Complexity Applications in Language and Communication Sciences [1st ed.] 978-3-030-04596-8, 978-3-030-04598-2

Table of contents :
Front Matter ....Pages i-xi
Introduction (Àngels Massip-Bonet, Gemma Bel-Enguix, Albert Bastardas-Boada)....Pages 1-13
Front Matter ....Pages 15-15
Science as a Social Self-organizing Extended Cognitive System. Coherence and Flexibility of Scientific Explanatory Patterns (Robert Hristovski, Natàlia Balagué, Pablo Vázquez)....Pages 17-30
The Paradigm of Complexity in Sociology: Epistemological and Methodological Implications (Alvaro Malaina)....Pages 31-42
How and Why to Model the Complexity of Thought Systems (Leonardo G. Rodríguez Zoya)....Pages 43-73
Front Matter ....Pages 75-75
Linguistic Variation and Change: Approach from the Perspective of Complex Adaptive Systems (Àngels Massip-Bonet)....Pages 77-92
Some theoretical Prerequisites for the Integrated Study of Linguistic “Macrochange” (Enrique Bernárdez)....Pages 93-105
The Impact of Social Reputation in Language Evolution (Gemma Bel-Enguix)....Pages 107-116
Front Matter ....Pages 117-117
‘Restricted’ and ‘General’ Complexity Perspectives on Social Bilingualisation and Language Shift Processes (Albert Bastardas-Boada)....Pages 119-137
Patterns of Linguistic Diffusion in Space and Time: The Case of Mazatec (Jean Léo Léonard, Marco Patriarca, Els Heinsalu, Kiran Sharma, Anirban Chakraborti)....Pages 139-170
Common Knowledge in Conversation of Bilinguals and the Ecology of Pressures. The Complex Processes of Using Language and Learning to Coordinate Actions with Other Speakers (Roland Terborg, Virna Velázquez)....Pages 171-184
Front Matter ....Pages 185-185
Discourse Analysis: The Constructivist Perspective and Transdisciplinarity (Esperanza Morales-López)....Pages 187-205
A Complex Approach to Prosodic Discourse Variation (Raquel García Riverón, Alejandro F. Marrero Montero)....Pages 207-225
Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic Performances (H. Samy Alim, Geneva Smitherman)....Pages 227-248
Front Matter ....Pages 249-249
Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem (Michael Zock)....Pages 251-272
The Emergence of Hubs in Complex Syntactic Networks and the DP Hypothesis: The Relevance of a Linguistic Analysis (Lluís Barceló-Coblijn, Maia Duguine, Aritz Irurtzun)....Pages 273-288
The World Color Survey: Data Analysis and Simulations (Peter Lewinski, Michal Lukasik, Konrad Kurdej, Filip Leonarski, Natalia Bielczyk, Franciszek Rakowski et al.)....Pages 289-311
Cognitive Meaning: Review of the Concepts of Imagination, Image Schema and Mental Image and Consequences on the Conceptualization of Emotions (Maria Antònia Font Fernández)....Pages 313-323
Correction to: The World Color Survey: Data Analysis and Simulations (Peter Lewinski, Michal Lukasik, Konrad Kurdej, Filip Leonarski, Natalia Bielczyk, Franciszek Rakowski et al.)....Pages C1-C1
Back Matter ....Pages 325-331

Citation preview

Àngels Massip-Bonet Gemma Bel-Enguix Albert Bastardas-Boada   Editors

Complexity Applications in Language and Communication Sciences

Complexity Applications in Language and Communication Sciences

Àngels Massip-Bonet Gemma Bel-Enguix Albert Bastardas-Boada •

Editors

Complexity Applications in Language and Communication Sciences

123

Editors Àngels Massip-Bonet Department of Catalan Philology and General Linguistics, Scripta (FFI2016-80482P), UBICS, CUSC Universitat de Barcelona Barcelona, Spain

Albert Bastardas-Boada Department de Lingüística General Universitat de Barcelona Barcelona, Spain

Gemma Bel-Enguix Grupo de Ingeniería Lingüística Instituto de Ingeniería, Universidad Nacional Autónoma de México Ciudad de México, Mexico

ISBN 978-3-030-04596-8 ISBN 978-3-030-04598-2 https://doi.org/10.1007/978-3-030-04598-2

(eBook)

Library of Congress Control Number: 2018962779 © Springer Nature Switzerland AG 2019, corrected publication 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Àngels Massip-Bonet, Gemma Bel-Enguix and Albert Bastardas-Boada

Part I 2

3

4

6

7

Interdisciplinary Approaches for Human Sciences

Science as a Social Self-organizing Extended Cognitive System. Coherence and Flexibility of Scientific Explanatory Patterns . . . . . Robert Hristovski, Natàlia Balagué and Pablo Vázquez

17

The Paradigm of Complexity in Sociology: Epistemological and Methodological Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . Alvaro Malaina

31

How and Why to Model the Complexity of Thought Systems . . . . . Leonardo G. Rodríguez Zoya

Part II 5

1

43

Language Change

Linguistic Variation and Change: Approach from the Perspective of Complex Adaptive Systems . . . . . . . . . . . . . . . . . . . Àngels Massip-Bonet

77

Some theoretical Prerequisites for the Integrated Study of Linguistic “Macrochange” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enrique Bernárdez

93

The Impact of Social Reputation in Language Evolution . . . . . . . . 107 Gemma Bel-Enguix

v

vi

Contents

Part III

Sociolinguistics

8

‘Restricted’ and ‘General’ Complexity Perspectives on Social Bilingualisation and Language Shift Processes . . . . . . . . . . . . . . . . 119 Albert Bastardas-Boada

9

Patterns of Linguistic Diffusion in Space and Time: The Case of Mazatec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Jean Léo Léonard, Marco Patriarca, Els Heinsalu, Kiran Sharma and Anirban Chakraborti

10 Common Knowledge in Conversation of Bilinguals and the Ecology of Pressures. The Complex Processes of Using Language and Learning to Coordinate Actions with Other Speakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Roland Terborg and Virna Velázquez Part IV

Discourse Analysis

11 Discourse Analysis: The Constructivist Perspective and Transdisciplinarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Esperanza Morales-López 12 A Complex Approach to Prosodic Discourse Variation . . . . . . . . . . 207 Raquel García Riverón and Alejandro F. Marrero Montero 13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic Performances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 H. Samy Alim and Geneva Smitherman Part V

Syntax, Semantics and Cognition

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 Michael Zock 15 The Emergence of Hubs in Complex Syntactic Networks and the DP Hypothesis: The Relevance of a Linguistic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 Lluís Barceló-Coblijn, Maia Duguine and Aritz Irurtzun 16 The World Color Survey: Data Analysis and Simulations . . . . . . . 289 Peter Lewinski, Michal Lukasik, Konrad Kurdej, Filip Leonarski, Natalia Bielczyk, Franciszek Rakowski and Dariusz Plewczynski

Contents

vii

17 Cognitive Meaning: Review of the Concepts of Imagination, Image Schema and Mental Image and Consequences on the Conceptualization of Emotions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 Maria Antònia Font Fernández Correction to: The World Color Survey: Data Analysis and Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peter Lewinski, Michal Lukasik, Konrad Kurdej, Filip Leonarski, Natalia Bielczyk, Franciszek Rakowski and Dariusz Plewczynski

C1

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

Contributors

H. Samy Alim Department of Anthropology, University of California, Los Angeles, CA, USA Natàlia Balagué National Institute of Physical Education (INEFC) Health and Applied Sciences, University of Barcelona, Barcelona, Spain Lluís Barceló-Coblijn Department of Catalan Philology and General Linguistics, University of the Balearic Islands, Palma, Spain Albert Bastardas-Boada Department of Catalan Philology and General Linguistics, Sociocomplexity—Complexity, Communication and Sociolinguistics Group, CUSC—Research Centre for Sociolinguistics and Communication, UBICS —Universitat de Barcelona Institute of Complex Systems, Universitat de Barcelona, Barcelona, Spain Gemma Bel-Enguix Grupo de Ingeniería Lingüística, Instituto de Ingeniería, Universidad Nacional Autónoma de México, Ciudad de México, Mexico Enrique Bernárdez Department of Linguistics, Universidad Complutense, Madrid, Spain Natalia Bielczyk Polish Academy of Sciences, Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands Anirban Chakraborti School of Computational and Integrative Sciences, Jawaharlal Nehru University, New Delhi, India Maia Duguine Centre National de la Recherche Scientifique, IKER (UMR 5478), Bayonne, France Maria Antònia Font Fernández Universitat de Barcelona, Barcelona, Spain Raquel García Riverón Grupo-Red de Complejidad y Lenguajes de La Habana, Havana, Cuba

ix

x

Contributors

Els Heinsalu National Institute of Chemical Physics and Biophysics, Tallinn, Estonia Robert Hristovski Faculty of Physical Education, Sport and Health, Ss. Cyril and Methodius University, Skopje, Macedonia Aritz Irurtzun Centre National de la Recherche Scientifique, IKER (UMR 5478), Bayonne, France Konrad Kurdej Faculty of Mathematics, Informatics and Mechanics, University of Warsaw, Warsaw, Poland Jean Léo Léonard Sorbonne Université, STIH, Paris, France Filip Leonarski Faculty of Chemistry, Interdisciplinary Centre for Mathematical and Computational Modelling, University of Warsaw, Warsaw, Poland Peter Lewinski Faculty of Law and Saïd Business School, University of Oxford, Oxford, UK Michal Lukasik Google, Zurich, Switzerland Alvaro Malaina Department of Sociology: Methodology Complutense University of Madrid, Madrid, Spain Alejandro F. Marrero Montero Universidad de Las Villas, Santa Clara, Cuba

Central

and

Theory,

“Marta

Abreu”

Àngels Massip-Bonet Department of Catalan Philology and General Linguistics, Projecte Scripta (FFI2016-80482P), Sociocomplexity—Complexity, Communication and Sociolinguistics Group, CUSC—Research Centre for Sociolinguistics and Communication, UBICS—Universitat de Barcelona Institute of Complex Systems, Universitat de Barcelona, Barcelona, Spain Esperanza Morales-López Universidad de A Coruña, A Coruña, Spain Marco Patriarca National Institute of Chemical Physics and Biophysics, Tallinn, Estonia Dariusz Plewczynski Centre of New Technologies, University of Warsaw, Warsaw, Poland Franciszek Rakowski Interdisciplinary Centre for Mathematical Computational Modelling, University of Warsaw, Warsaw, Poland

and

Leonardo G. Rodríguez Zoya Investigador del Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET)—Argentina, Comunidad de Pensamiento Complejo, Instituto de Investigaciones Gino Germani, Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina Kiran Sharma School of Computational and Integrative Sciences, Jawaharlal Nehru University, New Delhi, India

Contributors

xi

Geneva Smitherman Michigan State University, East Lansing, MI, USA Roland Terborg Escuela Nacional de Lenguas, Lingüística y Traducción, Universidad Nacional Autónoma de Mexico, Mexico City, Mexico Pablo Vázquez National Institute of Physical Education (INEFC) Health and Applied Sciences, University of Barcelona, Barcelona, Spain Virna Velázquez Facultad de Lenguas, Univesidad Autónoma del Estado de Mexico, Toluca, Mexico Michael Zock Aix-Marseille Université, CNRS, LIF (UMR 7279), Marseille, France

Chapter 1

Introduction Àngels Massip-Bonet, Gemma Bel-Enguix and Albert Bastardas-Boada

Abstract Based on the acknowledgment that many phenomena in human life are complex, there have been attempts to re-examine the conception of reality. Interdisciplines such as complex thinking, sciences of complexity or complex perspectives try to provide the “old” concepts with a new meaning. The complexity researches imply the restudy of reality, and they have cybernetics as precedent and partly as foundation: a transdisciplinary focus to explore the structures, restrictions and possibilities of regulatory systems. It intends to provide concepts, schemata and possibilities of thought and representation capable of expressing the interweaving and the multidimensional and systematic interdependence of the many phenomena of reality. Linguistics is one of the fields of knowledge that is making great progress under the new paradigm of complexity. The amount of contributions from physics and other scientific disciplines to linguistics is large, under which natural language has been addressed with theoretical and practical methods, both quantitative and qualitative. However, the conceptual resources and tools that are available nowadays are not completely suitable to perform all the tasks. Due to this, it is necessary to keep developing new theoretical and methodological tools that help understanding the dynamic interrelations of linguistic and sociocultural events. Simultaneously, the À. Massip-Bonet Department of Catalan Philology and General Linguistics, Projecte Scripta (FFI2016-80482P), Sociocomplexity—Complexity, Communication and Sociolinguistics Group, CUSC—Research Centre for Sociolinguistics and Communication, UBICS—Universitat de Barcelona Institute of Complex Systems, Universitat de Barcelona, Barcelona, Spain e-mail: [email protected] G. Bel-Enguix (B) Grupo de Ingeniería Lingüística—Instituto Ingeniería, Universidad Nacional Autonoma de México, Ciudad Universitaria, 04510 Ciudad de México, Mexico e-mail: [email protected] A. Bastardas-Boada Department of Catalan Philology and General Linguistics, Sociocomplexity—Complexity, Communication and Sociolinguistics Group, CUSC—Research Centre for Sociolinguistics and Communication, UBICS—Universitat de Barcelona Institute of Complex Systems, Universitat de Barcelona, Barcelona, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_1

1

2

À. Massip-Bonet et al.

lines of inter and transdisciplinary research that transcend the communicative and linguistic phenomenon, and that connect them and interrelate them with life and the world must be strengthened.

1.1 Complexity as a Transdisciplinary View The recognition that many phenomena relating to life are ‘complex’ in nature—i.e., that they are interwoven, self-organising, emergent and processual—has prompted a reexamination of how we have conceived reality, both the way we have looked at it and the images we have used to represent it. This constitutes the point of departure for the articulation of different interdisciplines engaged in refreshing such concepts and finding new ways of thinking that better fit the complex organisation of facts and events. New science perspectives have emerged under this recognition, giving rise to disciplines or paradigms that are referred to as ‘complex thinking’, ‘sciences of complexity’, ‘complex perspectives’, ‘complex [adaptive] systems’, network sciences, … The chose of the terms ‘complex’ and ‘complexity’ is, far from randomly motivated, precisely to suggest what their Latin etymology implies—complexus, to weave, braid, entwine—that is, a common characteristic of phenomena at this level of being, made up of a series of elements that are deeply interwoven and interdependent in their functioning. Fundamentally, the terms point to a need to go beyond approaches that are reductionist, one-dimensional or basically analytical in procedure—approaches that have been, and continue to be, useful at other levels and for other events—and to move towards perspectives that we have come to call systemic, holistic, ecological or networked, because they can more closely suit the kind of occurrences that we observe at this intermediate level of the universe. From this point onwards the terms ‘complexity’ or ‘complex’ and ‘complexics’ or ‘complexical’ (see Bastardas 2016) will be used to refer to a common research framework shared by different disciplines that suggest more all-embracing, contextual and dynamic perspectives. As a transdisciplinary view, the complexity perspective or complexics carries on the perspective of cybernetics: “Cybernetics deals with all forms of behaviour insofar as they are regular, or determinate, or reproducible. The materiality is irrelevant… The truths of cybernetics are not conditional on their being derived from some other branch of science. Cybernetics has its own foundations” (Ashby 1956: 1). Thus, it has a distinct transdisciplinary mission to provide concepts, schema and possibilities of thinking and representation able to express the multidimensional and systemic interwovenness and interdependence of the many, highly significant phenomena of reality which match these characteristics. Contemporary cybernetics began in the 1940s as an interdisciplinar study connecting the fields of control systems, electrical network theory, mechanical engineering, logic modeling, evolutionary biology and neuroscience (concepts related to the biological work of Ludwig von Bertalanffy in General Systems theory).

1 Introduction

3

Why do we talk about cybernetics in a book about linguistics? Because cybernetics is the science of systems in which we find a closed signaling loop—originally referred to as a “circular causal” relationship—following the circular causal process of (1) the action carried out by the system generates a change in its environment, (2) that change is, at the same time, reflected in the System (in the form of feedback) and finally (3) triggers a change in the whole system. Both the structural patterns of this process and the fields of study of cybernetics are to be found in human linguistic interactions, as well as in social systems: the study of feedback, black boxes and derived concepts such as communication and control in living organisms, machines and organisations including self-organisation. Moreover, central concepts such as learning, cognition, adaptation of learning, adaptation, social control, emergence, efficacy, convergence communication, efficiency, and connectivity, notions of realimentation, regulation, communication, autodirection and autocontrol, become crucial when society is understood as a sociocultural adaptative complex system (Buckley quoted by Parra Luna 1992: 387–390). From this perspective, it is possible to elucidate the foundations of Complexity theory. In fact, the discipline that dealt more with complex systems came to be Cybernetics, which has been defined by some authors as “complexity science” (Ashby 1956; Simon 1990). Studies in cybernetics, thus, provides means for examining the design and function of any system, including social systems, seeking to make them more efficient and effective. Cybernetics (in the sense we use it today) was lined up by Norbert Wiener in his book Cybernetics Or Control and Communication in the Animal and the Machine, where he points out that the behaviour of some systems “may be interpreted as directed to the attainment of a goal” (Wiener 1954: 89). Wiener popularized the social implications of cybernetics, drawing analogies between automatic systems (such as a regulated steam engine) and human institutions. The multiple definitions with which cybernetics is aligned are due to the richness of its conceptual base, founded in the common denominator of all its meanings: circularity. In order to synthesize the diverse constellation in which cybernetics is inscribed, we have selected particularly interesting insights adopted by different authors: while M. Mead (1968) states that “a form of cross-disciplinary thought which made it possible for members of many disciplines to communicate with each other easily in a language which all could understand”, Pask (1992) defends that “Cybernetics is the science of defensible metaphors”. Kolmogorov1 conceived it as the “science concerned with the study of systems of any nature which are capable of receiving, storing and processing information so as to use it for control”. Bateson (1972), epistemologist and anthropologist, considered cybernetics as a “branch of mathematics dealing with problems of control, recursiveness and information”. Rodney E. Donaldson,2 first president of the American Society for Cybernetics, described it as an “art of the understanding of understand-

1 Quotation 2 See

from Stanford Encyclopedia of Philosophy (online: https://plato.stanford.edu/). Footnote 1.

