no more than reproduce “its contours,” (1987a: 144) an idea reminiscent of Kant’s description of schematization as general “delineation”.42 Thus Vygotsky views the pseudoconcept as a transitional phase on the way to the “stable and constant word meanings” of the adult speaker. As such, it is part of the “development of generalizations” under the guidance of pedagogy. (1987a: 143) In this process the child is “not free to construct his own complexes”. They are “ready-made,” the child finding them “in completed form” as it begins to understand “unfamiliar speech”. The “form of generalization” at work in the pseudoconcept, however, differs from that operative in the “true concept”. A series of objects are “generalized by a given word”; the child “learns the developed speech of the adults around him”. As Vygotsky insists, “Everything is contained in this statement”. (1987a: 144f.) Stage three, lastly, at the point of the “maturation of the word itself,” (1987a: 241) is the attainment of actual concepts. Although Vygotsky privileges the “scientific concept,” in the form of the “social science concept,” from what he has to say of concept development, we can nonetheless glean the following. (1987a: 239) Amongst the kinds of concepts Vygotsky distinguishes, the “preconcept” of pre-school age, the “pseudoconcept,” the “spontaneous concept,” the 41

Horst Ruthrof, “From Kant’s Monogram to Conceptual Blending,” Philosophy Today 55, 2 (2011), 111-126. 42 Kant, CPR A141/B180.

The Linguistic Route: Saussure and Vygotsky

143

“scientific concept,” and the “everyday concept,” it is the everyday concepts in the domains of “social life” that are the most relevant to the semantics of natural language. Unfortunately he does not say very much about them. (1987a: 240f.) Nevertheless, his statements about scientific concepts permit a number of extrapolations relevant to “everyday concepts”. Concepts are treated as “word meanings” (1987a: 167) which function differently within the domains of everyday life and scientific thinking and according to the “measure of generality” as the place they occupy in a conceptual system and according to “concept equivalence” or the way concepts relate to one another. (227) The former “determines the set of possible operations of thought available for a given concept,” whereas the latter determines the “paths of movement from one concept to the next”. (228; 233) Scientific concepts are grasped within the coordinates of their respective paradigm of generality rather than by the order of perception and Vorstellung. Nevertheless, Vygotsky observes a feed-back relation between scientific concepts and their everyday equivalents in that the practice of the former produces an increase in general conceptual “performance”. (168). In spite of this relation, Vygotsky insists on the following “key difference in the psychological nature” of the two: scientific concepts are always part of a system, while everyday concepts are not. (234) This, I think, is a grave error in Vygotsky’s picture of concepts, as it is in Saussure’s Cours. It is a mistake to assume that the perceptual world and its modification in the Vorstellungswelt should be non-systemic. Nothing is more unlikely. The world coheres for us, as it does for animals, as a nonverbal system in which heterosemiotic readings together produce an intersemiotic system of multiple coordinates. It is against the socially reinforced background of that system that language evolved in the first place and against which the child, according to Vygotsky’s own description, acquires her concepts. Scientific paradigm dependent concepts cannot be taken as the master key for understanding concepts in general. The concepts of chemistry in no way illustrate how the olfactory concept of the scent of a jasmine blossom emerges and functions, though both are systemic rather than random. Like concept formation, language acquisition in the broader sense occurs in three stages, from egocentric speech, through inner speech to external speech. Egocentric speech in Vygotsky is “a primitive method of children’s thinking out loud in a difficult situation” (1987b: 195) and as such “fulfils an intellectual function”. (194) As a first step in the development of inner speech and as the first “autonomous speech function,” egocentric speech gradually turns into “inner speech”. (196) Compared to

144

Chapter Six

external speech, inner speech, or “endophasy,” for Vygotsky is a “distinct plane of verbal thought” (1991: 224; 248) which appears as truncated, “a dynamic, shifting, unstable thing, fluttering between word and thought”. (249) Other traits of inner speech include a “tendency towards predication,” concomitant with the frequent omission of the subject of the sentence, which is typically taken for granted. (236) With a condensation towards predication also comes “decreasing vocalization” to the degree that “inner speech is speech almost without words”. (244) As syntax and word sound recede, meaning as the “influx of sense” comes to the fore, so much so that Vygotsky describes it as “thinking in pure meanings”. (247; 249) While in external speech “thought is embedded in words,” in inner speech “words die as they bring forth thought”. (249) What makes Vygotsky’s elaboration of the three stages both in concept formation and speech forms important for our hypothesis is that it makes it highly unlikely that nonverbal thought, so powerfully present throughout the evolution of language in childhood, should suddenly dissolve into empty abstractions at the end of the process. When Vygotsky speaks of the attainment of the “true concept” of which the word is the bearer (1987a: 145) we need to know what sort and degree of abstraction has been achieved. I will return to this question towards the end of the chapter.