4

À. Massip-Bonet et al.

ing”. Finally, Louis Kauffman,3 President of the American Society for Cybernetics from 2005 to 2008, defined the discipline as “the study of systems and processes that interact with themselves and produce themselves from themselves.” New cybernetics will be more suited to the organizations which mankind discovers in nature, as biologists Maturana and Varela (2004) explain in their well-known book: The science and art of understanding. One characteristic of the emerging new cybernetics considered in that time by Felix Geyer and Hans van der Zouwen, according to Bailey (1994), was “that it views information as constructed and reconstructed by an individual interacting with the environment”. This provides an epistemological foundation of science, by viewing it as observer-dependent, that connects cybernetics directly with complexity theory. In this sense, we draw the lines whereby cybernetics can be conceived as the intellectual and scientific thought that gave rise to the actual complexity frame. Indeed, what the complexical perspective first undertook was to absorb the progress already made in disciplines such as physics—e.g., relativity and quantum theory—and biological ecosystems, as well as the foundations of cybernetics (Wiener 1948; Ashby 1956), as it has been explained in the previous paragraphs, and systems theory (Bertalanffy 1969). In the field of human and social sciences, the movement has been equally prevalent, although it has perhaps had less impact, despite the contributions of Gregory Bateson (1972), Morin (1973) and Elias (1982, 2000), whose works are central to the perspective applied in the area of human beings (cf. Bastardas-Boada 1996, 2013a, 2014). Others have had a hand in its construction as well. In Catalonia, for example, Munné (1995, 2013) was a driving force behind the creation and application of the perspective of complexity in social psychology, Serrano (1983, 2001) extended cybernetics and information and systems theories to linguistics and communication, and Aracil (1982, 1983) expressly developed an interdisciplinary, historical and discursive perspective in sociolinguistics. The complexity perspective brings together all contemporary efforts in any specific disciplines or by any researchers specifically devoted to constructing tools, procedures, models and concepts intended for transversal application that are aimed at understanding and explaining the most interwoven and dynamic phenomena of reality. This would encompass Edgar Morin’s theories of complex thinking (1992, 1999, 2005, 2007, 2008); the epistemological and theoretical contributions of physicists such as Bohm (1987), Prigogine and Stengers (1979, 1992), Capra (1982, 2002), and Wagensberg (1985), or of cognitive biologists such as Maturana and Varela (1999, 2004), and the proposals of ecologists such as Margalef (1991) and Allen and Hoekstra (1992). It also includes the most recent contributions of Barabási and Albert (1999) and of Solé (2009) in network theory, and of San Miguel et al. (2012), among others, in statistical physics and the study and computer simulation of complex systems. Without doubt, complexics currently lacks an integrated and unified body of theory to enable us to characterise a field in a general, widely agreed-upon manner. Nor can we dispel all doubts about its feasibility as a unifying paradigm, although we are 3 See

Footnote 1.

1 Introduction

5

convinced that we shall see important progress in coming years to confirm the wisdom of this approach. At a minimum, we are already witnessing a series of transversal concepts and models that are not only pushing forward specific disciplines with new images and perspectives that pass between them, but that are also forging a shared scientific lexicon useful in interdisciplinary communication and integration, which are made more difficult by the diversity of terminology. Complexics, as a scientific paradigm, needs to provide a set of principles, concepts and conceptual landscapes that can be applied transversally to distinct areas of knowledge and phenomena of reality, enabling us to gain a much firmer grasp of the complex aspects of their existence than we currently have. For this reason, our aim needs to be, as Morin says, not “to reduce complexity to simplicity, [but] to translate complexity into theory” (1994: 315). The complex or complexical approach is fully aware—as Morin put it—that our theories have the nature of ‘translations’ and not of ‘mirrors’. Bearing full responsibility for our ‘giving rise to a world’ (Maturana and Varela 1999), we must be hugely mindful of the often hidden assumptions that govern our paradigms, that is, the conceptual lenses through which we imagine the world. For example, some points of contrast between traditional scientific thinking and complex thinking can be summed up as in table 1.1.

Table 1.1 Comparison between the traditional and the complexity perspective. Taken from Bastardas-Boada (2014) Traditional perspective Complexity perspective Conceptual reification

There is no science without an observer (centrality of brain/mind)

Territory

Maps (we see by means of concepts and words)

Scientific truth Elements

Provisional theories Elements-and-contexts, interweaving, interdependences, networks

Objects

Events and processes

Steady-state

dynamic flux, change, evolution

Classical logic

Fuzzy logic

Linear causality

Circular and retroactive causality

Either/Or dichotomies

And/both integration and complementarity

Planned creation

Self-organisation and emergence

Unidimensionality

Inter-influential multidimensionality

“Explicate order” (things are unfolded and “Implicate order” (everything is folded into each thing lies only in its own particular region everything; a hologram, where the parts of space) contain information on the entire object) Fragmentation of disciplines

Inter- and transdisciplinarity

Structure, code

Meaningful and emotional Interaction

6

À. Massip-Bonet et al.

1.2 Complexity in the Study of Natural Language The appearance and/or consolidation of these new theoretical perspectives has an impact at the more practical level of methodology. New tools for the conception, apprehension and treatment of the data of experience will need to be devised to complement existing ones and to enable us to make headway towards practices that better fit complexical theories. Linguistics is one of the domains of knowledge that is undergoing a breakthrough under the new complexity paradigm, while Physics has become one of the pilot disciplines. Natural language has been treated with methods coming from Physics, both from a more theoretical and applied approach. On the one hand, we have the contributions of the more theoretical physicists, such as David Bohm, Ilya Prigogine and Fritjof Capra, and on the other hand, the contributions of more quantitativeoriented physicists from the field of statistical physics modelling, such as Murray Gell-Mann (1996), Maxi San Miguel and Albert Díaz-Guilera, for example. It will certainly be useful for us to gain familiarity with both of these major approaches, see their fruitful application in our disciplines and attempt to exploit them in a coherent and integrated manner. However, we must also be cognizant of the peculiarities of human phenomena, which are characterised by the existence not only of purpose and regularity in the control of behaviour, but also by the significant degree of agents’ cognitive and interpretive autonomy and by the powerful influence of the emotional dimension. This differential fact seems to pose a contradiction for the two fundamental orientations of complexics developed to date. On the one hand, the more epistemological and philosophical contributions lead us to postulate the inevitability of taking into account the brain/mind and everything that arises bio-cognitively from it in order to understand complex human behaviours. On the other hand, the proposals put forward by physics and computer science move in the opposite direction, postulating the selection of a few ‘practical’ parameters that can computationally ‘explain’ the observed facts. The human, sociocultural level has special features that make it even more complex, if we compare it to other existing organisations of phenomena. The elements or human ‘agents’, the units of the system that we want to understand, are not themselves simple, but rather the products of an enormous internal and external complexity. Not only do they contain the prior physical, chemical and biological levels, but also, in their interaction with environments and with one another, they develop extraordinary emotional, cognitive and symbolic capacities that enable them to produce social organisations of extremely high complexity. This explains why in developing knowledge about these societies, what has prevailed is a ‘separating’ view of the several domains present. However, such a view leads us to misunderstand the very phenomena that we want to grasp, because it does not enable us to readily capture their dynamic interactions and inter-influences. As Morin says, for instance, “the non-complex perspective of the human sciences, of the social sciences, is to think that there is an economic reality on one side, a psychological reality on another side,

1 Introduction

7

a demographic reality on yet another side, and so on. One thinks that these categories created by universities are realities, but forgets that in the economic, for example, there are human needs and desires” (1992: 92). This can similarly happen to us at the level of language, if in our study of the facts and events, we forget the intrinsic complexity of their existence and production in human beings. If we treat what we call ‘languages’ as if they were simple, decontextualised objects, we can make headway in our understanding of some of their more ‘mechanical’ aspects, but we can also entirely overlook the conditions of their existence, functionality, maintenance, variation, change or disappearance. At the same time, it will become more difficult for us to account for the major constitutive influence that a language has at the cognitive level and we will completely miss the social phenomenon of the continuous use (or the disuse) of languages at all levels of human life. In sociolinguistics and cognitive interactional linguistics, we have had to move towards the use of perspectives and metaphors relating to ecological complexity and complex adaptive systems in order to try to grasp the interdependencies among the different levels of organisation that can affect the determination of language behaviours (Bastardas-Boada 1996, 2013b, c, 2017; Junyent 1992; Ellis and Larsen-Freeman 2009; The Five Graces Group 2009, Massip-Bonet 2013a, b). The brain/mind, habits at the interactional level, demo-social groupings, the socioeconomic structure, the media and political power enter into constant relation with language forms and codes and can determine their course—through the pressures that individuals interpret as being exerted on them (Terborg and García-Landa 2013)—in an interdependent fashion, with conflict and tension that can vary by situation. We must not forget that, even though the public authorities often try to intervene through directed actions, languages are basically dynamic phenomena of social selforganisation and emergence that are interdependent with all of their contexts and not solely with political ones. Ways of speaking and languages are like organisms adapted to their setting and to their function within the whole (that is, to the purposes they must serve within that whole) (Bohm 1987: 37). Thus, if a specific language—or linguistic form—is being left without any communicative function in its society of origin because people are adopting an alternative, either some of its ‘own’ functions will be preserved, or ‘identity-based’, symbolic functions will be created to maintain its use, or ultimately it will be abandoned. Similarly, the speakers, if necessary, will autonomously create new forms and/or develop the existing forms according to their needs and the social meanings that they give to them (Bastardas-Boada 2004). Language and interaction are co-phenomena; the former is within the latter and the latter is within the former. In all likelihood, the conceptual resources and tools that we currently have are not yet entirely suitable to the tasks that must be undertaken. This is why it is necessary to continue developing new theoretical and methodological instruments that are able to help us more adequately imagine and understand the dynamic interweavings of distinct aspects of sociocultural and linguistic events. The challenge stands before us. From the socio-complexical perspective, we ought to strengthen the inter- and transdisciplinary lines of research that, from the biological

8

À. Massip-Bonet et al.

to the socio-political levels, cut across the communicative, linguistic phenomenon, which in turn is part—in its way—of each and every one of the interrelated domains of human life. A complexical, eco-co-dependent and processual view of sociocommunicative events—languages are in societies/cultures and in the brains/minds that are in the languages—can help us push towards a significant deepening in our understanding (Roggero 2013; Ruiz Ballesteros 2013; Vilarroya 2002; Steels 2000; Mufwene 2001, 2013). While every type of language study related to society, and language emergence has captured the attention of the researchers on complex systems, other branches of linguistics have not developed theories with explicative capabilities into this framework. This book is one of the first attempts to extend the paradigm of complexity to some areas of study of natural language that have not followed this theoretical perspective yet.

1.3 Structure of the Book As stated above, this volume adopts a new theoretical position for the study of natural language as a complex entity. Within this framework, where interdisciplinarity and interaction with other sciences are necessary as methodological options, we approach different disciplines of linguistics and introduce different applications. The book is divided into four parts. The first one is a theoretical section explaining why the concept of complexity has an important impact in human sciences, cognition and linguistics. After that, three more blocks are introduced, corresponding to three fields of linguistics—language change, sociolinguistics and a broader area including some developments in syntax, semantics and cognition. The first part, “Interdisciplinary approaches for Human Sciences”, consists of four contributions, which are oriented to philosophy of science, interdisciplinary approaches to complexity and historical interpretations of scientific theory. Robert Hristovsky, Natalia Balagué and Pablo Vázquez develop the idea that sciences are social self-organizing adaptive cognitive systems. They explain the rise of unifying themata in science overcoming the fragmentation of scientific language and illustrate the diversification and unification of scientific language with examples of different disciplines such as cosmology, chemistry, psychology and physics, among others. Alvaro Malaina presents what he calls the “paradigm of complexity”, from the Khun perspective, as a paradigm that incorporates both a worldview and models of scientific realizations. He proposes the integration between “general complexity” and “restricted complexity”, studying the implications of this process in sociolinguistics. Leonardo Rodríguez Zoya proposes a model to approach the principles organizing a thought system. To address the issue, he suggests a qualitative and quantitative study of scientific beliefs, understood as a form of social cognition produced through social practices and discourses.

1 Introduction

9

After theoretical foundations, the book introduces more concrete linguistic applications, starting with a second block especially focused on three models approaching language change. Àngels Massip-Bonet provides a general theoretical introduction to the topic of linguistic variation and change adopting the perspective of complex adaptive systems. The paper draws the main implications of taking the paradigm of complexity as a methodological framework and highlights the general lines of research that can be developed in the area. Enrique Bernárdez explains the idea of ‘macrochange’ in natural language, built in parallel to the concept of ‘macroevolution’ in evolutionary theory. This perspective of language change implies the consideration of language as a complex natural phenomenon, including not only structures and usage, but the whole ‘ecological niche’ where it exists and is in use by human beings in specific cultural and historical situations. Gemma Bel-Enguix studies language emergence and change using the applications and tools provided by agents theory, complex systems and simulations. She presents the results of some experiments that demonstrate how social structures influence language evolution. The third part of the book brings together three papers introducing several developments in sociolinguistics from the paradigm of complexity. Albert Bastardas-Boada proposes to take an ecological framework and bring sociocomplexity into the study of language contact. He provides a survey of the restricted and general perspectives of complexity that have been adopted to tackle natural language, and claims for the integration of both views for a more complete picture of the factors that affect language behaviour and evolution. Léo Léonard and colleagues provide a more specific case study, based on Mazatec dialects, an endangered Otomanguean language spoken in south-east Mexico by about 220,000 speakers. Closing the works on the area, Roland Terborg and Virna Velázquez tackle the problem of language and common knowledge from an ecological perspective. Individuals develop physical and history together. By doing it, they modify the state of the world. In the same way, language and human interaction modify knowledge of agents in a way that ideologies, values and beliefs can emerge. After the contributions dealing with the social aspects of language, three more papers deal with the issues related to discourse analysis. Esperanza Morales López explains how discourse analysis can take advantage of some postulates of complexity, among them the holistic perspective and transdisciplinarity. She highlights the fact that both features can boost the area with new methods and tools. In a similar range of thinking, García Riverón and Marrero Montero make use of a holistic perspective of research that leads them to find a new phonological interpretation of the prosody, lexical, grammar and speech systems in virtue of intention through speech multidimensional analysis. In this way, they describe the bases of a group of semantically and pragmatically founded attractors defined for the study of

10

À. Massip-Bonet et al.

intonation and a new concept of underlying structure formed by emergent features from the morphogenetic processes of the language systems. The last contribution in the area of discourse is conducted by Sami Alim. The author offers a study of the so-called raciolinguistics in the speech of Barack Obama, giving a number of examples of that kind of performances in ethnoracial contexts. The final part of this publication consists of several contributions on syntax, semantics and cognition. Michael Zock presents the problem of lexical access, the tip-of-the-tongue problem and cognition. Obviously, the human brain is a complex object and so is the process of accessing words in the mental lexicon. The goal of the paper is to describe a method that, once implemented, should help people to overcome the ToT problem. Lluís Barceló-Coblijn, Maia Duguine and Aritz Irurtzun connect the emergence of functional words as hubs in L1 acquisition with the DP theory in transformational grammar, with the help of graph theory. Dariusz Plewczynski and his coworkers offer a study of culturally-driven emergence of color categories, extending a model by Steels and Belpaeme (2005). They bring the discussion to the process of modeling the emergence of perceptual categories in human subjects. Closing the volume, Maria Antònia Font develops a review of the concepts of imagination, image schema, neural image and mental image, and discusses their implications in conceptualization about emotions.

References Allen, T. F. H., & Hoekstra, T. W. (1992). Toward a unified ecology. New York: Columbia University Press. Aracil, L. V. (1982). Papers de sociolingüística. Barcelona: La Magrana. Aracil, L. V. (1983). Dir la realitat. Edicions Països Catalans. Ashby, W. R. (1956). An introduction to cybernetics. London: Chapman & Hall. Bailey, K. (1994). Sociology and the new systems theory: Toward a theoretical synthesis. New York: New York State University. Bailey, K. D. (1994). Typologies and taxonomies: An introduction to classification techniques. Newbury Park: Sage cop. Barabási, A.-L., & Albert, R. (1999). Emergence of scaling in random networks. Science, 286, 509–512. Bastardas-Boada, A. (1996). Ecologia de les llengües. Medi, contacte i dinàmica sociolingüística [From language shift to language revitalization and sustainability: A complexity approach to linguistic ecology]. Barcelona: Proa; Barcelona: Publicacions de la Universitat de Barcelona (in press). Bastardas-Boada, A. (2004). Sociolingüística versus política y planificación lingüísticas: distinciones entre los campos y nociones integradoras. Revista de llengua i dret, 41, 175–194. Bastardas-Boada, A. (2013a). Complexitat i fenomen (socio)lingüístic. LSC—Llengua Societat i Comunicació, 11, 5–13. Bastardas-Boada, A. (2013b). Sociolinguistics: Towards a complex ecological view. In A. MassipBonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 15–34). Berlin: Springer.

1 Introduction

11

Bastardas-Boada, A. (2013c). General linguistics and communication sciences: Sociocomplexity as an integrative perspective. In A. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 151–173). Berlin: Springer. Bastardas-Boada, A. (2014). Towards a complex-figurational socio-linguistics: Some contributions from physics, ecology and the sciences of complexity. History of the Human Sciences, 27(3), 55–75. https://doi.org/10.1177/0952695114534425. Bastardas-Boada, A. (2016). Complexics as a Meta-Transdisciplinary Field. Congrès Mondial pour lapensée complexe. Les défis d’un monde globalisé. (Paris, 8-9 décembre. UNESCO). Available online at https://www.reseaucanope.fr/fileadmin/user_upload/Projets/pensee_complexe/ bastardas_boada_complexics_meta_transdisciplinary_field.pdf Bastardas-Boada, A. (2017). Complexity in language contact: A socio-cognitive framework. In S. Mufwene, et al. (Eds.), Complexity in language: Developmental and evolutionary Perspectives (pp. 218–243). Cambridge: Cambridge University Press. Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine Books. Bertalanffy, L. Von. (1969). General system theory. New York: George Braziller Inc. Bohm, D. (1987). La totalidad y el orden implicado. [Spanish translation of Wholeness and the Implicate Order. London: Routledge & Kegan Paul, 1980]. Barcelona, Kairós. Buckley, W. (1967). Sociology and modern systems theory. Englewood: Prentice-Hall, NJ. Capra, F. (1982). El punto crucial. Barcelona, Integral ed. [Spanish translation of The Turning Point. New York, Simon & Schuster, 1982)]. Capra, F. (2002). The hidden connections. New York: Doubleday. Elias, N. (1982). Sociología fundamental. Barcelona, Gedisa [Spanish translation of Was ist Soziologie? Munich, Juventa Verlag, 1970]. Elias, N. (2000). The Civilizing Process. Oxford, Blackwell. [English translation of Über den Prozess der Zivilisation. Soziogenetische und psychogenetische Untersuchungen. Basel, Haus zum Falken, 1939]. Ellis, N. C., & Larsen-Freeman, D. (Eds.). (2009). Language as a complex adaptative system. Oxford: Blackwell Publishing. Gell-Mann, M. (1996). El Quark y el jaguar. Aventuras en lo simple y lo complejo [The Quark and the Jaguar: Adventures in the Simple and the Complex]. Barcelona, Tusquets. Junyent, C. (1992). Vida i mort de les llengües. Barcelona: Empúries. Margalef, R. (1991). Teoría de los sistemas ecológicos. Barcelona: Publications of the University of Barcelona. Massip-Bonet, A. (2013a). El llenguatge: una visió des de la teoria de la complexitat, LSC—Llengua, societat i comunicació 11, 20–24 (Monograph on ‘Language and Complexity’). Massip-Bonet, A. (2013b). Language as a complex adaptive system: Towards an integrative linguistics. In Massip-Bonet & Bastardas-Boada. Massip-Bonet, A., & Bastardas-Boada, A. (Eds.). (2013). Complexity perspectives on language, communication and society. Springer: Heidelberg. Maturana, H., & Varela, F. J. (1999). El árbol del conocimiento. Las bases biológicas del conocimiento humano. Madrid: Editorial Debate. Maturana, H., & Varela, F. J. (2004). De máquinas y seres vivos. Autopoiesis: la organización de lo vivo. Buenos Aires: Lumen. Mead, M. (1968). Cybernetics of cybernetics. In H. von Foerster, J. White, L. Peterson, & J. Russell (Eds.), Purposive systems. New York: Spartan Books. Morin, E. (1973). Le paradigme perdu: la nature humaine. Paris, Éditions du Seuil. [Spanish translation of El paradigma perdido. Ensayo de bioantropología. Barcelona, Kairós, 1974]. Morin, E. (1992). Introduction à la pensée complexe. Paris, ESF. [Spanish translation of Introducción al pensamiento complejo. Barcelona, Gedisa, 1994]. Morin, E. (1994). La complexité humaine. Paris: Flammarion. Morin, E. (2005). Complexité restreinte, complexité générale, in: Colloque “Intelligence de la complexité: épistémologie et pragmatique”, Cerisy-La-Salle, 26 June 2005. Available online at http://www.intelligence-complexite.org/fileadmin/docs/1003morin.pdf.

12

À. Massip-Bonet et al.

Morin, E. (2007). Complexité restreinte, complexité générale. In J. L. Le Moigne & E. Morin (Eds.), Intelligence de la complexité. Épistémologie et pragmatique (pp. 28–50). La Tour d’Aigues, Éditions de l’Aube. Morin, E. (2008). La méthode. Paris: Éditions du Seuil. Morin, E., & Le Moigne, J. L. (1999). L’intelligence de la complexité. Paris: L’Harmattan. Mufwene, S. (2001). The ecology of language evolution. Cambridge: Cambridge University Press. Mufwene, S. (2013). The emergence of complexity in language. In A. Massip-Bonet & A. BastardasBoada (Eds.), Complexity perspectives on language, communication and society (pp. 197–218). Heidelberg: Springer. Mufwene, S. S., Coupé, C., & Pellegrino, F. (2017). Complexity in Language. Developmental and Evolutionary Perspectives. Cambridge: Cambridge Univerity Press. Munné, F. (1995). Las teorías de la complejidad y sus implicaciones en las ciencias del comportamiento. Revista Interamericana de Psicología, 29(1), 1–12. Munné, F. (2013). The fuzzy complexity of language. In A. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 175–196). Heidelberg: Springer. Parra Luna, F. (1992). Elementos para una teoria formal del sistema social: una orientación crítica. Madrid: Editorial Complutense. Pask, G. (1992). Introduction different kinds of cybernetics. In: van de Vijver G. (Ed.), New perspectives on cybernetics (Vol. 220). Synthese Library (Studies in Epistemology, Logic, Methodology, and Philosophy of Science). Dordrecht: Springer. Prigogine, I., & Stengers, I. (1979). La nouvelle alliance. Métamorphose de la science. Paris: Gallimard. Prigogine, I., & Stengers, I. (1992). Entre le temps et l’éternité. Paris: Flammarion. Roggero, P. (2013). Para una sociología según El método. In E. Ruiz Ballesteros & J. L. Solana Ruiz (Eds.), Complejidad y ciencias sociales (pp. 103–123). Seville: International University of Andalusia. Ruiz Ballesteros, E. (2013). Hacia la operativización de la complejidad en ciencias sociales. In Esteban Ruiz Ballesteros, José Luis & Solana Ruiz (Eds.), Complejidad y ciencias sociales (pp. 137–172). Seville: International University of Andalusia. San Miguel, M., Johnson, J. H., Kertesz, J., Kaski, K., Díaz-Guilera, A., MacKay, R. S., et al. (2012). Challenges in complex systems science. The European Physical Journal Special Topics, 214, 245–271. Serrano, S. (1983). La lingüística. Su historia y su desarrollo. Barcelona: Montesinos ed. Serrano, S. (2001). La semiótica. Una introducción a la teoría de los signos. Barcelona: Montesinos ed. Simon, H. A. (1990). Invariants of Human Behavior. Annual Review of Psychology, 41, 6. Solé, R. (2009). Redes complejas. Del genoma a Internet. Barcelona: Tusquets. Steels, L. (2000). Language as a complex adaptive system. In M. Schoenauer, et al. (Eds.), Lecture notes in computer science. Parallel problem solving from nature—PPSN-VI. Berlin: SpringerVerlag. Steels, L., & Belpaeme, T. (2005). Coordinating perceptually grounded categories through language: A case study for colour. Behavioural and Brain Science, 28(4), 469–489. Terborg, R., & García-Landa, L. (2013). The ecology of pressures: Towards a tool to analyze the complex process of language shift and maintenance. In A. Massip-Bonet & A. BastardasBoada (Eds.), Complexity perspectives on language, communication and society (pp. 219–239). Heidelberg: Springer. The ‘Five Graces’ Group: Beckner, C., Blythe, R., Bybee, J., Christiansen, M.H., Croft, W., Ellis, N.C., Holland, J., Ke, J., Larsen-Freeman, D., Schoenemann, T. (2009). Language is a complex adaptive system: Position paper, Language Learning 59(Supplementary 1), 1–26. Vilarroya, O. (2002). La disolución de la mente. Barcelona: Tusquets editores. Wagensberg, J. (1985). Ideas sobre la complejidad del mundo. Barcelona: Tusquets editores.

1 Introduction

13

Wiener, N. (1948). Cybernetics or control and communication in the animal and the machine. Cambridge, MA: The MIT Press. Wiener, N. (1954). The human use of human beings: Cybernetics and society. Boston: HoughtonMifflin.

Part I

Interdisciplinary Approaches for Human Sciences

Chapter 2

Science as a Social Self-organizing Extended Cognitive System. Coherence and Flexibility of Scientific Explanatory Patterns Robert Hristovski, Natàlia Balagué and Pablo Vázquez Abstract We conceptualize science as a social cognitive embodied-extended system with a perpetual action-perception-explanatory pattern formin cycle. This cognitive cycle encompasses the natural environment. The cycle is irreducible to inner cognitive processes of scientists. Its technologically embodied-extended nature necessarily makes the cognitive cycle to be context dependent, bringing about context dependence in the explanatory part shared via language. Despite of the context-dependence of scientific practices the past decades have witnessed a large-scale diffusion of explanatory concepts, i.e. themata, coming from dynamical systems theory and statistical physics into science fields which, till then, seemed totally disconnected. This trend increases the coherence of explanatory patterns and consequently enhances and diversifies the language communication possibilities between scientific practices. The structure that emerges is one which, on the one hand, possesses explanatory stability, that is, a coherent and pluri-contextual explanatory backbone that co-relates classically independent or weakly dependent scientific fields, and on the other hand, allows context-dependent flexibility and adaptivity of explanatory patterns to specific processes it strives to understand. The picture that emerges reveals the science as a social self-organizing adaptive cognitive system.

R. Hristovski (B) Faculty of Physical Education, Sport and Health, Ss. Cyril and Methodius University, 1000 Skopje, Macedonia e-mail: [email protected] N. Balagué · P. Vázquez National Institute of Physical Education (INEFC) Health and Applied Sciences, University of Barcelona, Gran via de les Corts Catalanes, 585 08007 Barcelona, Spain e-mail: [email protected] P. Vázquez e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_2

17

18

R. Hristovski et al.

2.1 Introduction. Science as an Embodied-Extended Social Cognitive System Science is a collective, social, cognitive endeavor. It seeks positive explanatory patterns. It seeks answers to questions such as what something is and what causes it. This epistemic dimension is not only a property of the highly abstract realm of science, but also part of the most basic pre-reflective biological perception-action pattern-forming systems (Gibson 1979; Araújo et al. 2006). Exercising an action in the world must correspond to the properties of the environment. The ability to detect these properties is the core of discovering law-like patterns in the world (Turvey 1992). Without it one can hardly imagine the survival of living systems. Actions and perceptions contain an epistemic dimension (Kirsh and Maglio 1994) on a basic, pre-reflective level. They test and adapt to the states of the world. They form the basis of biological cognition (Bruineberg and Rietveld 2014) . But for such law-like reproducible patterns to be detected, they must exist in some form. The discovery and systematization of such patterns on a reflective and conceptual level is what science predominantly does: it produces explanatory patterns that we call models. How the normative pre-reflective epistemic actions transform into reflective epistemic actions is one of the important questions that have to be answered in future. In what follows, we will consider science practices as a coordinated social cognitive system of action-perception-modeling processes, in which language plays an important role. We will also make some observations on ongoing self-organization within the modeling part of the system. Self-organization occurs despite the fact that there are substantial differences in the action-perception sub-systems that define scientific practices. Self-organization underpins the formation of stable, yet flexible, explanatory patterns, which is a hallmark property of adaptive systems. Scientific practices entail interactions with the environment through the design and performance of observations and experiments in order to explore, provide or test models of the world. It includes active engagement with the world. Therefore, the science-natural environment system, not merely science itself, could be the meaningful unit of consideration of science as a collective cognitive system. The differences between sciences lie in differences in the directedness of their perception-action systems, i.e. their attentional focus. The system contains a loop (see Fig. 2.1): science acts on a specific segment of the natural environment (be that proton-proton collisions or the dissemination of questionnaires), and then perceives and models the effects of that action, which in turn leads to another epistemic action. The effects of a scientific action are the active back-reactions, i.e. responses, of the natural environment to the action that has been applied. Actions, modeling and perceptions of this kind of collective cognition are provided by more or less sophisticated technological devices. Scientific cognition is, thus, embodied. It is embodied intra- and extra-somatically. The former refers to the biological body of cognizers, and the latter to technological devices. Technological devices are extensions of our bodily action and perceptual systems. Possible actions, modelings and perceptions depend on the sophistication, i.e. effectivities or abilities,

2 Science as a Social Self-organizing Extended Cognitive System …

19

Fig. 2.1 The science-natural environment loop of extended-embodied cognition. The natural environment has an active participatory role within the cognitive system

of the embodiment. Technological embodiment evolves through internal changes in a self-organizing fashion. In this way, science may be construed as an evolving extended-embodied social cognitive system. Extended cognition (see e.g. Clark and Chalmers 1998; Froese et al. 2013) concerns the suggestion that extra-somatic entities, not only internal brain processes, can be considered part of the cognitive system in cases when they reliably aid and enhance cognitive performance. Cognition is distributed rather than located within the agent’s brain. Here we assume that the natural environment itself can be considered part of a scientific collective cognitive system that consists of certain reproducible, invariant properties, so that whenever a certain action (e.g. an experiment) is performed by the scientific social system within a certain context (the boundary conditions) the scientific sensorium perceives similar events. This is the basis of (deterministic or statistical) natural laws, and the even deeper principles that connect them. In this way, the environment or nature may be considered an extended part of our social remembering process, i.e. cognition. The environment, i.e. nature, significantly drives scientific cognition. It strongly constrains the evolution and development of scientific cognition. Scientific cognition is more accurate, and in fact, it is possible only when coupled to the environment through the action-perception loop than otherwise. Natural processes have an ineliminable modifying role in scientific cognition. Evidence for this can be found in the difference between the pre-scientific and scientific era. If we remove all of the body of law-like relations that have been collectively discovered to date and the possibility of scientific epistemic action in the world, we find pre-scientific views waiting in the wings. In this case, extant collective cognition will be significantly impaired. In contrast, when we allow epistemic actions in and responsiveness from the world, the same law-like relations can be discovered anew. In this sense, the natural environment is an important and active part of scientific collective cognition. The natural environment is a reliable source of that information. Information can be reliably retrieved by repeatable clever manipulations. It can provide us with the information we seek for whatever the cognitive collective needs and decides. It functions as a memory aid. It virtually exists out there, and is available here and now as an information resource. In this way, a stable, coupled, distributed cognitive system is formed. This system is irreducible to inner mental processes of scientists. Of course, when we ponder the ‘here and now’ term, we must

20

R. Hristovski et al.

keep the sense of time scales in our considerations of environmental coupling within the social cognitive system. Social systems are more inert than individual agents, and characteristic time scales defining the ‘here and now’ are different. They may proceed on the level of seconds, but also on that of tens of years, depending on the specific extended social cognitive processes that are addressed. Social communication within and between science collectives takes place through language (Sutton et al. 2010; Wilson and Clark 2009). More generally, science collectives may be viewed as an embodied-extended cognitive system in which language is of great importance. It is important as a tool not only for articulating the intrinsic thought processes of individual agents, but also for coordinating the (actionperception-modeling) system on the collective, social level. This is the second level on which scientific cognition is extended/distributed. Scientific explanatory patterns are not confined within individual brains, but are shared, i.e. distribuited, among agents. Moreover, this language-based sharing profoundly aids problem-solving tasks, in the sense that isolated agents focused on their internal thoughts would often have impaired cognitive performance. Explanatory patterns can be defined as emergent entities since no elementary variable belonging to it contains the property of the pattern itself. Their meaning is emergent since it depends on the relations between elementary variables which are, in turn, part of the perception-action cycle of scientific practices. They may be considered as coordinative patterns since they entail co-ordinated elements and arise through the process of linguistic coordination (Raczaszek-Leonardi and Kelso 2008; Fusaroli et al. 2013; Raczaszek-Leonardi et al. 2014). They are also task specific because they are formed to satisfy certain scientific task goal constraints. Hence, explanatory patterns can be conceptualized as emergent task specific (functional and thus meaningful) dynamic coordinative language patterns at a social level, akin to their conceptualization at an individual level of the learner subject to a guided discovery learning process (Hristovski et al. 2014). Another language-based sharing mode of extended-embodied cognition, underpinned by the previous one, is important. The absence of linguistic information in printed or electronic format in journals, books, on the Internet or on computer displays of the scientific sensorium during experiments would utterly diminish the cognitive abilities of any scientific community. These embodiments serve as memory aids within the science community. It is important to note that all these forms of embodied-extended social cognition include context sensitivity. The form of technological embodiment and epistemic actions are all task-specific. They depend on the cognition of particular aspects of environmental states, which in turn determine the types of embodiment-extension pairs. To summarize, scientific collective cognitive practices can be considered cyclically coordinated action-perception-modeling activities of an embodied-extended social cognitive system (Fig. 2.1). Their extra-somatic embodied nature is reflected in the use and evolution of technology. The evolution of technology, i.e. the extrasomatic embodiment, may be construed as a special type of self-organization arising from intrinsic dynamics within the science-environment coupling. Its evolving, extended nature is at least two-fold. The social cognitive system, through its embodied

2 Science as a Social Self-organizing Extended Cognitive System …

21

properties, entails a reliable coupling between socially shared cognition and different properties of the natural environment. The natural environment and its law-like properties is the part of the cognitive system that has the role of a reliably accessible memory aid, while socially shared cognition itself entails cognitive extension coordinated by language. All the intertwined and interdependent levels of scientific embodied-extended cognition are context-dependent; they focus on different aspects of states of the natural environment, and hence their technological embodiment and cognitive extensions differ widely. Here some important questions arise concerning the areas of science communication and language: 1. Since successful cognitive extension in social systems presupposes a common language for communicating and using the information, can it spread in all scientific communities from physics to sociology, given the fact that scientific practices are heavily context-dependent? In other words, can scientific explanatory patterns be formulated in such a way that any scientific community can understand the core principles from another science, at least at basic level, so that positive transfer can occur in science learning? 2. Can this be achieved without utterly homogenizing the scientific language, which would contradict the context-dependency of different science-environment couplings? 3. Are positive answers to the above questions possible at all, bearing in mind that the differences in which the sciences as collective cognitive systems are embodied and extended lead to context-sensitivity, and consequently fragmentation, in the action-perception-modeling cycle of the science-environment coupling?

2.2 Fragmentation of Scientific Language Fragmentation of scientific language arises from the fragmentation of scientific practices, and this in turn is a consequence of different properties of matter at the different layers of organization on which a particular science action-perception-modeling system is focused. Due to the coupling of science and the environment, the environmental levels of organization map onto the scientific practices, and consequently onto the scientific language space. The diversity of phenomena constrains the scientific practices of each discipline to form a specific vocabulary for explanatory patterns of natural properties and processes, and to communicate the knowledge among scientists. This is a major reason for the formation of context-dependent scientific vocabulary. For example, a cosmologist, cell biologist, sports scientist and sociologist would hardly understand each other when they explained the basic processes within their fields. In turn, this disables the possibility of sharing and transferring knowledge between scientific communities, which is the main mechanism of language-based extended social cognition. In other words, extended cognition remains condensed in small and fragmented pockets of the scientific society. The diversity of the phenomena

22

R. Hristovski et al.

and properties of organized substance was recently ascribed to the existence of what are known as “mesoscopic protectorates”(see e.g. Laughlin et al. 2000), i.e. emergent organization levels of substance, whose key properties cannot be formally, i.e. mathematically, deduced from the laws that govern the behavior of the more microscopic components (e.g. Mainwood 2006). Hence, each upward or downward level possesses novel structures and properties that need a specific language for dealing with them. Therefore, these languages use context-dependent concepts to name and explain the process under scrutiny. Context dependence is essentially viewed as the major cause of fragmented vocabulary among scientific practices. In other words, knowledge communication and, hence, language-driven extended cognition, is easy within specific scientific fields and subfields, as it is enabled by common vocabulary. However, it becomes increasingly difficult between more distant disciplines.

2.3 The Rise of Unifying Themata Indeed, given this state of affairs, one may well ask why there are not as many totally disconnected models as there are different phenomena in the world. Logically, this is not an absurd question. The difference in scientific practices underpinned by different types of science-environment cognitive couplings would support exactly this state of affairs. Why isn’t there a totally different model for each phenomenon, akin to polytheistic worldviews where behind each phenomenon acts a separate god or goddess? However, the fact is that models of different levels of substance organization are increasingly being based on some general themata.1 The hypothesis that themata are artifacts of mathematical coincidences can be ruled out by a simple chi-squared test. The problem remains if we try to explain this fact by stating that these themata are used as metaphors due to the economy of thinking, because we would then ask why these and not a vast number of other metaphors survive the different levels of organization. These (and possibly a relatively small number of other) themata show a high degree of context-independence. They are context-free. In a sense, they are also pluri-contextual, since they play a significant role in explanatory patterns at each level (context) of organized matter. The fact is that, in the past several decades, explanatory patterns that capture different levels of organized matter by means of common, universal, dynamical context-free concepts have been increasingly successful. The search for minimum principles that explain the maximum number of phenomena is a tacit motive in all sciences, at least in theoretical disciplines. In addition, this has proven very successful at theoretical and empirical levels. Newton started this trend by unifying disparate phenomena such as celestial mechanics, earthly tides and falling bodies into the law of universal gravity. It was the first time that these phenomena were expressed in a dynamical form. The trend was followed by Maxwell, who unified the apparently 1 Themata, mostly in physics, were explored and discussed extensively by Gerald Holton. However,

he discusses only a few of the general themata from this chapter.