Linguistic meaning What then does Vygotsky’s meaning consist of? The semantic realization of language occurs as a mental process of “filling” word sounds with “the extraction of meaning from what we see, hear, and read”. (1987a: 320; my emphasis) This crucial passage identifies in a nutshell the essential features of what turns an arbitrary, empty signifier into a motivated signified. According to Vygotsky, the non-semantic sound of words and their assemblage in syntactic strings are semantically activated by our nonverbal resources of sight, hearing and, we should add, all other perceptual modalities and their modifications in and by Vorstellung. This is why Vygotsky can say that “a word devoid of thought is a dead thing”. (1991: 255) The distinction between word and thought remains sharp throughout his work. Recalling the situation of having seen “a barefoot boy in a blue shirt running down the street,” Vygotsky says that he conceived of this “in one thought, but I put it into separate words”. (1991: 251) This justifies his controversial insistence on the word as “the most elementary form of the unity between thought and word,” which “cannot be further analysed”. (1991: 212) Because nonverbal thought remains an

The Linguistic Route: Saussure and Vygotsky

145

active component of meaning, speech comprehension depends on whether we are able to imagine a portion of a “world” as a result of hearing the sounds of linguistic expressions. Vice versa, in order to convey our thoughts, verbal and nonverbal, we must find linguistic expressions that guide our speech partners to imagine the sort of “world” we want to project. But the relationship between thought and word is by no means straightforward. “Neither the external-verbal nor the semiotic-meaningful aspect of speech appears suddenly in its final form”. (1987a: 320) They do not “move in parallel,” nor is either a “copy of the other”. (1987a: 320) Rather, they “are the reverse of one another”. (1987a: 321) This suggests a dialectically dynamic, chiastic structure of word and imaginability. Initially, Vygotsky observes an opposition between words and their heterosemiotic contents of tactile, visual, olfactory and other origins, which gradually approach one another, cross over and move apart in the opposite direction, with adult speech characterized as an oppositional association of abstracted thought and word sound. If “the meaning of the word is a generalization” (1987a: 238) and the evolution of verbal meaning is a generalization of “previous generalizations”, (1987a: 229) then once more the nature and degree of abstraction involved becomes an urgent issue to address, as does the relation between Vygotsky’s dynamic conception of meaning and the assumption of its immutability in definitional semantics. Like Peirce, Vygotsky accentuates the evolving character of meaning. Ontogenetically, “the meaning of the child’s word develops”. (1987a: 322) Much the same can be said from a phylogenetic perspective. In either case, semantic development is to be associated with “the intellectualization of the mental functions”. (1987a: 324) In opposition to describing linguistic meaning by way of “definition,” Vygotsky points us away from the inevitable infinite regress inherent in definitional approaches and towards the heterosemiotic ground that makes word sounds semantic. “A major deficiency of the method of definition,” Vygotsky notes, “is that the concept is torn from its natural connections”. (1987a: 123) This remark contains a profound critique of an entire paradigm of semantics. Definitions only work if we already have semantic mastery over a significant portion of their “as-structure,” which begs the very question of meaning. So we should jettison semantic explications via verbal substitutions and instead argue for links between word sounds and nonverbal mental scenarios. “A word,” he says, “acquires its sense from the context in which it appears; in different contexts it changes its sense. Meaning remains stable throughout the changes of sense”. (1991: 245)

146

Chapter Six

The “stability” of meaning in Vygotsky’s sense of “meaning” as provided in the dictionary is of course not semantic at all. Only in a metaphorical sense can the dictionary meaning be called “a stone in the edifice of sense”. (1991: 245) After all, without the “filling” of linguistic expressions by nonverbal, mental materials it is nothing but a syntactic relation between two sound sequences. To be sure, there remains a tension in Vygotsky’s account between the emergence of “the connection between thought and word” and its consolidation in the adult native speaker where we are dealing with “stable and constant word meanings”. (1987a:143; my emphasis) On the one hand, we cannot assume any semantic immutability because “word meanings develop,” (1991: 212) and because “new word meanings” facilitate “new paths from thought to word” (1991: 251) and also because “there is no rigid correspondence between the units of thought and speech”. (1991: 249) On the other hand, at the end of the long process of language acquisition, Vygotsky stabilizes meaning as “a generalized reflection of reality” semantically informing “the basic character of words”. (1991: 255) This leaves us with the following issues still to be clarified. One is how precisely nonverbal materials can be generalized to suit linguistic meaning; the other, how nonverbal mental states can be aligned with public discourse. When Vygotsky speaks of the generalization that nonverbal materials undergo in the process of intellectualization, this must not be understood as a process of formalization, which would mean the elimination of nonverbal content from language. Instead, “the verbal character of thought is inherent to both imagination and realistic thinking” and “both imagination and realistic thinking are often characterized by a high degree of affect and emotion”. (1987a: 348) Nonetheless, what is missing in Vygotsky is an argument in favour of the kind of schematization that we must stipulate nonverbal mental materials undergo to fit the bill of intellectualized verbal thought which he sees as the end point of language acquisition in the child. As to Vygotsky’s formulation of “Locke’s dilemma,” that “direct communication between minds is impossible” and that “communication can be achieved only in a roundabout way,” this points to a pivotal disagreement in the struggle over semantics in linguistics and the philosophy of language. (1991: 252) Because of the simultaneous external and internal mediation of thought by signs, Vygotsky cannot but commit himself to “study the inner workings of thought and speech, hidden from direct observation”. (1991: 252; 254) Yet it is precisely because of the hidden nature of thought that an entire tradition from Frege to this day has rejected the Lockean path of mental resemblance relations and instead