2 Science as a Social Self-organizing Extended Cognitive System …

23

distinct phenomena of light, electricity and magnetism into a unified electromagnetic dynamical theory, and shortly after by the unification of notions of space, time, matter, energy and gravity into an even more general dynamical theory by Einstein. He started a program of unifying the fundamental forces of nature in a framework today known as “The Grand Unification”. It is this level that some physicists, (e.g. Weinberg 1995) think of as the proper level of explanatory unification (but see Anderson 1972). Another type of unifying tendency that is closely related in its rationale to the above emerged at the start of the second half of the twentieth century, with studies in condensed matter physics (Landau 1969). In this trend, sciences are connected by finding deep analogies among seemingly disparate phenomena. At the start of the 1960s, a very fruitful link was established between statics and the dynamics of condensed matter physics and elementary particle physics (e.g. Englert and Brout 1964; Higgs 1964). Around that time, this trend was also present in non-equilibrium physics (Haken 1964). The successes of the explanation of phase transitions in the late 1960s and early 1970s (Kadanoff 1966; Wilson 1975) was paralleled by similar progress in non-equilibrium chemistry (Glansdorff and Prigogine 1971), and shortly afterwards in nonlinear dynamics (e.g. Cvitanovi´c et al. 1984). Initial explanatory patterns containing unifying themata on the organization of animal and human movement (Kelso et al. 1979)and the sociology of groups had already appeared by the end of the 1970s and in the early 1980s (Isnard and Zeeman 1974). The advent of neural networks and protein folding models strongly boosted the spread of these ideas in general biology (Hopfield 1982; Wolynes et al. 1995; Pollack and Chin 2008). It is interesting that network explanatory patterns are not only applied to higher level phenomena such as those in biology or sociology, but also to microscopic physical processes like quantum gravity and field theory (see e.g. Rovelli and Smolin 1995). What distinguishes this type of unifying tendency from the previous one is not the reduction of the world to microscopic fundamental physical laws as a unifying ground for science, as Weinberg (1995) sees it, but the survival of general principles at each level of existence of the world, including the microscopic level of basic physical laws. The discovery of fundamental microscopic physical laws would not bring about a unified understanding of phenomena at different levels in nature without destroying them. In contrast, the discovery and use of general principles at each level of existence of a general or fundamental theory may be formulated. The trend of applying general explanatory patterns in diverse fields of research is growing at an increasing rate, as if it is being guided by a positive feedback loop, which is suggestive of spontaneous self-organizing phenomena.

2.4 The Self-organization of Scientific Explanatory Patterns A preliminary study was recently conducted to form a clearer picture of the ongoing self-organizing process within scientific explanatory patterns and their properties, including the degree of mutual coherence and flexibility (Hristovski 2013; Hristovski

24

R. Hristovski et al.

et al. 2014). The conceptual content of explanatory patterns in high school and higher education textbooks and contemporary scientific modeling papers were compared in 10 traditionally disparate science fields. First, the vocabularies of the science fields were extracted from high school and higher education textbooks. Then, an Internet search was carried out in relevant databases of pairs of these vocabularies with the themata concepts. The following themata were used: self -organization (self assembly or soft-assembly), collective modes (order parameter, collective coordinate or variable, reaction coordinate), control parameter or variable, phase transition, bifurcation, symmetry-symmetry breaking, stability, instability (loss of stability), metastability, criticality (critical point or manifold), gradients, scalar field, vector field, attractor, repeller, entropy-information, and network. The search in relevant databases and the co-word analysis showed that the extracted science papers consisted of mathematical models, theoretical reviews and position papers, because these were the most likely media to include all three sub-systems of the action-perceptionmodeling cycle of science-environment coupling. Most of the purely empirical papers referring to the experimental, i.e. the action-perception, component of that cycle did not contain clear, unequivocal references to the general modeling issues. The degree of inter-pattern coherence was defined as the overlap value between the vocabularies of any two science fields. The inter-pattern coherence was then subjected to principal component analysis, and the structure obtained under the presence and absence of general themata was then compared. Remarkable differences were found between the conceptual structure of high school and higher education textbooks on the one hand, and modeling scientific papers on the other (see Figs. 2.2 and 2.3).

Fig. 2.2 Basins of attraction and passages between science fields in the absence of general themata, in the space spanned by the first and the third principal components. Communication between fields is constrained to a narrow channel between conceptually neighboring disciplines. A state of low linguistic coherence. EFP—elementary fields and particle physics, CL—cosmology, MP—molecular physics, CR—chemical reactions, CB—cell biology, NB—neurobiology, MB—motor behavior, PP—psychological processes, CS—collective sports, and SG—sociology of groups. Legend: linguistic overlap between science fields (with kind permission from Research in Physical Education, Sports and Health)

2 Science as a Social Self-organizing Extended Cognitive System …

25

Fig. 2.3 Basins of attraction and saddle points between scientific fields in the presence of general themata in the space spanned by the first two principal components. In contrast to Fig. 2.2, the communication between fields is now spread on a surface, due to lowered language barriers. A state of higher linguistic coherence. EFP—elementary fields and particles physics, CL—cosmology, MP—molecular physics, CR—chemical reactions, CB—cell biology, NB—neurobiology, MB—motor behavior, PP—psychological processes, CS—collective sports, and SG—sociology of groups. Legend: linguistic overlap between science fields (with kind permission from Research in Physical Education, Sports and Health)

To clarify, Figs. 2.2 and 2.3 may be interpreted as potential landscapes in which the dark gray regions represent deep valleys (basins of attraction) belonging to a certain scientific vocabulary, and the pale and white regions are the high plateaus of the landscape that maximally separate each of the two scientific vocabularies. All other gray shades represent possible passages from one vocabulary to another that require a less than maximal conceptual (informational) change for mutual communication. Passing from one valley to another means transforming the vocabulary (passing over the hill or barrier) for some quantity proportional to the distance d, as defined later in the text. In other words, it roughly corresponds to the communication or learning effort required to acquire the vocabulary of other science fields. The absence of general themata, which is typical in high school and some higher education textbooks, led to the compression of the original 10 science vocabulary vectors into four principal components. Information compression through a reduction of dimensionality was mostly due to the language similarities between neighboring science practices. The primary principal components (PC) were weakly correlated and resulted in one secondary PC, which was saturated mostly by natural science concepts, while other scientific disciplines shared less information with this component, which had low projections. More information on the interesting structure of the PCs is given in Hristovski et al. (2014). In the absence of general themata, inter-science conceptual communication forms a narrow channel and exists only as a consequence of linguistic nearest-neighbor interactions. In this case, there are maximal barriers between other disciplines, as shown by the pale white plateau (see Fig. 2.2). In network parlance, this structure corresponds to a formed giant component. The dominance of the nearest-neighbor

26

R. Hristovski et al.

and absence of long-range connectedness between scientific disciplines arises as a consequence of concepts that mainly overlap between neighboring disciplines. The attractor basin structure resembles a canyon-like configuration organized in an arc in which the path leads through small barrier saddles connecting neighboring scientific disciplines. Explanatory communication between more distant disciplines, say CR and PP or CS, needs a large, sometimes maximal conceptual (i.e. language) restructuring, and thus represents pathways of exploration and communication with vanishing probability. This means that a learner would hardly detect a connection and possibility of communication between distant scientific disciplines. The structure depicted in Fig. 2.2 corresponds to the extant reductionism-emergentism explanatory channel. On the one hand, as reductionism purports (e.g. Weinberg 1995), the vocabulary for higher level properties and processes, e.g. social groups (SG), may, in principle, be reduced to the explanatory vocabulary of elementary fields and particles (EFP). In contrast, emergentism argumentation is constructionist, and argues for the non-deducibility of higher order explanatory vocabulary from more elementary ones (the path from ELP to SG). However, the presence of general themata, which is characteristic of science modeling papers, leads to compression of the original 10 science vocabulary vectors into three principal components. This reduction of dimensionality (from four to three) and information compression was a consequence of language similarities between neighboring disciplines, but was also due to sharing common explanatory concepts between widely separated science practices (Fig. 2.3). The primary PCs were strongly correlated and resulted in one secondary PC, which was highly saturated by all scientific disciplines almost equally (for further information, see Hristovski et al. 2014). However, the presence of the general themata lowered the barriers and formed a more coherent linguistic domain. This enabled direct linguistic communication, even between fields that have traditionally been deemed distant and unconnected, such as PP, MB or CS and CB, on the one hand, and EFP and CL on the other. In contrast to the giant component structure in Fig. 2.2, the structure depicted in Fig. 2.3 corresponds to a fully connected network. This is a third and alternative way of connecting the different science vocabularies with respect to reductionism and emergentism. First, contrary to reductionism, connecting the sciences through general explanatory themata does not ‘destroy’ the higher level phenomena, properties and associated vocabulary of these sciences. Second, contrary to emergentism, the lower level vocabularies of EFP, MP and CL, for example, are not ‘forgotten’ because each level and discipline is left to be partly autonomous. Third, the general explanatory themata that are valid for EFL and for SG, and the whole spectrum within, are kept intact. The integration between scientific practices is underpinned by these general themata. Hence, what occurs is a social cognitive system that contains two coexistent properties: integration and autonomy of its components. This property is a hallmark of metastable adaptive systems, in which integration and segregation of information brings about the coexistence of stability and flexibility of its functioning (Tognoli and Kelso 2014).

2 Science as a Social Self-organizing Extended Cognitive System …

27

2.5 Scientific Practices as a Complex Adaptive System According to (Haken 2000), information compression within a system is a hallmark of self-organization and of the existence of increased coherence. Context-free themata lower the barriers between scientific explanatory patterns, which eases communication. Generally speaking, separate scientific fields maintain their context-dependent language (inter-scientific conceptual distances do not drop to zero—see Fig. 2.3). In contrast, context-free themata form a general embedding explanatory attractor basin, within which stabilizing coherent explanatory patterns become a feasible perspective. The linguistic communication process, then, may be defined as hopping, metastable dynamics within the general basin of attraction. It becomes obvious that the communication dynamics would be different in the two cases depicted in Figs. 2.2 and 2.3. Whereas in Fig. 2.2 the hopping dynamics are severely constrained within the narrow path connecting neighboring scientific fields, the basin of attraction depicted in Fig. 2.3 enables much more versatile communication among fields, and hence greater exploratory and communication breadth. Within this model, scientific research may be envisioned as a self-organizing process embedded in the action-perception-modeling space. Cooperative, synergic processes between scientific fields have led to the development of a new emergent explanatory pattern in the last few decades, which enables a novel synthetic world view. Context-free explanatory themata play the role of correspondence, and form a stable link among the models of organized matter at different levels. Synthetic thinking, viewed epistemologically, becomes an emergent property of science, based on the enhanced coherence revealed by the information compression of science vocabularies. The property is emergent since the whole of scientific understanding becomes different from the specialized knowledge within each of the fragmented scientific disciplines. However, it still contains specific, context-dependent explanations within each scientific discipline as special cases. Science reveals itself as a complex learning system. Empirical scientific concepts are basically context-dependent, where the context is formed by the length-time scales of the natural system under study. Length-time scale context determines the type of embodiment of the action-perception scienceenvironment cycle (that is, which technology embodiments are used to implement the experimental perceiving-acting cycle). However, recent developments clearly progress towards the explanation of concepts through a more general, and thus unifying, theoretical framework. This theoretical framework may be called context-free, but also pluri-contextual, depending on whether we use a bottom-up or top-down approach to its definition. If we use a bottom-up approach, the more general and hence more abstract explanatory terms become truly context-free. Conversely, if we apply a top-down approach, then, since the general explanatory concepts are valid for all levels, they become pluri-contextual, pleiotropic i.e. functionalfor each of the specific levels of organized substance. Therefore, the emergent explanatory pattern is characterized by both coherence of the explanatory skeleton, and flexibility through its context dependence.

28

R. Hristovski et al.

In this way, an interesting multi-scale structure of scientific explanatory patterns emerges. On the largest scale are the pluri-contextual principles of the widest scope (all length-time scales), and at the shortest scale are context-sensitive concepts of the smallest scope. Each of these levels has a characteristic spatiotemporal scale (the scope of functionality). The most general themata i.e. stabilizing concepts, constrain the context-sensitive concepts of smaller scope to function in a certain way within the explanatory patterns. The smaller scope, context-sensitive concepts give the more general concepts a specific context (context endowing concepts). The global scope principles stabilize the lower scope concepts by creating a global attractor in which more local concepts function. In this way, a coherent, yet flexible, adaptive system is being formed. Context-sensitive concepts respond to small changes in context (i.e. length-time scales) and provide flexibility, while general concepts provide the coherent backbone. The above-mentioned process of generalization is similar to the renormalization procedures used in statistical mechanics. In statistical mechanics, the renormalization procedure tells us which system parameters remain valid for different description levels of the system under study. One renormalizes, or transforms, the system’s model from a microscopic to macroscopic description, and finds out that some parameter of interest, for example temperature, remains relevant for all these levels. Other parameters of the microscopic description fade out as the scale of description gets larger. Hence, an accurate macroscopic description would be impossible without taking into account temperature as a relevant parameter, but other microscopic variables may be neglected. Let us use this analogy for explanatory vocabularies instead of physical variables. General themata survive the renormalization procedure, i.e. the change of levels of organized matter, and remain relevant and structurally stable at each level separately. They survive the coarse-graining transformation procedure of explaining the levels of description from micro- to macroscopic, from elementary fields and particles to social systems. This may be envisioned as a special case of language evolution (see Massip-Bonet 2013), i.e. as scientific evolution under selective pressure, which eliminates context-dependent concepts and stabilizes context-free ones as the fittest. Seen epistemologically, the role of selective pressure here is played by the process of change in the level of observation-description. From an ontological perspective, that role is played by the change in the level of substance organization, i.e. the length-time scale. Such general themata form a correspondence principle that penetrates across the levels and timescales of organized matter. In this sense, the interplay of such general themata at different levels of substance organization forms the coherent backbone of that very organization. On the other hand, the interplay of general themata and context-dependent concepts enables the context-sensitivity and adaptability of scientific language to all levels of substance organization. Having said this, the answers to the first and the third questions posed at the end of the introduction seem to be yes. The language coherence induced by general themata may enhance socially extended cognition among different science communities, so that positive learning transfer becomes a viable perspective. In the future, enhanced communication abilities between scientists from different fields could form a much more coherent, synthetic world view, which could serve basic educational purposes,

2 Science as a Social Self-organizing Extended Cognitive System …

29

among others. However, the degree of linguistic coherence in such explanatory patterns cannot be maximal, due to the existence of context-sensitive vocabulary that is specific to different scientific practices. Hence, despite the increased linguistic coherence, there is no threat of language homogenization. As a result, science communication practices can have two coexisting and complementary (Kelso and Engstrøm 2006) properties: coherence and flexibility. In summary, we attempted to explain science cognitive practices as a selforganizing adaptive system on at least two levels: the science-natural environment coupling cycle was defined as a unified embodied-extended cognitive system, in which both components causally constrain each other. Because of the multi-leveled phenomena in the natural environment, the technological embodiments, and consequently the cognitive extensions of science, lead to differences in explanatory practices that are mostly shared by the language in the form of models. However, the growing discovery of general themata that are functional for different scientific practices brings about a second level of unifying tendencies. A careful reader will also notice that the very model of self-organization of scientific practices that is presented here is, in fact, based on the same explanatory themata. This interplay between context-free unifying themata and context-sensitive explanatory practices provides stability, as well as flexibility, to the science-natural environment coupling that acts as a complex adaptive cognitive system. Acknowledgement This research was partly financed by the “University of Sts. Cyril and Methodius” program for research projects, No. 02-663/28 from 14.09.2012.

References Anderson, P. W. (1972). More is different. Science, 177, 393–396. Araújo, D., Davids, K., & Hristovski, R. (2006). The ecological dynamics of decision making in sport. Psychology of Sport and Exercise, 7, 653–676. Bruineberg, J., & Rietveld, E. (2014). Self-organization, free energy minimization, and optimal grip on a field of affordances. Frontiers in Human Neuroscience, 8, 1–14. https://doi.org/10. 3389/fnhum.2014.00599. Clark, A., & Chalmers, D. J. (1998). The extended mind. Analysis, 58, 7–19. Cvitanovi´c, P., Jensen, M. H., Kadanoff, L. P., & Procaccia, I. (1984). Renormalization, unstable manifolds, and the fractal structure of mode locking. Physical Review Letters, 55, 343–346. Englert, F., & Brout, R. (1964). Broken symmetry and the mass of gauge vector mesons. Physical Review Letters, 13(9), 321. Froese, T., Gershenson, C., & Rosenblueth, D. A. (2013). The dynamically extended mind. In Presented in IEEE Congress on Evolutionary Computation (CEC), Cancún, México. https://doi. org/10.1109/cec.2013.6557730. Fusaroli, R., Raczaszek-Leonardi, J., & Tylén, K. (2013). Dialog as interpersonal synergy. New Ideas in Psychology, 32, 147–157. Gibson, J. J. (1979). The ecological approach to visual perception. New York: Lawrence Erlbaum Associates. Glansdorff, P., & Prigogine, I. (1971). Thermodynamics theory of structure, stability and fluctuations. London: Wiley-Interscience.