The Linguistic Route: Saussure and Vygotsky

147

opted for externalist explanations. In opposition to semantic externalism and in support of Vygotsky’s position, a solution to Locke’s paradox can be found in an argument for the indirectly public nature of linguistic meaning.43 Thus, native speakers can be defined as persons that have been trained to activate Vorstellung within the strict constraints of the speech community under the rules of sufficient semiosis and the linguistic linkage compulsion. 44 This allows us to argue the reconciliation of public signifiers with indirectly public signifieds in the social production of meaning without either having to defend semantic privacy or opt for an untenable externalism. The non-reductive transformation of thought into its abstracted verbal form is implicated also when Vygotsky speaks of the “plane” of “thought itself” as the “last step in our analysis”. (1991: 249; 252) It is at this point that such diametrically opposed research enterprises as Vygotsky’s and Donald Davidson’s all of a sudden appear to meet one another. As observed earlier, toward the end of his career Davidson surprises his readers by saying that “truth thus rests in the end on belief and, even more ultimately, on the affective attitudes”. 45 Similarly, towards the end of Thought and Language, Vygotsky writes that thought cannot be separated out from our “desires and needs, our interests and emotions” since “behind every thought there is an affective-volitional tendency, which holds the answer to the last ‘why’ in the analysis of thinking”. (1991: 252; my emphasis) Without grasping the affective motivation underlying the words of speakers we cannot fully understand their speech.

Conclusion Although Vygotsky’s focus is primarily on the acquisition of language in childhood rather than on a natural language semantics, the chapter has gleaned from his Collected Works what appear to be essential features of linguistic meaning in the adult native speaker. I sum these up as follows. (1) Linguistic meaning is a generalized nonverbal complex in the unification of thought and word sound. (2) Thought consists of nonverbal materials and their abstraction in language. (3) The combination of thought and word takes place in the minds of the speakers of a natural 43

Cf. Chapter One. Ruthrof, Semantics and the Body, 48f. and “Metasemantics and Imaginability,”, 23, respectively. 45 Donald Davidson, Truth and Predication (Cambridge, Mass.: Harvard University Press, 2005), 75; my emphasis. 44

148

Chapter Six

language. (4) This mental event occurs under the guidance of the speech community and so is fundamentally social. (5) Nonverbal materials include perceptual contents and their productive modifications in and as Vorstellung. (6) Given their heterogeneous sources, such mental contents are heterosemiotic. (7) As such they are homogenized into general concepts in linguistic meanings. (8) In linguistic meaning, the initial dominance of nonverbal thought over words is reversed such that generalized verbal meanings govern those contents. As we shall see in the next chapter, it took a while since Vygotsky’s path-breaking findings for intersubjective mentalism to become respectable once more, as it has in cognitive linguistics and foremost in the research conducted by Michael Tomasello.

CHAPTER SEVEN COGNITIVE LINGUISTICS AND CONCEPTUAL BLENDING

Introduction I have postponed until now an exploration of the relation of my metasemantic principles and the body of literature known as cognitive linguistics. On the one hand, the imaginability thesis belongs squarely to cognitive theory in the broad sense; on the other hand, there are some significant differences between cognitive linguistics and the imaginability project. They have in common the rejection of the commitment to an autonomous language faculty in the mind and the idea of a modular brain where a “language organ” will once be discovered. Since it is likely that language emerged as an economising matrix to cover, abbreviate, and make more efficient all human mental capacities, the opposite assumption looks much more attractive. Namely, the view of the likelihood of the bulk of neurons playing some role in the production and comprehension of language and that the totality of our brain has evolved to service the needs of the social complexity of the human mind. As Ray Jackendoff has put it bluntly, “language shares the general character of other faculties of the mind”.1 Another broad agreement concerns my critique of propositional and formal approaches, as discussed in Chapter Four. Similarities can also be found in the commitment in Language and Imaginability to meaning as “use”, except that its Wittgensteinian version is argued, as I did in Chapter Five, to require revision with a view of re-incorporating mental states in the form of imaginability. From the joint emphasis on the description of natural language as emerging from a broadly conceived use follows that neither cognitive linguistics nor a semantics of imaginability could support any neat separation of semantics from pragmatics. At the same time, as I 1