30

R. Hristovski et al.

Haken, H. (1964). Theory of coherence of laser light. Physical Review Letters, 13, 329–331. Haken, H. (2000). Information and self-organization: A macroscopic approach to complex systems. New York: Springer Berlin Heidelberg. Higgs, P. (1964). Broken symmetries and the masses of gauge bosons. Physical Review Letters, 13, 508. Hopfield, J. J. (1982). Neural networks and physical systems with emergent collective computational abilities. Proceedings of the National Academy of Science, 79, 2554–2558. Hristovski, R. (2013). Synthetic thinking in (sports) science: The self-organization on the scientific language. Research in Physical Education Sports and Health, 2, 27–34. Hristovski, R., Balagué, N., & Vázquez, P. (2014). Experiential learning of the unifying principles of science through physical activities. In F. Miranda (Ed.), Systems theory: Perspectives, applications and developments (pp. 37–48). New York: Nova Science Publishers. Isnard, C. A., & Zeeman, E. C. (1974). Some models in the social sciences. In L. Collins (Ed.), The use of models in the social sciences (pp. 44–100). London: Tavistock. Kadanoff, L. P. (1966). Scaling laws for Ising models near Tc. Physics, 2, 263. Kelso, J. A. S., & Engstrøm, D. (2006). The complementary nature. Cambridge: MIT Press. Kelso, J. A. S., Southard, D., & Goodman, D. (1979). On the nature of human interlimb coordination. Science, 203, 1029–1031. Kirsh, D., & Maglio, P. (1994). On distinguishing epistemic from pragmatic action. Cognitive Science: A Multidisciplinary Journal, 18, 513–549. Landau, L. D. (1969). Collected papers. Moscow: Nauka. Laughlin, R. B., Pines, D., Schmalian, J., Stojkovi´c, B. P., & Wolynes, P. (2000). The middle way. Proceedings of the National Academy of Science, 97, 32–37. Mainwood, P. (2006). Is more different? Emergent properties in physics. Oxford: PhilSci Archive. Massip-Bonet, A. (2013). Language as a complex adaptive system: Towards an integrative linguistics. In Á. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 35–60). Berlin: Springer, Heidelberg. Pollack, G. H., & Chin, W. C. (2008). Phase transitions in cell biology. Springer. Raczaszek-Leonardi, J., & Kelso, J. A. S. (2008). Reconciling symbolic and dynamic aspects of language: Toward a dynamic psycholinguistics. New Ideas in Psychology, 26(2), 193–207. Raczaszek-Leonardi, J., Debska, A., & Sochanowicz, A. (2014). Pooling the ground: Understanding and coordination in collective sense making. Frontiers in Psychology, 5(1233), 1–13. https://doi. org/10.3389/fpsyg.2014.01233. Rovelli, C., & Smolin, L. (1995). Spin networks and quantum gravity. Physical Review D, 53, 5743–5759. https://doi.org/10.1103/PhysRevD.52.5743. Sutton, J., Harris, C. B., Keil, P. G., & Barnier, A. J. (2010). The psychology of memory, extended cognition, and socially distributed remembering. Phenomenology and the Cognitive Sciences, 9, 521–560. Tognoli, E., & Kelso, J. A. S. (2014). The metastable brain. Neuron Perspective, 81, 35–48. https:// doi.org/10.1016/j.neuron.2013.12.022. Turvey, M. (1992). Affordances and prospective control: An outline of the ontology. Ecological Psychology, 4, 173–187. Weinberg, S. (1995). Reductionism redux. The New York Review of Books, 15, 39–45. Wilson, K. G. (1975). The renormalization group: Critical phenomena and the Kondo problem. Reviews of Modern Physics, 47, 773. Wilson, R. A., & Clark, A. (2009). How to situate cognition: Letting nature take its course. In M. Aydede & P. Robbins (Eds.), The Cambridge handbook of situated cognition (pp. 55–77). London: Cambridge University Press. Wolynes, P., Onuchic, J. N., & Thirumalai, D. (1995). Navigating the folding routes. Science, 267, 1619–1620.

Chapter 3

The Paradigm of Complexity in Sociology: Epistemological and Methodological Implications Alvaro Malaina

Abstract This article seeks to present a unified frame of what we might call a “paradigm of complexity” from the definition of Thomas S. Kuhn, i.e., as a paradigm that incorporates both a worldview and models of scientific realizations. This dual nature of the paradigm of complexity is expressed by Edgar Morin with the distinction drawn between a more epistemological “general complexity” (complex thinking, second order cybernetics, autopoiesis, dissipative structures, etc.) and a more methodological “restricted complexity” (complex adaptive systems, multi-agents systems, cellular automata, etc.). We pose the respective limitations of both approaches and the need for their integration into a common paradigm of complexity that incorporates inseparably philosophy and science. In the second part of our article we study the implications of the paradigm of complexity in sociology. We propose, as examples, a number of approaches to the social, both epistemological and methodological, from the perspective of the “general complexity” and from the perspective of the “restricted complexity”.

3.1 Towards a Paradigm of Complexity. Bringing Together the Two Complexities According to Morin (2007), we can distinguish between the two following approaches to the phenomenon of complexity: “general complexity” and “restricted complexity.” On the one hand, a “general complexity,” is a fundamentally epistemological approach developed by scientists and philosophers such as Edgar Morin, Ilya Prigogine, Heinz von Foerster, Humberto Maturana, Francisco Varela, among others. It was developed primarily between the 70s and 80s from new disciplines such as cybernetics, systems theory, dissipative structures theory, catastrophe theory or A. Malaina (B) Department of Sociology: Methodology and Theory, Complutense University of Madrid, Madrid, Spain e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_3

31

32

A. Malaina

autopoiesis theory. Morin’s complex thinking would be one of its best syntheses. “General complexity” implies “an epistemological rethinking, that is to say, bearing on the organization of knowledge itself” (Morin 2007: 32). On the other hand, a “restricted complexity,” is a primarily methodological approach developed by scientists such as Murray Gell-Mann, John Holland, Stephen Wolfram, Stuart Kauffman, and Robert Axelrod since the creation in 1984 of the Santa Fe Institute in the United States and the improvement and sophistication of computational technologies. Complex Adaptive Systems Science is currently its dominant trend. The difference with “general complexity” is that, within “restricted complexity”, “complexity is never questioned nor thought epistemologically” (Morin 2007: 32).

3.1.1 The “General Complexity” “General complexity” approaches the phenomenon of complexity from a natural language. It draws its epistemological implications from the point of view of the subject who knows: complexity would compose a “new paradigm” (Morin 1977) or “new alliance” (Prigogine and Stengers 1979), which is potentially transdisciplinar. Therefore it gives a theoretical account of the properties of self-organization and autonomy of the physical, biological, and social systems from the perspective of the process of their observation. Complexity would express the extent of ignorance of an observer who is unaware of the information of the observed system itself (Atlan 1979) and the process of “construction” (von Foerster 1981) of an external object that is unattainable by the cognitive system of a subject. It is characterized more by their own “operational closure” and “internal consistency” (Varela 1979) than by the faithful representation of the external reality. This approach, going back to the historic Macy Conferences (1946–1953) on Cybernetics (Dupuy 1994), was widely developed in the 70s since the transition from a “first-order cybernetics” or cybernetics of observed systems (Wiener 1948) to a “second-order cybernetics” or “cybernetics of observing systems” (von Foerster 1981). Complexity, at the same time, expresses the self-organized and systemic nature of the world and the cognitive limits of human observers and it would call into question the deterministic, reductionist, and positivist principles of classical science. This approach has even included ethical proposals through authors such as Edgar Morin or Fritjof Capra. For these authors, the “paradigm shift” would not only have epistemological implications and imply the change of view of science or reality, but it would have ethical implications that tend toward a more harmonious relationship with nature, other people, and cultures from a “holistic” or “systemic” capability. Such a capability, integrating several complementary elements, makes up the biosphere and the entire humanity into a harmonious whole.

3 The Paradigm of Complexity in Sociology …

33

3.1.2 The “Restricted Complexity” “Restricted complexity,” instead, approaches the phenomenon of complexity from a formal language by trying to model using new computational techniques. It models are based on, such as cellular automata or multi-agent simulation (Miller and Page 2007), the “objective” and “observable” global structures and functions emerging in unpredictable dynamic “complex adaptive systems” (Gell-Mann 1994; Holland 1995) of the physical, biological and social, by means of the local interaction of its components. The restricted complexity algorithms are radically different from the classical science algorithms that are based on linear systems and differential equations. The methodology of simulation employed in restricted complexity shows the limits on cognition and prediction capabilities of the modeler. The structures and functions of the systems overcome the human computing capabilities and emerge as “surprising” (Wolfram 2002) or “counterintuitive” (Axelrod 1997) computational outcomes. A widely extended interdisciplinary project of research on modeling and simulation of complex systems has been launched since the foundation of the Santa Fe Institute in 1984, the first center exclusively devoted to the study of complex systems (Waldrop 1992). Unlike general complexity, restricted complexity is encouraged by the classical scientific spirit, and it does not cross the scientific borders. It looks closer at the hidden regularities of complexity and the refinement, as much as possible, of complex system modeling.

3.1.3 Towards a Unification of the Two Complexities There is a deep separation between these two ways of approaching the phenomenon of complexity. Each has its own references, schools, publications, and conferences. Each is presented as exclusive and often neglects or ignores the other. Many, who work in one, do not know the other, and vice versa. This scenario of dichotomy and polarization seems to contradict the very meaning of complexity as the linking of opposing and complementary principles, and demands a reunification of both perspectives in the context of a comprehensive paradigm of complexity including a “worldview” and “models of scientific realizations” (Bastardas-Boada 2014; Malaina 2012), the two dimensions of a paradigm according to Kuhn (1962).

34

A. Malaina

3.2 The Paradigm of Complexity in Sociology. Epistemological and Methodological Implications 3.2.1 The “General Complexity” in Sociology In the next section we consider how to apply the paradigm of complexity in sociology. We are going to approach the social both from the epistemological perspective of the “general complexity” as from the methodological perspective of the “restricted complexity.” We will first present an onto-epistemological model of the complex social system combining theoretical approaches of “general complexity” by Edgar Morin, Niklas Luhmann, Jesús Ibáñez and Anthony Wilden. To describe the human social systems from the point of view of the “general complexity” we must refer first to two fundamental concepts that ensure the cohesion of the organizational units. These are the notion of subject (central in Morin) and the notion of sense (central in Luhmann). We can not conceive of a subjectivity without sense or sense without subjectivity. We can not conceive of a society without the quality of a subject (collective) embedded in sense. The anthropo-social complex system being considered is a subject that makes sense that makes it. The sense is a possibility and a selection that constitute the boundary between the system and the world, which serves to “reduce complexity” through complexity and observe the world through the lenses of the concept of system. The subject is not only the effect of sense, but it is both cause and effect of sense. We move from a structural perspective, where the subject is the effect of sense (as in structuralism), to a dialectical perspective, where the subject is both cause and effect of sense. The proto-subject observed and distinguished indicates, and through its selective operation emerges itself as subject. This subject is not just an isolated individual observer, but it is also the anthropo-social observing system as a whole, made by relationships and interactions, a whole that is more than the sum of its parts, parts that can only be parts through the whole (through the sense that the whole gives them). But the boundary of sense may be exceeded, the sense is based as much in the system as in the ecosystem (it is based in the social system and in the psychic system, but also in the the ecosystem that is also social), although in the form of nonsense (for the system, for the observer in the system). The eco-system is a field of potential possibilities of sense and meaning. We can identify the sense of the system with the in-formation of the system (an extract from the sense) and the nonsense of the eco-system with the noise of the ecosystem (an extract of nonsense). But noise can generate new in-formation in the system and thus the nonsense can mutate into sense. Noise can be transformed into information: desorder can cooperate in the formation of a new order, as order and disorder always relate to a context (which is always relative to an observer). Moving further in the “general complexity” approach to the social, we find that the anthropo-social system is a communicative system (in the classical sense of information transmission). The objectives of the system (system of information) must

3 The Paradigm of Complexity in Sociology …

35

be communicated, spread by the network of interactions that gives form to the system. “Communication is a directed activity goals (goalseeking),” Wilden (1972a: 124) said. But in addition, communication creates the system. If there is a communication between the system and the eco-system where the “difference creates a difference”, in words of Gregory Bateson (1972), there is also a communication between parts (or subsystems) of the system. It is precisely the communication between parts of the system that creates the system. The anthropo-social systems are composed by elements that are subjects. The structural perspective had eliminated the subject, but the complex dialectical perspective places the subject at the heart of reality. It is the subject that changes the system by changing the sense. To do so, it activates its innate faculty of self -reflexivity. The relationships and interactions between subjects make up the system and let the system change. Society is a collective subject that changes and evolves from the change and development of the interrelations between the individual subjects that compose it. This change is mediated by the power of self-reflexivity and, in general, by the “evolutionary conquest” that owns the anthropo-social, consisting on the faculty to compute computations, that is to say, the property of consciousness. The subject is defined by presenting self-referential loops. Thus the various physical, biological and anthropo-social systems are “subjects”. They each feature for self-originated and self-referential activity centered on the self. What differs is the degree of awareness of the self. Strictly speaking subject is the system that computes its own self-referential activity (the living system) and the system that computes its own computation and becomes self-aware (the anthropo-social system) but the physical systems are already proto-subjects when, as the atom or the star, they present a self-centered activity. According to a complex thinking, the individual-society relationship would be hologramic, recursive and dialogic (Morin 2005): hologramic (the individual is in society that is in the individual), recursive (there is no determinism in this relationship, but a mutual production loop of individuals-society where interactions between individuals produce society, whose emerging qualities return as feedback on people and integrate them in the whole), and dialogic (the relationship is complementary—there is no society without individuals and there are no individuals without society—antagonist—the society represses individual impulses that transgress the norms and prohibitions of society—ambivalent—any society is both “community”—expression of solidarity between individuals—and “rivalrous”—expression of conflicts between individuals-). According to Anthony Wilden (1972a), in a set, there are three levels: the level of elements, the level of the structure (relations between elements) and the level of the system (relations between relations or relations between structures). The subject is therefore embedded in structures that are embedded in systems. The self-reflexive subject is entangled in a fixed network of relationships that links him to other subjects, producing a sense that finally defines the subject who embodies it. Self-reflexivity is only the ability to exercise freedom subject to constraints, a “semiotic freedom” (Wilden 1972a), a “freedom within an order” (Ibáñez 1994), an “autonomy/dependence” (Morin 1977). But this self-reflexivity is still able to

36

A. Malaina

open a gap in the structure, which can promote the leap to a meta-system, always unpredictable. A revolution or structural change may be started by subjects, but the end of the revolutionary process is always unpredictable. As in the formulations of deterministic chaos, the new stability that follows the fluctuation is undecidable. The social system is structured by linguistic codes. The linguistic code (which combines semantics and pragmatics) is, ultimately, the structure of the system. There are many kinds of linguistic codes that structure societies. Ibáñez (1979: 193) has studied such an evolution of three main Texts in humankind, three linguistic codes that have structured the social system during history: the religious Text (typical of the Old Regime), the legal Text (typical of Modernity) and the publicity Text (typical of Postmodernism). In general, the texts are dictated by the ruling classes of the system in order to keep their domination over the entire system. It is a strategy by which the ruling classes become the eco-system and the dominated classes the system. It responds to the model of an organizationally open (programmed by the dominated classes) and informationally closed (no reflexive, which makes the dominated classes unable to produce information) system (Ibáñez 1994: 97). But the system always returns to what it is by nature: organizationally closed (self-organized) and informationally open (reflexive, information producer): the dominated classes can subvert and reorganize the system (Ibáñez 1994: 98). The linguistic codes, such as the genetic codes, are reproduced by the individuals, but they can also evolve, mutate or be revolutionized by the subjects themselves. Through this way the social system is capable of restructuring itself by changing its linguistic code. The possibility of change happens as we know from the transformation of noise (ecosystemic) into information (systemic). The structure is not fixed; it can always be changed, specifically from the top of the systemic level that encompasses the previous levels, from the level of the system itself, the level of relations between relations, the level of change of structures. Communication, as we said, is the fundamental quality of the anthropo-social system (Luhmann 1995). All communication is carried out by a structure defining the range of communicative possibilities. However, communication may also go beyond the structure, becoming meta-communication, becoming noise (relative to the structure, relative to the “coded” information). This is the event. This metacommunicative noise can leave a trace in the system, trace that is remembered and would be the basis of restructuring and morphogenesis of the system. The ultimate source of morphogenetic restructuring will always be the subject, as the central device for self-reflexivity of the anthropo-social system. But the new form, originated by the subject, “self-transcends itself” in the complex and uncertain process of emergence. Morphogenesis takes the form of a negative feedback of the second order, which usually brings to the meta-system the following properties: an enhanced adaptive scope, increased sustainability, variety (complexity) increased, structural innovations, a new order of organization, the generation of modified subsystems, a broader selectivity, changes in the order of adaptability and learning, a symbol storage increased, more varied possibilities for simulation of increased opportunities to change goals, an increase of the system sensitivity to “noise” (Wilden 1972b: 62).

3 The Paradigm of Complexity in Sociology …

37

Social morphogenesis could be broken down into two successive moments: a moment of crisis (blockage of the flow of energy/information at the molecular level) and an event time (emergence at the molar level of a new regulation of flows of energy/information) (Morin 1972). Both are complex moments in the system, since they precede the formation of a new structure (unpredictable). Formally, the event itself, which epitomizes morphogenesis, would be the revolution as a “global morphogenetic event” (Malaina 2012: 99), which affects the structure of the system as a whole, projecting the referring system to a meta-system; but it may also be that there are not revolutionary events on a global level, but on the local one (e.g., the end of slavery, universal suffrage, the conquest of civil rights by women or ethnic minorities, etc.). The crisis would be the “pre-event”, the stage immediately preceding (and somehow already mixed with it) the triggering event. In the luhmannian conceptual framework, the crisis would express a rupture of the sense within the system, whereas an event would be a selective reconstruction of sense. At the base of social morphogenesis (crisis/event) there is noise, but the noise which complexifies social organizations is of a very special kind, since it incorporates an intentionality and self-reflexivity, unlike the noise that affects biological organisms. Social morphogenesis through noise appears as randomness. It is randomness for the observer located within the self-reproductive path of the system, because it is an anticipated sense. The shadow of a becoming new sense, not possible to be seen in the current code. But, seen from outside the system, and in a time subsequent to its impact, it is considered as a historic event in the dynamics of the system, a historical event with meaning. If morphogenesis seems randomness, it is also because it is a creation. The event is creation (creation of a new sense) and the creation always appears as randomness for the observer. The observer is finally the last variable in the model of general social complexity. Through it we introduce the epistemological problems developed by complex thinking and second-order cybernetics. The human cognitive system is an autopoietic system and thus epistemology must be constituted as an observation of the second order, that is to say, as an observation of the observer, derived from his capacity of self-reflexivity, which allows this ongoing journey to move to a meta-level of observation, which allows external observation of the autopoietic circle. The social system thus should be exposed to second-order cybernetics involving an observation of the observation and in turn to a third-order cybernetics involving an observation of the observation of the observation and so on… As the loop of observations has no end, it is the modeler who must operate the final cut operating a distinction of the social system by the trace of a border and a selection of one of the two sides of the system to set itself apart, distinghised from its ecosystem. We have seen therefore how society can be understood from the “general complexity.” However, we found that the description is of a purely philosophical order, finally appealing to an epistemology of the observer. Is it not also necessary to measure and quantify to speak of a complex perspective not only of social theory but of social science in its broadest sense? We need then models of social complexity, which we will study in the next section.

38

A. Malaina

3.2.2 The “Restricted Complexity” in Sociology We have conceptualized society from the more epistemological perspective of “general complexity.” But we can also approach methodologically the social from the paradigm of complexity. Artificial Societies are an area of research in the science of complex systems. Axelrod, Gilbert, Epstein and Axtell are some of the leading names within this field. The Artificial Societies are simulation models of social systems made from multi-agent systems. They are written in computing language (such as Java, NetLogo, Visual Basic, C++ or Pascal) and represent the emerging global dynamics of a social system that arise from the set of interactions between agents. They have been already used to study many anthropo-social phenomena, such as cooperation, trade, group formation, conflict, interaction of human society with the natural environment, the spread of epidemics, the transmission of culture, migration, urban growth and population dynamics. The agents at the base of Artificial Societies may be entities of various kinds: they may be individuals, but also families, ethnic groups, businesses, cities or entire countries. The environment in which they work can be economic, social, political or cultural. It is the modeler who establishes the ultimate characteristics of the agents and the environment of the Artificial Society. One of the simplest models of Artificial Societies is the model Sugarscape, introduced by Epstein and Axtell (1996). Agents act in a grid of 50 × 50 cells. Each cell has a renewable amount of sugar that can be consumed by the agent that reaches it. However, the amount of sugar varies according to geographical location and the amount of sugar that has already been consumed. The agents must eat sugar to survive. If the agents sugar level drops to zero they die and are replaced by agents with random initial endowment of sugar. If agents get more sugar than they need for their survival, they can accumulate it and consume later, or even talk to other agents, and thus establish communication and commerce between them. Agents can look north, south, east and west of their current position. Depending on the “gene pool” of each agent, one can see a longer or shorter distance. The rule is simple: the agent continues to move toward the nearest cell with the largest amount of sugar and consumes it (the rule M). Agents also have differentiated “metabolic rates”. A rule of “survival of the fittest” follows from the system: agents who die faster are the agents less adapted to their environment, those with short sight, a high metabolic rate or living in geographical areas where sugar is not available in large quantities. Agents who have had success also die once they have reached a life limit assigned randomly. Epstein and Axtell applied the model Sugarscape to modern societies, by substituting the element of sugar by the element of wealth. The conclusion they get is alarming: even if the initial distribution of sugar or wealth is relatively symmetrical, after successive iterations, we find that the final distribution of sugar or wealth remains highly unequal: the best fitted agents will hold the largest amount of wealth, while the rest dies or hardly manages to survive. It would be a phenomenon of self-organization and emergence of a collective stable structure from strictly local, simple rules and actions. Epstein and Axtell have developed many experiments with this “laboratory” of social complexity, as they call it, adding extra dimensions to

3 The Paradigm of Complexity in Sociology …

39

their original model: sexual reproduction, culture, conflict, market or trade. They always seek to identify the emergence of collective complexity from the local simplicity, how global structures arise from local interaction between the agents that make up the system. The Sugarscape model, besides the of unequal distribution of wealth, discovers emerging waves of migration, the formation of cultural groups or, by introducing spices along with sugar, of stable market prices. Sugarscape has also give rise to many social simulations following his example (e.g. the Anasazi model, which simulates the behavior of the Anasazi culture in Long House Valley of Arizona, during the years 400–1450 AD). Modelling seeks the “skeleton” of the organizational principles of natural systems. Modelling Artificial Societies also seeks to strip the social logic to the end of its skeleton, until its fundamental matrix, the “micro-social mechanisms” (Axelrod 1997) that generate social complexity. The goal is to mathematize the substrate/fundamental mechanism of social complexity that would explain the micro/macro transition. In his book, The Complexity of Cooperation, Robert Axelrod introduces complexity theories in search of a social skeleton, integrating the model of cellular automaton, multiple bands, the noise in the Prisoner’s Dilemma, the emergence of new political actors in a pluralistic world and the emergence of culture. The attempt is always to look for the fundamental social substratum that is expressed by an inherent tendency to cooperation applicable to multiple contexts. “The complexity of modeling based on agents must be in the simulated results, not in the presupposed model.” (Axelrod 1997: 5) “An extremely simple model captures an essential feature of many interactions.” (Axelrod 1997: 6) We can feel the presence of the complexity theory by Murray Gell-Mann and of the model of cellular automata by Stephen Wolfram in Axelrod’s background. The multi-agent models studied by Axelrod seek to study the emergence of cooperation among social systems. The environment is competitive and agents are adaptive (not rational). The logic of the system is non-linear, nonapprehended by the classical mathematical analysis: hence the need for simulation. An example of cooperation that produces social complexity is the model of dissemination of culture by Axelrod. We will present this model as an example among the many models of complex adaptive social systems developed by Axelrod. He sought the model of social influence that may explain how people become more similar to each other through interaction, so that they increase the probability for cooperation between them. He finally was seeking to model the emergence of a culture shared by members of a given group. The model is very simple. Culture would be described by a series of cultural dimensions (such as religion, the way people dress, music, cooking, etc.). Each dimension of culture is expressed by a set of traits, alternative values of each dimension of culture (if one takes the dimension of religion, it would be Christianity, Buddhism, atheism, Islam, etc.). The dimensions and characteristics must be quantified and realized in the model (which implies to make cuts). Axelrod states that such a culture would have 5 dimensions and 10 possible features for each dimension. A culture may then be described by a set of numbers, such as 27,891 or 75,482. In the first case, the culture for the first dimension has the feature 2, whereas in the second case has the feature 7, and so on. Two agents have the same culture if they have the same 5 traits for

40

A. Malaina

each dimension. The degree of cultural similarity between two agents is expressed as the percentage of the dimensions that have the same trait. The model distributes the agents geographically in a grid with variable dimensions. Axelrod takes such a set of 100 spaces in a grid of ten to ten dimensions. Its initial model is static (subsequent models can develop the possibility of migration of agents along the grid), so in each place there is an agent and each agent interacts with his immediate neighbors, located in North, South, East and West (the “von Neumann neighborhood”). The agents located in the sides of the grid have only three neighbors and those located in the four corners have only two (one could also in other models transform space into a torus). The simple rule of the model is that the more the agents are culturally similar, the more their probability of interaction increases. Each event of iteration of the model works as follows: the program randomly selects an agent that is activated and one of his neighbors. With probability equal to their degree of cultural similarity these two agents interact. The program randomly selects one dimension where agents have different traits, and then changes the feature of the original agent by the feature of his neighbor. The cultural similarity of the two further increases, so that in the future the possibilities of convergence between the two increase. The model of culture of Axelrod, iterated thousands of times, shows the formation of cultural regions in a space battle between cultural homogeneity (convergence) and cultural heterogeneity (polarization). Axelrod discovered that these two trends continue, and that on one hand there is a general trend towards the abolition of cultural borders, but also a certain tendency to conservation of cultural differences. The model stabilizes after 81,000 events, with an ample overall homogeneous sector and some locally differentiated cultural areas. Simulations of this Artificial Society show that the number of local homogeneous cultural regions decreases with the number of dimensions of culture (more dimensions, more likely to share features, more likely to interact and be influenced), increases with the number of alternative features for each dimension (more features, less likely to share with neighbors and therefore interact), decreases with the radius of interaction (more neighbors, more likely to share traits, more chances to interact and be influenced), and decreases when the size of the space increases beyond a certain threshold (which is a counter-intuitive result, unexpected, evidence of the emergence of complex structures unpredictable at the starting point: the small territories have little space for stable homogeneous cultures, the middle size territories have sufficient space, while large territories also, but the settling time and thus the interaction is so long that eventually there is a tendency to dissolve cultural boundaries). Complexity, shows Axelrod, is ultimately a “mechanism”, the mechanism of the universe to generate variety and evolution so that a quantitative conglomerate of elements are able to produce a qualitative novelty to adapt to their environment. The cellular automaton artificially reproduces this mechanism. So there is homeomorphism. The universe is a cellular automaton and the cellular automaton is a universe, paraphrasing Edward Fredkin. But we must ask the question, to capture the complexity, do we simply need to go back to the organizational basic substrate leaving aside its concrete manifestations, do we not lose at the same time the complexity itself? The adaptive/behaviorist micro mechanisms are the only source of

3 The Paradigm of Complexity in Sociology …

41

macro social complexity? Can we avoid considering the pre-structuring of the system and subjectivity itself that does not appear in the models and that has been well studied for example by Foucault or Bourdieu? Do not also run the risk of falling into some neoliberal ideology (Malaina 2014) if we forget taking into account the morphogenetic and revolutionary capacity of the subjects that make up the social system? And also, do we not need to introduce in the model the consciousness of the observing subject, the only operator that gives “sense” to the dynamics of the cellular automaton/multi-agent system/Artificial Society? We have seen how the dimensions of the subject and its capacity to transform the structure of the system become fundamental. There is however an important difference between these reflexive approaches and the behaviorist approach of complex adaptive science and Artificial Societies. The solution that we propose would be to bring together both approaches into a common “complex sociology” (Malaina 2012).

3.3 Conclusion: The Paradigm of Complexity in Sociology In this article we have tried to present some snapshots of the implications of the paradigm of complexity in sociology, highlighting its two fundamental aspects that Edgar Morin calls “general complexity ‘and’ restricted complexity.” We have seen on the one hand how from the “general complexity” perspective, society can be seen as a complex system consisting of sense and composed by selfreflexive subjects. The social system is structured by a code that is reproduced by subjects in a context of inequality and domination. But the subjects in their interactive dynamics are also able to change the code and restructure the system, making noise mutate in new information in order to transform the system. We have finally seen how all this description of the system should take into account the observer as last variable of the model. We have seen on the other hand how society from the “restricted complexity” perspective can also be modeled using multi-agents simulations. These models are called Artificial Societies, and seek to study the fundamental mechanisms in the simplest models to explain the observable social complexity. The model Sugarscape by Axtell and Epstein is one of the pioneering models, as well as the work of Robert Axelrod. In sum, we find that both the ‘general’ and ‘restricted’ approaches have limitations and should be combined in a unified paradigm of complexity. They can serve if they operate in a convergent way to study the social from a very novel transdisciplinary perspective, which is rooted in the physical and natural sciences and can provide much light and rigor to the sociological study.

42

A. Malaina

References Atlan, H. (1979). Entre le cristal et la fumée. Paris: Seuil. Axelrod, R. (1997). The complexity of cooperation. Princeton: Princeton University Press. Bastardas-Boada, A. (2014). Toward “Complexics” as a transdiscipline. In M. A. Martí & M. Taulé (Eds.), Homenatge a Sebastià Serrano (pp. 63–77). Barcelona: University of Barcelona Press. Bateson, G. (1972). Steps to an Ecology of Mind. Chicago: University of Chicago Press. Dupuy, J. P. (1994). Aux origines des sciences cognitives. Paris: La Découverte. Epstein, J. M., & Axtell, R. L. (1996). Growing artificial societies: Social science from the bottom up. Cambridge: MIT Press. Gell-Mann, M. (1994). The Quark and the Jaguar. Adventures in the simple and the complex. New York: W. H. Freeman and Co. Holland, J. (1995). Hidden order. How adaptation builds complexity. London: Helix Books. Ibáñez, J. (1979). Más allá de la sociología. El grupo de discusión: técnica y crítica. Madrid: Siglo XXI. Ibáñez, J. (1994). El regreso del sujeto. La investigación social de segundo orden. Madrid: Siglo XXI. Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press. Luhmann, N. (1995). Social systems. Stanford: Stanford University Press. Malaina, A. (2012). Le paradigme de la complexité et la sociologie. Paris: L’Harmattan. Malaina, A. (2014). Complex Adaptive systems and global capitalism. The new ideology of global complexity. World Futures, 70(08), 469–485. Miller, J., & Page, S. (2007). Complex adaptive systems. Princeton: Princeton University Press. Morin, E. (1972). L’événement-sphinx. Communications, 18, 173–192. Morin, E. (1977). La Méthode I: La nature de la nature. Paris: Seuil. Morin, E. (2005). Introduction à la pensée complexe. Paris: Seuil. Morin, E. (2007). Complexité restreinte, complexité générale. In J. L. Le Moigne (Ed.), Intelligence de la complexité. Epistémologie et pragmatique (pp. 28–50). Paris: Editions de l’Aube. Prigogine, I., & Stengers, I. (1979). La nouvelle alliance. Métamorphose de la science. Paris: Gallimard. Varela, F. (1979). Principles of Biological Autonomy. Amsterdam: North Holland. von Foerster, H. (1981). On constructing a reality. In H. von Foerster (Ed.), Observing systems (pp. 234–235). Seaside, CA: Intersytems Publications. Waldrop, M. M. (1992). Complexity: The emerging science at the edge of order and chaos. New York: Simon and Schuster. Wiener, N. (1948). Cybernetics. Paris: Hermann. Wilden, A. (1972a). System and structure. Essays on communication and exchange. London: Tavistock. Wilden, A. (1972b). L’écriture et le bruit dans la morphogenèse du système ouvert. Communications, 18, 48–71. Wolfram, S. (2002). A new kind of science. Wolfram, cop.

Chapter 4

How and Why to Model the Complexity of Thought Systems Leonardo G. Rodríguez Zoya

Abstract The aim of this paper is to sketch out an epistemological model for empirical research on paradigms, i.e., the principles organizing a thought system. This model is a theoretical and methodological device for examining questions such as how scientists think the epistemic, social, political and ethical dimensions of their own scientific practices; how a paradigm, or way of thinking, determines a mode of problematization, that is, the horizon of possible research questions, conceivable ideas, acceptable problems, imaginable concepts, and how a science or a scientific discipline is able to observe, problematize and criticize their own paradigms. To address these questions, the model proposes a qualitative and quantitative study of scientific beliefs. A belief is a form of social cognition produced through social practices and discourses. The analysis of language and social communication among scientists provides a way of understanding scientific beliefs. Moreover, beliefs are not discrete and isolated entities. On the contrary, they implicate each other forming networks or assemblages of heterogeneous beliefs. A system of scientific beliefs is a complex structure constituted by different types of clusters of beliefs called attitudes. Such a system is a kind of social representation, that is, a mental model socially produced and shared by members of a group which structures the way they perceive, question, understand and objectivize reality. Modeling a system of scientific beliefs is a kind of metacognitive strategy, a way of producing a reflexive, self-critical meta-point of view for scientific research.

L. G. Rodríguez Zoya (B) Investigador del Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET)—Argentina, Comunidad de Pensamiento Complejo, Instituto de Investigaciones Gino Germani, Universidad de Buenos Aires, Matheu 1225, C1249AAA Ciudad Autónoma de Buenos Aires, Argentina e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_4

43

44

L. G. Rodríguez Zoya

4.1 Introduction Contemporary societies are confronted with a myriad of complex problems: environmental problems (climate change and biosphere deterioration), economic problems (the regulation of financial markets), social problems (poverty, hunger, inequality) and political problems (terrorism, violence, human rights violations), among others. Transforming complex problems into more manageable situations requires (a) the production of more appropriate knowledge for understanding the complex problems at hand (epistemic challenge); (b) democratic and strategic actions aimed at improving the current state of a complex problem and collectively shaping a better future (ethical-political challenge), and (c) a reconsideration of the fundamental problems we face, the knowledge we produce and the decisions we make as part of larger intervention strategies intended to modify reality (reflexive and meta-cognitive challenge). This last challenge poses an invisible problem can be hard to solve, precisely because it involves a process of objectification and self-examination of our own system of thought, or paradigm. Our mode of thinking plays a crucial role in the questioning and conceptualization of complex problems, as well as in the strategies and decisions we make in order to deal with them. For this reason, the problem of thought is one of the greatest political, educational and scientific challenges facing humanity in the 21st century. Imaging and constructing an alternative world-system, that is, an ethically, ecologically and socially sustainable form of living, is dependent on our capacity of learning to think in a different way thus transforming our paradigms. Could science help us in our quest to think the complexity of our most pressing problems? Possibly. However, in order to do so, science should first of all be able to think itself, that is, problematize its own structure of thought. How can a particular science investigate the paradigms organizing its own knowledge construction practices? To tackle this problem, the present work outlines a strategy for modeling the complexity of thought systems. Building a model of paradigms or organizing principles for a system of scientific thought (or any other form of human organization) constitutes a meta-point of view through which scientists are able to self-examine, problematize and criticize their knowledge constructions strategies. Constructing this meta-point of view requires a metacognitive strategy allowing the reflective inclusion of the subject in the production of knowledge. In doing so, we are developing a second order of knowledge and a science with consciousness of the complexity of its own practice and way of thinking (Morin 1984). This paper is organized as follows. First, we problematize the complexity of thought as an object of study and as a method of knowledge. Second, we propose and develop the category of “complex problems”, demonstrating the importance of the reflective dimension of thought in empirical studies. Finally, we outline an epistemological model for empirical research on complex problems, which is briefly illustrated through a practical application to the field of complexity science and social simulation.

4 How and Why to Model the Complexity of Thought Systems

45

4.2 Thinking as a Complex System In his seminal work, The Architecture of Complexity, Simon (1973) argues that a system can be described from two different points of view: as a state or as a process. The former concerns a structural analysis aimed at understanding the organization of a system at a particular moment of its development, whereas the latter entails a dynamic analysis of the system itself, that is, its evolution over time. Complex systems (physical, biological, socio-anthropological) are historical systems to the extent that their organization is the emergent result of an evolutionary process. For this reason, Jean Piaget (1979: 117) argues that there is a “fundamental necessity of a dialectic between genesis and structures”, so that “there is no structure without history nor history without structure” (García 2006: 81). Even more, it can be stated that “an structure is organized history” (Rodríguez Zoya 2013: 84). Methodologically, the evolutionary dynamics of a system entails: (i) the structuring phases corresponding to a generative or genetic process through which a new organization is constituted; (ii) structured phases involving mechanisms that intervene in keeping systems organized; and (iii) distructuring and restructuring phases, that encompass the breakup of a given structure and the emergence of a new one. The key question in studying thinking as a complex system is: “How is thought constructed, how is it organized and how does it change overtime?” The problem concerning the organization and change within thought systems has been tackled from different perspectives and with such varying concepts as psychogenesis and sociogenesis (Piaget and García 2008), paradigm (Morin 1998), archeology and genealogy (Foucault 1998), thought structure or thought framework (Koyré 1999), social representations (Moscovici 1979), mindscape (Maruyama 1980), mode of thinking (Whitehead 1944), among others. Here we are neither able to provide a deep discussion of these perspectives nor to formulate an answer to the question at hand. Rather, we would like suggest that the problem of constitution, organization and change within thinking may be considered a central issue of an interdisciplinary research program on complexity of thought systems. The following rationale offers some pointers for thinking the complexity of thought.

4.2.1 The Recursive Loop Between Thought and Reality There is an adage stating that “he or she who only knows a hammer sees nails everywhere”.1 This visual metaphor contains a heuristic value to problematize the link between reality and thought. Our instruments of knowledge (i.e. our thinking styles, theories, concepts, scientific models and research methods) determine our way of perceiving and organizing reality. 1I

am thankful to Professor Gustavo Montengro, an organizational psychologist from Universidad Nacional de Córdoba (Argentina), for suggesting this metaphor.