Ray Jackendoff, “Word Meanings and What It Takes to Learn Them: Reflections on the Piaget-Chomsky Debate,” in The Nature and Ontogenesis of Meaning, ed. Willis F. Overton (Hillsdale, N.J.: Lawrence Erlbaum, 1994), 129-144; 60.

150

Chapter Seven

argue in Chapter Nine, pragmatics itself is in need of review from the perspective of the neglected mental process of imaginative projections. After all, I claim, conceivability and all other propositional moves in a fundamental way supervene on imaginability (see Chapter Eight). Apart from such broad accord, the chapter will draw the reader’s attention to specific correspondences as well as a number of specific differences.

Parallels and incompatibilities Using the work of Vyvyan Evans as my first reference,2 I list the following agreements. There is fundamental agreement on the idea that without granting primacy to meaning “our understanding of language cannot be advanced in a significant way”; that “mental representation” plays a crucial role in linguistic meaning; that meanings cannot be “captured in propositional terms”; that words function like “prompts for meaning construction”; that words cannot “contain” meanings; that “truthconditional semantics” has failed to be convincing; and that it is an illusion to think that we can have a neat separation of “semantics and pragmatics”. 3 At the same time, the agreement does not extend to explanations why I share those broad convictions. My emphasis on the iconicity of mental representations finds support in Esa Ikonen’s singling out “analogy” as “the lifeblood of human thinking”, 4 in the sense that analogy can be viewed as a special case of my broader and central notion of resemblance relations. As to the work of Charles Fillmore 5 on Construction Grammar, “frame semantics” can be used to shore up the claims of the projectionist thesis of imaginability, in giving prominence to the grammatical holism of “grammatical construction” and its embodied character over piecemeal syntax, except that I want to place the accent on 2

Vyvian Evans, The Structure of Time: Language, Meaning and Temporal Cognition (Amsterdam: John Benjamins, 2003); How Words Mean: Lexical Concepts, Cognitive Models, and Meaning Construction (Oxford: Oxford University Press, 2009); A Glossary of Cognitive Linguistics (Edinburgh: Edinburgh University Press, 2007). 3 Evans, The Structure of Time, 39; 53; 54. 4 Esa Itkonen, Analogy as Structure and Process: Approaches in Linguistics, Cognitive Psychology and Philosophy of Science (Amsterdam: John Benjamins, 2005). 5 Charles Fillmore, “Towards a Descriptive Framework of Spatial Deixis,” in Speech, Place and Action, ed. R. Jarvella and W. Klein (London: Wiley, 1982); “The Case for Case,” in Universals in Linguistic Theory, ed. E. Bach and R.T. Harms (New York: Holt, Rinehart and Winston, 1968). [1975 and 1976]

Cognitive Linguistics and Conceptual Blending

151

the protosyntactic organisational principles which must already be part of a pre-linguistic Vorstellungswelt out of which grammatically organized language emerges both phylogenetically and ontogenetically. Likewise, my project is in sympathy with Leonard Talmy’s emphasis on gradation in semantics in contrast to the freeze frame results of formal frames of investigation, and his findings concerning the augmentability of the lexicon as an “open class”.6 Where I part company with Talmy’s project is by emphasizing imaginability as a precondition of language rather than a result of linguistic categories, though I take Talmy’s point that from the perspective of the open and closed categories which we can observe in all language, it is linguistic structures that impose a specific cohesion on the heterosemiotic sign systems that are constitutive of the Vorstellungswelt. That is, I sharply distinguish his “schematic categories”, which are language driven, from my own iconic schematizations, which are generated at the level of imaginability. The reader will discover a strong affinity of the imaginability project with the writings of Michael Tomasello, especially in his foregrounding of the mental states that must be involved in “intention-reading”, goal recognition in others, “pattern finding” beyond single utterances, and the kind of mental processes we must stipulate to make sense of the child’s ability to develop communicative motivation. This emphasis is reminiscent of what Husserl pointed to as introjection, comprising “all acts that a hearer may introject into a speaker”. 7 Attractive too is Tomasello’s explanation of the acquisition of grammar in children as the result of imitation of recurrent patterns of signifiers in specific social situations.8 Here his work accentuates a usage-based linguistics, in the wake especially of Ronald Langacker.9 What should be pointed out, however, is that Tomasello’s notion of “use” is not entirely compatible with that of the later Wittgenstein, since unlike the author of the Philosophical Investigations Tomasello is granting special prominence to the mental side of the meaning event. And it is precisely “this mental dimension that gives