46

L. G. Rodríguez Zoya

Let us develop a thought experiment. Someone shows you a picture of a jungle. Let us call this object the observed system. Let us now bring in many different observers. Then, someone asks: what do you see? An individual says: “a jungle, threes, plants”. A student answers: “chlorophyll and photosynthesis”. A poet pronounces: “the colors of life before men”. An agronomist states: “there are, at least, five plants species”. A physicist and chemist asserts: “a thermodynamic process of energy dissipation”. An ecologist declares: “an ecosystem, a fragment of the biosphere and GAIA”. All of them were looking at the same object but they did not observe the same phenomenon (Hanson 1958). Different observing systems grant different points of view over the same empirical domain. Thinking is a point of view structuring a way of perceiving and organizing the experience of reality. But, at the same time, the phenomena of the world of experience determine the very structuring of thinking. Thought and reality are not tow discrete and separate entities, nor is it possible to establish an ontological priority between them. Both ontological constructivism (Woolgar 1991; Latour 1999) and ontological realism (Bunge 2009; Putnam 1994) entail forms of simplification. On the other hand, according to empirical findings of social psychology (Moscovici 1979; Duveen and Lloyd 1990), genetic epistemology (Piaget and García 2008; García 2000) and biology of knowledge (Maturana and Varela 1972, 2003), among others scientific disciplines, the organization of thought and reality amount to a correlative and mutually constitutive process. Complexity lies in the recursive process through which thinking constitutes reality and, thereafter, is constituted by it. Reality is the name given to a historical structuring of that recursive process. Therefore, the reorganization of a thought system is concomitant with a deep transformation of reality’s organization. For instance, Aristotelian cosmology is radically different from modern cosmology emerging with Tycho Brahe, Galileo, Kepler, among others, whose synthesis is accomplished by Newton. The reorganization of the Aristotelian-Thomistic’s thought system into a modern way of thinking has entailed what Alexandre Koyré called “the destruction of the cosmos” (i.e., the substitution of a finite, closed and hierarchical conception of the world as a wellorder whole for an indeterminate or even infinite universe) and “the geometrization of the space” (the replacement of the Aristotelian space with Euclidean geometry) (Koyré 1999). The constitution of the modern thought system involved a fundamental process as the result of which man lost his place in the world, or, more correctly perhaps, lost the very world in which he was living and about which he was thinking, and had to transform and replace not only his fundamental concepts and attributes, but even the very framework of his thought (Koyré 1999: 6).

4.2.2 Thinking as a Model-Building Process The poet and philosopher Paul Valéry states: “We reason only on models”. Thinking can be conceived as a process through which we construct models to observe, question and interpret the world, our relations with others and with ourselves. The concept of a mental model is a plausible hypothesis of cognitive science, suggesting that

4 How and Why to Model the Complexity of Thought Systems

47

individuals construct cognitive representations of the concrete situations where they act (Johnson-Laird 1983, 1987). According to the logician Jean-Blaise Grize, “all action, all behavior and, particularly, all discourse lies in a mental model of a specific reality” (Grize 1993: 3). As a consequence, a mental model represents the point of view of an observer X about an experience Z. Following the observation of Valéry, systemic thinker Jean-Louis Le Moigne (1990: 15) raised the question of “how we construct the models through which we reason”. It is assumed that models are historical, social and cognitive structures built through language and social communication. Models are historical because they are not immutable entities, but they change over time; they are social both because they are irreducible to mental individual activity, and because they are made by social matter emanating from language and culture; finally, they are cognitive constructions because they amount to a form of knowledge. Mental models are individual mental constructions in the sense that they constitute a personal elaboration that is not shared by other people (van Dijk 1999). However, this assertion does not imply that mental models are asocial cognitive productions. On the contrary, building a mental model necessarily entails the use of internal language as the symbolic material inherent to thinking (Voloshinov 1976: 25–26). Cultural-historical psychology suggests that thinking is internalized language (Vygotski 1995). This approach allows us to go beyond the principle of the “excluded middle” in classical logic, which leads to the disjunction between mental and social, cognitive and cultural, individual and collective. By contrast, culture and society are an “included middle” constituting the individual thought. Even if mental models are constructed by internalized language, they do not amount yet to communicative structures, because discourse and socio-verbal interactions are not presupposed by them. In this regard, Grize (2012) has coined the concept of schematization to refer the discursive formulation of a mental model. The communicative dimension of language is precisely what allows us to understand the constitution of social representations, that is, mental models socially produced and shared through discourse, communication and social practices (van Dijk 1999; Moscovici 1979). Following these arguments, it can be asserted that a model is a symbolic structure built out of a language. There is no model without language. The idea of models as linguistic constructions is useful for thinking and distinguishing between models built of natural language and artificial language. Social representations are natural language-based models emerging from social, cognitive and communicative interactions in the life-world. They are also implicit or tacit models since they are not intentionally and deliberately produced by social actors. On the other hand, modeling as a scientific practice is the work of building explicit models, that is, a deliberate activity intended to conceive, design and construct models. According to Minksy (1965), a model can be defined as follows: “To an observer B, an object A* is a model of an object A to the extent that B can use A* to answer questions that interest him about A”. This valuable heuristic conceptualization allows for the distinction of modeler subject (B) from the object of modeling (A) and, thus, conceives the model as an instrument of knowledge (A*). Besides, it suggests that

48

L. G. Rodríguez Zoya

every model is constructed by virtue of a question and, consequently, entails the conception of a problem. This observation reveals the practical nature of models, since a model is always built by someone (the subject of modeling), and is always about (the object of modeling) and for something. Even further, we can question why, for what and for whom a model is constructed. This consideration suggests that models are interest-laden constructions and, for this reason, models are not neutral (Rodríguez Zoya 2013). Scientific models can be formulated both in natural and formal languages. The first are discursive models built of a theoretical and conceptual language belonging to a particular discipline. The adjective ‘discursive’ is used in order to highlight the fact that they are no formal models.2 Qualitative models produced by social sciences and humanities are an example of discursive models. The second are formal models based on artificial language, for instance, mathematical models (usually called equations based models), statistical models (variable centered models) and computer simulation models (models written in a programming language). A formal model is always an abstraction and a simplification of the conceptual or mental model on which it is founded. To conclude, both social and scientific thinking entail model-building processes, differing from each other in terms of the language with which they are built, their goals and levels of formalization.

4.2.3 Thinking as a Process of Problematization Problems are not given in our immediate reality. A problem is not a datum of reality but is something real that can be empirically investigated (García 2006). What we usually call “problems” is actually the result of a process of problematization. Thinking can be characterized as a practice or activity through which human beings reconstruct fragments of our experience as problems (Foucault 1999). To think is to practice the art of problematization, that is, the art of conceiving new questions and problems (Morin 1986). The concept of thought system refers to mental and social, 2 Natural thought and discursive logic (Grize 1996, 2012), often used to characterized common sense

thinking, are also essential in science, even if scientific thinking can be described as formal thought based on logical-mathematical operations (Piaget 1978). Let us point out that: (i) no thought, no conception, can ignore ordinary language” (Morin 1986: 203); (ii) scientific practice needs dialogue, communication and understanding between scientists, hence, natural language; (iii) theories and scientific conceptualizations use specialized language derived from natural language. The phenomenologist Alfred Schütz states that knowledge produced by social science amounts to second-order constructions, i.e., models of typification of our common sense. Logical positivist Otto Neurath coined the term ballunguen to point out that scientific language is intrinsically inexact: a combination of precise terms stemming from scientific work and others more vague and ambiguous deriving from ordinary language (Neurath 1983; Gómez 2008). All in all, scientific thought is not equivalent to formal thought (logical-mathematical reasoning), nor can it be reduced to it. On the contrary, scientific thought expresses the complex articulation between natural and formal languages.

4 How and Why to Model the Complexity of Thought Systems

49

cultural and cognitive processes through which individuals, groups and human societies problematize their experience. A problem, whatever its type (social, political, economic, scientific), is an experience construed as an object of thought through a process of problematization. Therefore, different thought systems entail different forms of problematization. If the foregoing rationale is correct, then, it can be argued that our thought system determines both the type of problems we construct and their proposed solutions. Thus, a problematization delimits not only the domain and scope of what is conceived as a problem, but also the range of possible answers. Therefore, transforming a problem is not only associated with the search for new solutions but, essentially, with a critical analysis of paradigms (Morin 1998), that is, the principles organizing the thought system which grounds what is experienced and conceptualized as a problem. Let us illustrate this idea with a brief example taken from the history of science. Physics form Aristotle to Galileo asked the same question: What is motion? According to Aristotle, the natural state of bodies is stasis, so that, in the absence of an external force, there is a concomitant lack of motion. The inertia principle, formulated by Galileo and expressed in the First Law of Motion by Newton,3 was unimaginable and inconceivable by the western thought system for more than two thousand years. Furthermore, before Galileo, the idea of a permanent motion not caused by the constant action of an external force was systematically rejected as absurd (Piaget and García 2008: 232). A Chinese text dating from about five centuries BC states that “the cessation of a motion is due to an opposite force. If there is not an opposite force the motion will not stop […] This is so evident as a caw is not a horse” (Piaget and García 2008: 232). More than two thousand years before Galileo and the principle of inertia, the Chinese civilization had envisioned a conception of motion which was very similar to the moderns’; however, they were not capable of developing either the mechanics or the physics as Galileo or Newton. What happened? In the Chinese thought system, becoming and change were the natural states of the world, thus, they did not require an explanation. In the Aristotelian-Thomistic’s thought system, world is static, consequently, the principle of inertia was not conceivable. Both thought systems block, albeit by different means, a mode of objectification—developed, in fact, by moderns—in which inertial motion became the object of a new science: classical physics. The scientific revolution systematized by Newtons’ Principia has mainly consisted in the invention of a new question: “How motion is produced?” The answer is not formulated any longer in terms of substances and ends, but relations between forces affecting bodies independently of time. A paradigmatic change of a thought system is not produced either by finding new answers to established questions or by establishing a cumulative body of knowledge, but due to the invention of “new questions allowing formulating problems in a different manner” (Piaget and García 2008: 228). Having said that, changing a 3 The

First Law states: “An object at rest stays at rest and an object in motion stays in motion with the same speed and in the same direction unless acted upon by an unbalanced force”.

50

L. G. Rodríguez Zoya

question can be a difficult thing to do, since it implies “changing the bases of the starting point of a reasoning or a theory” (Morin 1998: 238), that is, transforming the paradigm of a thought system. A paradigm determines a form of problematization, that is to say, the scope of thinkable questions, acceptable problems, conceivable ideas and imaginable concepts. For this reason, theoretical and empirical research on complex problems—as we shall see further ahead—is inseparable from the problematizing paradigms that structure our thought system and the way we make sense of experience.

4.2.4 Complexity of Thinking and Complex Thought Thought is a complex phenomenon and a distinctive feature of human complexity. Indeed, human culture in all its forms is mediated by thinking: art, poetry, science, philosophy, economics, and politics and so on. Thinking is an emergent phenomenon generated by biological and sociocultural complexity, irreducible to either dimension. From an evolutionary point of view, hominization involves a process of brain development whereby it becomes bigger and more complex. The increasing sociocultural complexity of human societies was made possible by, though not solely as a result of, the biological complexity of the brain. In fact, the emergence of mind, consciousness and intelligence is inconceivable without the development of a double articulation of language, which is, in turn, impossible to conceive without an increase of cerebral complexity. Thinking, brain, language and culture are connected in a recursive process (Morin 1973, 2001). A distinction can be made between thinking as an object of knowledge and as a method of knowledge. On the one hand, a process of objectification is conducted in order to construct thinking as an object of inquiry, reflection or experimentation. In doing so, thinking can be investigated on a theoretical or empirical level, scientifically or philosophically, with the purpose of producing new knowledge. The complexity of thinking as object of knowledge rests upon the fact that thought is formed by interrelated processes stemming from different disciplinary domains. Thinking is simultaneously a neurological, biological, chemical, linguistic, psychological, social, cultural and historical phenomenon. Each discipline must provide a necessary input, but their isolated contributions are insufficient to elucidate the complexity of thinking. For this reason, there is an increasing demand to conceive interdisciplinary research programs on thought systems as complex systems. On the other hand, the idea of a method of thought can be conceived as a “dialogic art of conception” (Morin 1986). Why an art? The idea of art refers both to praxis (action) and poiesis (production). Genetic epistemology has empirically demonstrated that thinking begins with the action of a subject upon an object. The development of thought begins through the coordination of schemas of action and then through conceptualizations. For that reason, Piaget argues that thinking is an internalized action. Moreover, the notion of art intertwines the artist’s creativity in the conception of new ideas with the skill of the artisan in the production of new

4 How and Why to Model the Complexity of Thought Systems

51

objects. Creativity and thinking skills are learnt through habit, experience and practice within a specific social and cultural context. Hence, a method of thought is a way of practicing a kind of rational thinking that encompasses creativity, praxis, learning and art. Why a dialogic art? The concept of dialogic elaborated by Morin refers to the interplay of two complementary and antagonistic logics, ideas or concepts (Morin 1986). The praxis of thought entails a dialogic between the pluralities of cognitive operations (e.g., analyze—synthetize, distinguish—relate, induce—deduce, connect what is separated—separate what is connected). The notion of dialogic refers to the idea of dialogue, and it can be associated with the word ‘dialectics’, whose etymology comes from the Greek verb dialégomai: ‘the art of the dialogue and argumentative dispute’. Dialogue is a discursive process implying a link between a person with an interlocutor. Although thought always entails an individual cognitive dimension, the development of thinking involves a reflective mediation with the other, as it is suggested by Ricoeur (1996) when he affirms that “the other constitutes oneself”. Thus, the praxis of a method of thought requires both learning how to think for oneself and how to think with the other. What’s more, the constitution of thinking cannot be reduced to a dyadic relation between subject and object, but rather, constitutes a triadic type of relation: subject-object-other. The category of ‘other’ invokes a variety of instances: context, culture, society, language, others subjects. The other is an “included middle” constituting both subject and object of knowledge. Effectively, thinking does not problematize or assimilate isolated objects but socially and culturally interpreted objects. Finally, why an art of conception? The notion of conception refers to the creation of novel qualitative results. It can be argued that thinking is conceived of new ideas, new concepts, new questions, and new objects. The notion of conception weaves together the ideas of creativity, imagination and innovation: to think is to creatively imagine something new, that is, to conceive what has not yet come into to existence. For this reason, we affirm that the practice of thought is our most valuable methodological resource. The main function of thinking is neither contemplation nor reflection about the external world but, transformation (Piaget 1973). Thinking is a constructive practice of both cognitive and social character through which we create, organize and transform relations between objects, actions and concepts. A method of thought is a practice of thought which allows for the construction of new knowledge. A practice or method of thought entails two different levels of analysis: firstorder thinking and second-order thinking. On the one hand, first-order thinking is an object-oriented thought. In this sense, the practice of thinking rests upon a process of objectification through which an experience is constituted as an object of thought. All science, all art, as well as all action and decision entails a first-order thinking through which we problematize, organize and interpret experience. On the other hand, second-order thinking implies the development of a thought about thought, a reflective movement through which thinking becomes an object for itself. This level of analysis has received different names: reflexivity, meta-cognition, second-order knowledge, complex epistemology or knowledge of knowledge.

52

L. G. Rodríguez Zoya

The formation of a complex thought, as suggested by Morin (1977, 1990), entails a kind of method or practice stemming from rational thinking. Complex thought involves a creative and rigorous articulation of both a method of objectification (first-order thinking) and a reflective method (second-order thinking). A complex practice of thought requires thought being able to think about itself in order to selfproblematize, self-discover, self-critique. Consequently, the development of a complex thought is a strategy for building reflective knowledge. The goal of thought about thought is to elucidate, problematize and critique the mental structures with which we perceive, question, understand and organize the world of experience.

4.3 Objectivity and Reflexivity in the Interdisciplinary Study of Complex Problems Over the last three decades, theoretical production, empirical research and philosophical reflections on complex systems, complexity science and complex thought have been vigorously developed. (Prigogine and Nicolis 1997; Gell-Mann 1995; Morin 1977). However, a conceptual lacuna can be discerned in the category of complex problems which has received limited attention. In fact, efforts to further develop the concept of complex problems in an epistemologically rigorous and methodologically operative manner, with the purpose of conducting collective researches on concrete problems of contemporary societies, have been scant (Rodríguez Zoya 2012). The goal of this section is to outline the category of complex problems. It is stated that empirical research on complex problems implies the articulation both of objectification methods (i.e. modeling and simulation of complex systems) and reflective methods (i.e. knowledge strategy of complex thought). Both methods allow us to identify two dimensions of the concept of complexity: objectified complexity and reflective complexity. First, the objectification of complexity is essential to explain the organization, functioning and dynamics of a complex problem, as well as to imagine and make visible strategic alternatives for its transformation. Secondly, reflective complexity is required to build a meta-point of view which allows for the objectification of the paradigm organizing our thought system. This meta-point of view enables us to make sense of how we think the complexity of a problem, its mode of objectification and strategic alternatives. In short, there is a recursive loop between objectification and reflexivity in the empirical and critical research on complex problems. Why a problem is called ‘complex’? What makes a problem ‘complex’? There is a long debate concerning the definition of ‘complexity’ (Edmonds 1999). However, defining a term is not always a good theoretical and epistemological starting point when conducting a research. In fact, in science not all concepts are defined explicitly. For instance, mathematics defines ‘natural number’, ‘rational number’, etc. but it does not define the term “number” (García 2000). As a consequence, instead of formulating a definition of complexity, we propose to conceptualize four dimensions in order to

4 How and Why to Model the Complexity of Thought Systems

53

demarcate, more broadly, the notion of complex problems. We do not claim that this is the only possible way to conceptualize complex problems; on the contrary, what we claim is that this conceptualization is relevant for thinking the complexity of fundamental problems in contemporary societies.

4.3.1 The Ethico-political Dimension of Complex Problems As previously stated, a problem is not given in immediate reality, but it is an emergent construction stemming from a process of problematization. This process implies that an aspect of experience (the ‘observed system’) is perceived, evaluated and conceived as problematic for someone—an individual, a group, a community, a society—(the ‘observing system’), by virtue of certain values, beliefs and knowledge (Foerster 1996; Longino 1990; Morin 1977). Therefore, the objectification an experience as a problem presupposes value judgments. A process of problematization is an evaluative process through which an observing system judges and evaluates a given experience as problematic. Throughout this process, an observing system identifies, selects, and abstracts certain aspects, relations, data and information of empirical reality in order to reconstruct experience as a problem. The label ‘complex problem’, associated with ‘empirical phenomena’, is an ‘observed system’ emerging from the constructive process carried out by an ‘observing system’. Why does an experience become problematic? This is due to many different reasons. In effect, various factors of diverse character are intertwined in a process of problematization: political, economic, social, cultural and scientific factors, as well as knowledge opinions, ethical judgments, values, beliefs, among others. The key point is that social problems are complex because they imply what Putnam (2004) called an entanglement of fact and value judgments. For instance, “soil salinity” entails no only a fact judgment about objective reality but also an ethicopolitical standpoint about the implications of salinity for the environment and food production within a given agro-system. Factual components (descriptive-explanatory judgments) and axiological components (ethical-political judgments) cannot be separated in the study of a complex problem. Complex problems are experiences of social life which are judged as problematic because of they affect life itself (life of human beings and natural life). A complex problem is an ethical, undesirable or unbearable situation. Empirical and theoretical studies of complex problems call for an axiological reflexivity in order to make explicit the ethical and political aspects intervening in the conception, observation and objectification process of a complex problem. The problematization of the axiological dimension can be summarized in the following question: why and for whom investigate a complex problem?