6

Leonard Talmy, Towards a Cognitive Semantics (Cambridge, Mass.: MIT Press, 2000) 7 Husserl, Logical Investigations I §7. 8 Michael Tomasello, Constructing a Language: A Usage-Based Theory of Language Acquisition (Cambridge, Mass.: Harvard University Press, 2003). 9 Ronald Langacker, Foundations of Cognitive Grammar, vol. 1 (Stanford: Stanford University Press, 1987); Foundations of Cognitive Grammar, vol. 2 (Stanford: Stanford University Press, 1991);

152

Chapter Seven

linguistic symbols their unparalleled communicative power”.10 Note here, however, that Tomasello does not employ the term symbolic in the formal sense but rather in the sense of pointing to an extra-linguistic, perceptual reality. In other words, Tomasello’s “linguistic symbol” appears to be much closer to the concept of the “linguistic sign” as consisting of an arbitrary signifier and a motivated signified. Otherwise he could not say, as he does, that “linguistic symbols are fundamentally perspectival”.11 At the core of Tomasello’s enterprise seems to lie the conviction that perhaps the most crucial event in the evolution of modern humans as well as in child development is when “human beings understood one another for the first time as intentional and mental agents - which then led them to attempt to manipulate one another’s intentional and mental states for various cognitive and competitive purposes”.12 Again, what I want to stress here is that the experimental side required for such manipulation cannot be overestimated. We need to add a strong dose of imaginability to this process, on which all specific forms of conceivability supervene of necessity. Of particular interest for the question of the emergence of language under the primacy of imaginability is Tomasello’s linking of prelinguistic acts such as “pointing” with later verbalisation, a link sedimented for example in the German term for meaning, Bedeutung (consisting of the active prefix be- and deuten, to point). It is tempting to take Tomasello’s observation as an indication that gestural communication not only predates language in ontogenesis but that the same sequence informed the kind of transformations that are likely to have occurred in phylogenesis, a topic he is resuming in his forthcoming book Origins of Human Communication. 13 What I would stress is that the mental projections stipulated to allow for Tomasello’s observations such as imitative behaviour in the process of attention-sharing must also and importantly involve an experimental mental dynamics. While Tomasello’s work in part draws on Ronald Langacker’s notion of use, the imaginability thesis finds support in the following areas: Langacker’s accent on the social control of language via conventions; grammar as part of the mastery of conventions; that language must be a 10

Tomasello, Constructing a Language, 8. Tomasello, Constructing a Language, 12. 12 Ibid.; cf. Anna Wierzbicka’s Lingua Mentalis: The Semantics of Natural Language (New York: Academic Press, 1980) where she insists that semantics must be able to “account for intuitively felt semantic relations”. 13 Michael Tomasello, Origins of Human Communication (Cambridge, Mass.: MIT Press, 2014). 11

Cognitive Linguistics and Conceptual Blending

153

socio-psychological product; that schemas and sub-schemas play an important role in the organization of Vorstellung by language; that language is not well described as an algorithmic set; that grammar cannot be viewed as a formal system; and that grammar reflects a nonverbal manner of conceiving the world. 14 I offer a different explanation on language as a heterosemiotic rather than a symbolic sign system; I foreground the entire spectrum of perceptual modalities to avoid the visual bias inherent in the image schema; and I curtail his psychological emphasis via my linguistic linkage compulsion.15 In this respect, Yanna Popova’s exploration of the “cross-modal nature of image schemas” in her article on verbal synaesthesia is congenial, synaesthesia being conceived of as “involuntary perceptual experience of a cross-modal association”.16 My general criticism, though, is that it would be preferable to move the notion of schema to a meta-level to allow for thermal, gravitational, gustatory, olfactory and other non-spatial readings. Partially compatible too is Jordan Zlatev’s “mimetic schema” as part of his discussion of embodiment and language, in that it acknowledges Aristotelian resemblance relations as the basis of language.17 Again, though, I want to qualify this compatibility by suggesting that the mimetic dimension of embodiment is no more than the visible tip of the iceberg, leaving the bulk of mental states, their imaginative extensions, that is, imaginability, remarkably undertheorized.18 Imaginability, it can be argued, is heavily involved also in one of the important non-visual iconic ingredients, usually referred to as “emotion”. A strong case that emotion does indeed play a powerful role in the phylogenesis of language has recently been made in The First Idea: How Symbols, Language, and Intelligence Evolved from our Primate Ancestors

14 Ronald Langacker, Cognitive Grammar: A Basic Introduction (New York: Oxford University Press, 2008), 3f. 15 Cf. Chapter Ten; see also Ruthrof, “Metasemantics and Imaginability,” 23f.; 25. 16 Yanna Popova, “Image Schemas and Verbal Synaesthesia,” in From Perception to Meaning: Image Schemas in Cognitive Linguistics, ed. Beate Hampe (Berlin: Mouton de Gruyter, 2005), 421- 442; 396. 17 Jordan Zlatev, “Embodiment, Language, and Mimesis,” in Body, Language and Mind, vol 1, ed. Tom Ziemke, Jordan Zlatev, and Roslyn Frank (Berlin: Mouton de Gruyter, 2007), 297-338; cf. also Jordan Zlatev, Situated Embodiment: Studies in the Emergence of Spatial Meaning (Stockholm: Gotab Press, 1997). 18 Cf. the obverse criticism made by Leonard Talmy that the notion of embodiment covers too much and remains vague, in Towards a Cognitive Semantics.

154

Chapter Seven

to Modern Humans by Stanley Greenspan and Stuart Shanker. 19 The authors argue the plausible case of emotions being one of the main driving sources in the invention of language. If they are right on this basic claim, as I think they are, then it is likely that emotional traces still form part of the heterosemiotic background on which linguistic meaning depends to this day. This area of resemblance relations carried by language is as significant in understanding the complexities of the meaning process as it is difficult to argue. Yet, as indicated above, Tomasello’s rich notion of “intention-reading”, for example, could not get off the ground without the assumption that emotional attunement is a necessary component of linguistic exchange. Given certain parallels between such recent work and the imaginability project, I need to address some of the early cognitive impulses on my thinking. They came in particular from the writings of George Lakoff, Mark Johnson, Mark Turner, Eve Sweetser, and Gilles Fauconnier over the last thirty years. I want to touch on some of these writings in this section, before addressing specific cognitive notions more critically. Foremost there is Mark Johnson’s well-received book in which the author revived a number of Kantian principles for a cognitive approach to mind and language. In Johnson’s “cognitive semantics,” conceived as a “semantics of understanding,” linguistic meanings are largely nonpropositional, yet public. 20 More recently, Johnson has brought his description of image schemas up to date in “The Philosophical Significance of Image Schemas”.21 Johnson’s position is to a large extent compatible with the imaginability thesis, except for my arguments on the indirectly public nature of linguistic meaning and the objection to the visual bias which Johnson shares with other cognitivist linguists. I will have more to say about image schemas below. Another approach close to the convictions that inform Language and Imaginability is a paper entitled “Design for a Theory of Meaning” in which Mark Turner returns us to a phrase by the Presocratic philosopher 19

Stanley Greenspan and Stuart Shanker, The First Idea: How Symbols, Language, and Intelligence Evolved from our Primate Ancestors to Modern Humans (Cambridge, Mass.: Da Capo Press, 2004). 20 Mark Johnson, The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason (Chicago: University of Chicago Press, 1987), 175-190. 21 Mark Johnson, “The Philosophical Significance of Image Schemas,” in From Perception to Meaning: Image Schemas in Cognitive Linguistics, ed. Beate Hampe (Berlin: Mouton de Gruyter, 2005), 15-34.

Cognitive Linguistics and Conceptual Blending

155

Protagoras, “Man is the measure of all things” in order to question referential as well as formal theories of language. In referential theories of meaning, Turner says, “a semantic express train shoots straight from the linguistic symbols to an objective reality without passing through the human brain,” while in formalist theories of meaning the human brain is no more than an “implementational detail”.22 Turner accepts none of the traditional dichotomies, such as the distinctions between objective and subjective meaning, mind and body, culture and biology, innate and acquired, or genetics versus experience. Instead, all we are dealing with is the neurally grounded, “imaginative mind”. In his other writings, Turner has tried to show how the study of literature and culture in general are in need to be reinvigorated by a cognitive approach based on the fundamental insights of the Lakoff School. “Our concept of rational argument would appear to be a product of poetic thought,” for “without metaphor, we would not have the concept of argument we have” and “without a body, we could not have the concept of argument we have”.23 Here and in some of his other writings, Turner goes beyond Lakoff by demonstrating the explanatory power of cognitive rhetoric for language analysis, including literary meaning and interpretation.24 A particularly exciting aspect of cognitive linguistics is the investigation into “mental spaces” explored by Gilles Fauconnier. While the relation between language and perception has already received a significant boost by the turn to neurobiology, imaginability, as perceptual modification, has been dealt with mainly by way of spatial analogy. Nevertheless, Gilles Fauconnier’s discussion of the notions of mapping and mental spaces is of special interest to this book. In Espaces mentaux: aspects de la construction du sens dans les languages naturelles Fauconnier announced a programme of investigating the principle of “construction mentale” in relation to linguistic expressions, indicative of “a certain plane of cognitive organisation”. 25 The role of “mental spaces” is extended in Fauconnier’s Mappings in Thought’s Language where he explores the likely links between the way language reflects the world indirectly by way 22

Mark Turner, “Design for a Theory of Meaning,” in The Nature and Ontogenesis of Meaning, ed. Willis F. Overton and David S. Palermo (Hillsdale, N.J.: Lawrence Erlbaum, 1994), 91-107; 92. 23 Mark Turner, Reading Minds: The Study of English in the Age of Cognitive Science, (Princeton: Princeton University Press, 1991), 100. 24 Mark Turner, Death Is the Mother of Beauty: Mind, Metaphor, Criticism (Chicago: University of Chicago Press, 1987). 25 (Paris: Minuit, 1984), 13.

156

Chapter Seven

of cognitive construals. Meaning is argued to occur at lightening speed, with much of its operations remaining unconscious. The idea of mapping as a device is taken as central for linking mental spaces.26 But while in Fauconnier’s perspective mental spaces are as much “constructions associated with discourse” as they are a “manner of speaking” not to be “taken literally,” the present study makes a case especially for imaginability, its mental modification and multiplication, in terms of mental, iconic resemblance functioning as indispensable sources and ingredients of natural language.27 I will resume Fauconnier’s and Turner’s contribution to cognitive linguistics in the last section of this chapter. Another seminal contribution to the basics of cognitive linguistics is the research by Ann Sweetser from the perspective of language philosophy in arguing that we can discover neurobiological traces in abstract terms and function words.28 This is important for two reasons. First, her argument significantly enlarges our knowledge of function words, or syncategorematic terms, such as the connectives “and,” “but,” “if,” “or” and others. She makes a convincing case, for example, of the claim that the conjunction “and” has a symmetrical as well as an asymmetrical application, the first of which permits a reversal of the sentence so conjoined, while the second does not. Importantly, the formal conjunction “and” allows no such alternative. In natural language, Sweetser shows, “and” can fulfil several functions, such as the temporal connective “and then,” a causal conjunction “and so,” as well as an epistemic function, when meaning is the result of interpretive choices on the part of the listener or reader. Second, she suggests that not only the content of natural language terms but their syntactic combinations are likewise traceable to “human perception”. There are two things that strike me as particularly fruitful in her remarks. One is the not so obvious implication that there may be something like a protosyntax informing the way natural languages have developed their specific syntactic constructions. 29 The other important implication seems to me to point to the epistemic side of perception, a 26

Gilles Fauconnier, Mappings in Thought’s Language, (Cambridge: Cambridge University Press, 1997). 27 Gilles Fauconnier, Espaces Mentaux, 193f. 28 Anne Sweetser, From Etymology to Pragmatics: Metaphorical and cultural Aspects of Semantic Structure (Cambridge: Cambridge University Press, 1990). 28 Gilles Fauconnier, Espaces Mentaux: Aspects de la Construction du Sens dans les Languages Naturelles. 29 Horst Ruthrof, “From Kant’s Monogram to Conceptual Blending,” 118; “Semantics of Imaginability – Vorstellungssemantik: 13 Theses”, 167.

Cognitive Linguistics and Conceptual Blending

157

feature that lives on in the linguistic sign by having gone underground, a topic I deal with under the term implicit deixis or enunciative modality in a broad sense.30 A groundbreaking paper is George Lakoff’s “What Is a Conceptual System?” in which he attempts to show how we can trace abstract linguistic expressions back to our perceptual performance in the brain by identifying metaphor as the “main mechanism through which we comprehend abstract concepts and perform abstract reasoning”. 31 To demonstrate this claim, Lakoff moves metaphor out of the domain of language into the more general domain of conceptualisation where he says we can observe “cross-domain mappings,” an activity characterized by the conceptualisation of “one mental domain by another”, a forerunner of the more recent notion of conceptual blending. (43) Nor, according to Lakoff, is this mechanism an isolated occurrence. Quite the contrary, “everyday metaphor is characterised by a huge system of thousands of cross-domain mappings”. No surprise then that in cognitive linguistics the term metaphor is routinely translated as “cross-domain mapping in the conceptual system”. (43) Linguistic metaphor, from this perspective then, is no more than a surface phenomenon of such conceptual mapping. At bottom, says Lakoff, metaphor is “conceptual, not linguistic in nature”. (84) When Lakoff insists that all metaphorical grasp is “grounded in nonmetaphorical understanding,” we want to know more about this stipulated baseline of human comprehension. Are metaphors combinations of perceptual readings freed from their sensory base? Are they perceptual variations in Vorstellung, the forum of imaginative modification of perception? If so, then imaginability as a neuron based faculty by which humans vary and so multiply perceptual input nonconsciously, as well as at will, should be identified as the ground from which language evolves both phylogenetically and ontogenetically.

30

Ruthrof, The Body in Language, 48-53; “Semantics of ImaginabilityVorstellungssemantik: 13 Theses,” Review of Contemporary Philosophy 10, 2011, 165-183. 31 George Lakoff, “What Is a Conceptual System?” in The Nature and Ontogenesis of Meaning, ed. Willis F. Overton and David S. Palermo (Hilldale, N.J.: Lawrence Erlbaum, 1994), 41-90; 84. Cf. also “The Invariance Hypothesis: Is Abstract reasoning Based on Image-Schemas?” Cognitive Linguistics, 1 (1990), 39-74; “Cognitive Semantics,” in Meaning and Mental Representation, ed. U. Eco, M. Santambrioggio, and P. Violi (Bloomington: Indiana University Press, 1988).

158

Chapter Seven

Where my project differs markedly from some cognitive linguists is on the notion of the “lexical concept” conceived, for instance, by Evans as a “subset of concepts which are paired with linguistic symbols”, or “the conventional meaning associated with a particular lexeme (or collocation)”. 32 Instead of “concept” in Evans, I retain the Saussurean distinction between signifier and signified, except that my signified is understood throughout as motivated. Furthermore, I split the motivated signified into two components, mental iconic materials providing content and the concept which is responsible for the socially agreed upon regulation of those materials. This distinction obviously revives the separation which we find in Kant between Anschauung, Vorstellung and Einbildungskraft, on the one hand, and, on the other, their social conceptual regulation. This allows for the important differences by which different cultures and their languages cut up physiognomically shared iconic mental materials by specific ways of regulating those contents. As to the notion of “cognitive models” as an “integration of lexical concepts” in Evans, I prefer the looser idea of the projection of iconic mental scenarios, which gives a good deal more room to the imaginative variations on perception that I believe are underrated in cognitive linguistics. This criticism also applies to Alva Noe’s Action in Perception, Representation and Mind which, though, marginal to cognitive linguistics, adds an important new perspective to an embodied conception of consciousness and its mechanisms with the idea of an “enactive approach” and the author’s commitment to “a neuroscience of embodied activity”.33 Alva Noe’s point that perception and thought are driven to a large extent by our actions in the world needs to be accepted without however having to give up on my emphasis on the asymmetrical relation between perceptual input and internal monitoring, heavily in favour of the latter, as a result of the evolution of the human brain. I do not accept the implication that this should in any way entail a “demotion” of imaginability. While our Vorstellungen are continually modified by our bodily interaction with the actual world, this interaction only strengthens the power of Vorstellung and our capacity of producing an in principle infinite number of virtual variations of perceptual input, including those that “bounce back” from enaction. 32

Evans, The Structure of Time, 6; 223. Alva Noël, Action in Perception, Representation and Mind (Cambridge, Mass.: MIT Press, 2004), 227f. 33

Cognitive Linguistics and Conceptual Blending

159

Where my perspective differs from the cognitive picture drawn by Fauconnier and Turner in The Way We Think (2002), to be addressed in detail at the end of the chapter, is on the generation and mappings between mental spaces in relation to Vorstellung. Here I hold the obverse position. It is Vorstellungen as the result of our faculty of imaginability that is the ground for recognizing and relating both mental spaces and their iconic reappearance in language. I stipulate mental spaces as part of iconic schematizations to account for conceptually regulated content of the motivated signified. Fauconnier and Turner extrapolate them from already linguistically organized Vorstellung. As I speculate, without an extremely rich Vorstellung, we would not have developed language in the first place. If this is so, then imaginability is the ground on which non-linguistic as well as linguistic meanings stand, and well in advance of conceivability.34 Other differences between cognitive linguistics and Language and Imaginability include the vital distinction I draw between habitual language and interpretive use. It is only the latter that I would want to classify as a form of conscious mental acts of interpretation. Habitual language, in contrast, is strongly circumscribed by the linguistic linkage compulsion, according to which we cannot be aware of two meaning events simultaneously. In habitual speech, the signifier cue of “door-knob” denies the native speaker of a language any option during the split second event of meaning. The idea that we are in a position to entertain two or several different meaning constructions is an illusion, just as brain research is now confirming. The typical time span of comprehension of