54

L. G. Rodríguez Zoya

4.3.2 The Practical Dimension of Complex Problems There are two principles that must be articulated in any research on complex problems: ‘investigate to know’ and ‘investigate to act and transform’. Indeed, a complex problem needs to be transformed because it is judged ecologically, ethically, socially and even humanly as undesirable by virtue of certain values. Consequently, a complex problem connects two imperatives: an epistemic and a practical imperative. The first one calls for explanation and understanding of complex problems, and presupposes the need for an integrated and rigorous diagnosis about the history of the system (its structuring process) as well as its present (its current structure and functioning). The second one entails the challenge of acting upon the system in order to transform it and make it evolve into a more desirable status (in social, ecological or ethical terms) in the future. As a result, a complex problem involves three times dimensions: past, present and future. This epistemic and practical take on complex problems can be conceptualized, following Varsavsky (1975), as a constructive approach, since its strategic aim is to transform a complex problem into a more desirable situation. The crucial question is: More desirable for whom—who decides and defines what is desirable? In order to tackle this issue, it is convenient to distinguish the three-fold logic of this constructive approach: the desirable, the possible and the probable. The logic of ‘the desirable’ implies the construction of models about future, that is, the representation of a prospect about the future. How do we want the situation experienced as problematic to actually be in the future? Building a model of the future involves making ‘purposes’ and ‘values’ clear. Whereas a purpose defines a goal or objective to be achieved, values are necessary to choose and decide between different purposes that can be evaluated as more or less desirables. A model of the future is an ethico-political concept, thus a particular kind of normative model. The key question is: Whose are values taken into account in the design and construction of a model for the future? This question enables us to distinguish between top-down and bottom-up models. The first are hierarchically built by managers, policy makers, planners, technicians; the second are emergent models built through a participatory process led by multiple actors involved in the problematic situation. The logic of ‘the possible’ implies two types of models: diagnosis and possibilistic models. The goal of the first is to explain the historical evolution and the present structure of a system. This kind of model revolves around questions such as: “How does a system have to evolve in order to be what it is?” and “how is a system organized, how does it work, how does it actually behave?” On the other hand, possibilistic models tackle questions such as: “What may or may not happen to a system taking into account its present organization?” “How could a system evolve in the future?” Possibilistic models allow us to explore ‘what if’ hypotheses and, therefore, are useful to make visible different scenarios and evaluate their feasibility. Possibilistic models are valuable tools for the strategic planning and management of complex problems due to the fact that they are help to compare, evaluate and decide

4 How and Why to Model the Complexity of Thought Systems

55

between possible alternatives by virtue of the models of the future that have been constructed. The logic of ‘the probable’ is inherent to statistical models, usually generated with a predictive goal and based on the calculus of probability and confidence interval. All in all, the constructive approach of complex problems presupposes a succession of models: first, a normative moment directed at building a model of the future; then, an explanatory moment encompassing diagnosis models; next, a strategic moment orientated to the production of possibilistic models; and finally, a tacticaloperative moment aimed at conducting a practical intervention on the system in order to transform it.

4.3.3 The Epistemic Dimension of Complex Problems A complex problem can be empirically investigated through a modeling process. There is a difference that needs to be established between the notion of complex problems and the notion of complex systems. A complex problem is a domain of reality that is conceived as an observable system through a process of problematization. Thus, a complex problem is an object of modeling, whereas a complex system is the model of a complex problem. For any given complex problem, different models (i.e. complex systems) can be constructed. Complex problems are phenomena compounded by: – Many interrelated processes involving different disciplines. For instance, a social agro-environmental system amounts to a phenomenon where physical-biological, ecological, technical, productive, socio-economic and juridical factors intervene. – Nonlinear interactions between many social, political, economic actors. – Multiple special scales (local, regional and global) and temporal scales (short, mid and long term). – Multiple levels of organization: nonlinear dynamics between the micro, meso and macro levels of complex systems. – Multiple consequences: economic, ecological, political, ethical and so on. According to Warren Weaver’s (1948) pioneering work, complex problems are problems of organized complexity “which involve dealing simultaneously with a sizable number of factors which are interrelated into an organic whole” (Weaver 1948: 3). Further, Weaver distinguished what he has called problems of simplicity and problems of disorganized complexity. The first refer to phenomena where fewer variables are involved and can be analytically examined by mechanical models (e.g. classical physics), whereas the second one involve a large number of variables or elements that can be analyzed through statistical models (e.g. thermodynamics). Problems of organized complexity cannot be studied either by mechanical or statistical models, but through systemic models. Thus, a complex problem can be characterized by epistemic attributes usually identified to describe the structure and

56

L. G. Rodríguez Zoya

dynamics of complex systems: self-organization, emergence, non-linearity, sensibility to initials conditions, adaptive behavior, among others. A complex system is an ensemble of interrelated elements whose dynamic interaction over time produces an emergent behavior at the macro level. These macro properties cannot be linearly deduced from the analytical study of parts at the micro level. For this reason, complex problems can be studied by modeling and simulating methods of complexity science.4

4.3.4 The Methodological Dimension of Complex Problems The systemic character of complex problems has important methodological consequences. Complex problems are closely related to non-decomposable systems, according to Simon (1973). The elements and processes forming these kinds of systems are mutually dependent and, consequently, cannot be separated in order to be independently studied. Moreover, the components of a complex problem stem from the theoretical domains of different disciplines. From a methodological point of view, the complexity of a problem is deeply connected with the impossibility of achieving a systemic understanding from one particular discipline or a mere juxtaposition of disciplinary perspectives. This approach to complex problems has methodological consequences for the type of research needed in order to study them. Following the work of Rolando García (2006), complex problems require an interdisciplinary research methodology. This methodology does not entail either suppressing or fusing disciplines. Specialized, disciplinary knowledge is necessary but insufficient for the study and diagnosis of complex problems. This observation permits to establish an essential difference with respect to ‘transdisciplinary’ research (Nicolescu et al. 1994; Vilar 1997), that is, aiming for discipline integration (Wallerstein 1996) by cultivating the idea of ‘systemic generalist’ (Bertalanffy 1976). According to García, the basic feature of interdisciplinary methodology is the integration of disciplinary approaches at the beginning of a research project, in order to define a common problem and establish a shared conceptual framework. Interdisciplinary is a process that articulates phases of integration and differentiation between disciplines. This consideration enables us to establish another essential difference with respect to what is usually called ‘multidisciplinary’ and where there is a coordination or juxtaposition of results stemming from disciplinary-based research. To conclude, complex problems demand interdisciplinary research strategies.

4 Complexity

science uses different types of formalisms to model the structure and simulate the dynamics of complex systems, such as: cellular automata, complex networks, agent-based models, and system dynamics, among others.

4 How and Why to Model the Complexity of Thought Systems

57

4.3.5 A Provisional Synthesis of the Concept of Complex Problems A complex problem is a meta-system including multiple points of view onto different observing systems founded on a common experience to be known and transformed, since it is evaluated as undesirable. The observing systems are formed by specialists from different disciplines, by social actors involved in the problematic and by policy makers and stakeholders. The problematic experience constitutes an observed system that emerges through an objectification process produced by the articulation of multiple viewpoints. That observed system can be conceptualized as a complex system and studied with an interdisciplinary methodology. The expression ‘complex problem’ is the meta-system articulating both the observing system and the observed system. Whereas complexity science offers a wide range of techniques and tools for objectifying the complexity of an observed system, complex thought offers a reflective strategy for modeling a meta-viewpoint and allowing the observing system to observe its observation, that is, to objectify its mode of objectification, to think its way of thinking and to conceive its conception.

4.4 An Epistemological Model for Empirical Research on Thought Systems The concept of paradigm elaborated by Edgar Morin is a useful theoretical tool for constructing a meta-viewpoint capable of stimulating a self-reflective and selfcritical activity concerning a thought system. However, Morin’s work also displays a methodological limitation, since it does not offer an explicit strategy for empirical research on paradigms. As a result, a key question arises: “How can a paradigm be empirically investigated and how a meta-viewpoint modeled in order to help an observing system produce a self-observation and think about itself?” To advance a possible answer to this problem, an epistemological model of complex thought [EMCP] is outlined here. This model proposes qualitative and quantitative research schema on scientific beliefs as a strategy for modeling a paradigm, that is, the principles organizing a thought system. This model has been applied to an empirical research on beliefs system of complexity science and social simulation.

4.4.1 The System of Scientific Beliefs Following the idea of thinking as a modeling process, as it was previously conceptualized, scientific beliefs are social representations shared by members of a research community. Therefore, scientific beliefs are a form of social cognition, that is, social mental models. A system of scientific beliefs is a complex structure formed by het-

58

L. G. Rodríguez Zoya

erogeneous beliefs concerning different dimensions of scientific practice and their relations with society. The main dimensions of scientific beliefs are: – Ontological beliefs: representations on objects of knowledge and beliefs about the nature of reality. – Epistemic beliefs: conceptions about what is and what should be considered scientific knowledge, the idea of science accepted as valid. – Methodological beliefs: conceptions about methods and instruments of knowledge construction. – Logical-cognitive beliefs: conceptions about cognitive strategies for knowledge construction, representations on the very strategy of thought and the mode of reasoning. – Social beliefs: conceptions about the relationship between science and society; ideas about how science determines scientific work and its social consequences; normative beliefs about how the relationship between science and society should be; conceptions concerning the social responsibility of science and scientists. – Axiological beliefs: conceptions about the role of values (ethical, social, political) in the process of knowledge construction. – Anthropological beliefs: conceptions about the role of the subject in the process of knowledge construction. A system of scientific beliefs is an organized assemblage of ideas, beliefs, representations, values and attitudes of a scientific community. As socio-cognitive constructions, scientific beliefs are a form of knowledge which is socially produced and shared and enables the organization a functional viewpoints on the world. In this sense, scientific beliefs amount to shared schemes of action and communication enabling the practical development of scientific life. As a consequence, scientific beliefs play a crucial role in structuring practices, since they allow for coordinated actions and understanding about what there is, what is done, how it is done and with what purpose. Scientific beliefs constitute the nooshpere of scientific thinking. Beliefs intervene in scientific reasoning and form mental models with which scientists elaborate interpretative schemas for the practical contexts in which they act. In short, scientific beliefs, socially constructed and shared, are a socio-cognitive framework of scientific thinking.

4.4.2 Systems of Beliefs as Propositional Networks Cognitive psychology states that beliefs are placed in a section of the mind called semantic memory, that is, a long-term memory which stores permanent mental representations (Best 2002: 116–119; Bruning et al. 2005: 46–52).5 According to 5 Cognitive

psychology has distinguished three types of memory: sensorial memory (SM), shortterm memory (STM) and long-term memory (LTM) (Best 2002: 107–119). The latter was divided by

4 How and Why to Model the Complexity of Thought Systems

59

van Dijk (1999: 48), social beliefs are shared mental representations stored in the social memory. Following this contribution, it can be stated that scientific beliefs are located within social semantic memory. Therefore, scientific beliefs are mental, social and shared constructions. Cognitive science and psychology have coined different models to describe how the content of the semantic memory is represented and organized. Ross Quillian (1968) has proposed one of the first models claiming that information is organized in networks of propositions. A proposition is a link between two concepts or nodes based on a significant relation (Hernandez Forte 2005: 61). The proposition is the minimum unit of memory information processing (Gagné 1991: 77). Propositions “represent meaning in a semantic code, regardless of lexical, grammatical and syntactical aspects of natural language sentences” (Rodríguez 2010: 220). This is the reason why propositions are not equivalent to phrases or sentences. The difference rests on the fact that “words, syntagms and sentences represent forms of communicating ideas, whereas propositions represent the ideas in strict sense” (Gagné 1991: 82). From this approach, propositions are articulated in propositional networks that meaningfully represent the content stored in the semantic memory. Therefore, semantic memory has a reticulated structure (Cabeza 1987: 84–85). According to this perspective, scientific beliefs can be described by significant propositional networks located in the social memory of a group (van Dijk 1999: 39–40). Consequently, scientific beliefs are not discrete and isolated mental entities, but an organized assemblage of propositions.

4.4.3 Communication and the Sociogenesis of Scientific Beliefs Scientific beliefs are the emerging result of social cognitive practices. Beliefs are elaborated by way of a practical process (communicative, discursive and symbolic) in which the mental activity of individuals and the social dimension of the group are articulated. We coined the concept of sociogenesis of scientific beliefs to account for the socio-cognitive process through which beliefs are constructed and transformed. The crucial link between, on the one hand, socially-shared scientific beliefs and, on the other hand, individual mental models, must be conceived taking into account the communicational dimension of schematized processes. In effect, the concepts of mental model and schematization establish a connection between cognition and thinking at the individual level and the empirical situation where individuals communicate through discourse supported by natural language. Thus, dialogue gains

Tulving (1972) into two: episodic and semantic memory. Episodic memory stores concrete events or episodes of individual life., semantic memory represents “organized knowledge that a person possesses about words and other verbal symbols, their meaning and referents, about relations among them” (Tulving 1972: 386).

60

L. G. Rodríguez Zoya

relevance as a practical process in which the social and cognitive components of scientific beliefs are intertwined. Therefore, the categories of dialogue and socio-verbal interaction, proposed by Voloshinov (1976), are crucial epistemological concepts to problematize the sociogenesis of scientific beliefs. In concrete scientific practice, researchers establish social relations with other members of their group through interdiscursive practices in which the schematizations of mental models formed by scientific beliefs play an important role. Interdiscursivity is a dynamical and controversial process through which scientists may arrive at consensus (share beliefs) or disagreement (conflicting beliefs) about the practical situation where they act. Therefore, personal beliefs involved in the construction of mental models, “may be seen to be shared by others, and thus are generalized as social beliefs” (van Dijk 1999: 115). An explicit articulation may be drawn between, on the one hand, the generalization process of socially-shared beliefs suggested by van Dijk and, on the other hand, abstraction and generalization mechanisms proposed by genetic epistemology to account for the psychogenesis of formal structures of thought (García 1997: 47–53). Following Castorina (2007: 167–168), abstraction and generalization mechanisms developed by Piaget to describe operative logic can be extended to the analysis of natural logic developed by Grize. Taking into account this reasoning, a theoretical hypothesis on the sociogenesis of scientific beliefs can be proposed: scientific beliefs are constructed and transformed through the abstraction and generalization of mental models schematization emerging from interdiscursive practice between scientists. Therefore, interdiscursive practices constitute the arena of acquisition, construction and transformation of scientific beliefs. As for genetic epistemology, thought is the organization of interiorized actions (Piaget 1973: 89–92), whereas scientific beliefs are the organization of interiorized interdiscursive practices.

4.4.4 Organizational Complexity of Beliefs Systems: Paradigm, Epistemic Framework and Attitudes A paradigm is the pattern that connects heterogeneous assemblages of beliefs. It is an organizational concept used to characterize the type of relationship established between networks of heterogeneous beliefs. For this reason, a paradigm lacks an explicit definition and it should not be confused with a theory or body of knowledge. Furthermore, a paradigm includes “the essential relationship of exclusion/association between fundamental concepts, that is, preliminary alternatives and associations that control and guide the knowledge, the thinking and thus the action” (Morin 1977: 430). Paradigms are organizing principles of the thought that regulate “operations of union (conjunction, inclusion, implication) and operation of separation (differentiation, opposition, selection, exclusion)” between ideas, beliefs or concepts (Morin 1999: 26). Consequently, a paradigm is endogenous to thinking, theory, reasoning and discourse; it functions as a generative principle modulating the organization of the

4 How and Why to Model the Complexity of Thought Systems

61

thought and the reasoning. Thus, a paradigm does not have a visible existence (as a material thing or substantial entity) but a virtual one. Consequently, a paradigm only exists as it is generated and regenerated by practice, by thought and discourse. Therefore, a paradigm “rests on the phenomenological reality that it generates and it needs that reality to be regenerated” (Morin 1998: 236–237). An epistemic framework is an ensemble of beliefs that is organized by virtue of a given paradigm. Genetic epistemology has coined the concept of “epistemic framework” to refer to a tacit system of thought or worldview that modulates theories and conceptualizations across different disciplines without determining its content (Piaget and García 2008; García 2000). In this context, it is important to point out that a system of scientific beliefs in a specific spatial-temporal context may be composed of different epistemic frameworks. Even more, rival epistemic frameworks can coexist in a system of scientific beliefs, expressing different conceptions about reality, knowledge, science, methods and so on. For instance, two research groups from the same discipline may have opposite ontological and axiological beliefs and, nevertheless, share similar methodological beliefs. Therefore, there are complex modes of articulation and opposition between epistemic frameworks constituting a belief system. In short, paradigms refer to how beliefs are connected and form an epistemic framework. In order to empirically investigate paradigms organizing a system of scientific beliefs, a conceptual device is required to make an interface between two levels of analysis (scientific beliefs and epistemic framework). In other words, from a theoretical point of view, we need a concept to account for how beliefs are grouped together and related to one another inside an epistemic framework. The concept of attitude is a theoretical category used to describe clusters of scientific beliefs forming an epistemic framework. Attitudes are “specific, organized, clusters of socially shared beliefs” (van Dijk 1999: 65). More precisely, attitudes are “socially shared evaluative beliefs” (1999: 55). An evaluative belief is a proposition with which one may agree or not. For this reason, attitudes express degrees of agreement on a specific cluster of beliefs. As a result, a cluster of beliefs forming an attitude form around a subject; for instance, beliefs about the role of values in science (axiological beliefs), beliefs about the nature of reality (ontological beliefs), and so on. To summarize the discussion thus far: – A system of scientific beliefs is organized by one or more epistemic frameworks. – An epistemic framework is a compound of interrelated attitudes. Each attitude is an organized cluster of beliefs. – An attitude or cluster of beliefs is a propositional network belonging to the social memory of a group. – A paradigm is the organizational pattern connecting assemblages of beliefs that form an epistemic framework. – Paradigms are organizing principles between epistemic frameworks, systems of beliefs and thought system. – Empirical research on scientific beliefs constitutes a strategy for studying paradigms and the organization of systems of thinking.

62

L. G. Rodríguez Zoya

4.4.5 How to Empirically Investigate a Paradigm? Qualitative and quantitative methods can be used in order to produce empirical evidence on scientific beliefs and model the organization of a system of thought. Firstly, qualitative methods are essential for analyzing the discursive logic by which the natural thinking of scientists is expressed. Specifically, conversation techniques, such as in-depth interviews or focus groups, are a valuable methodological resource to produce empirical evidence on scientific discourse. Discursive practices are the milieu from which forms of reasoning, beliefs and attitudes emerge. Additionally, conversation grants the possibility to examine how scientists think their own way of thinking and thus infer meta-cognitive strategies. Secondly, psychometric techniques for attitude measurement, such as Likert scales (Likert 1932), are useful tools for producing quantitative evidence on scientific beliefs. Scales of attitude measurement are a cluster of items. Each item is an empirical indicator of the attitude that seeks to be measured. Items are written “as opinions with which one can agree or not agree. An opinion is a verbalized attitude, thus, through opinions we can infer the underlying attitude” (Morales Vallejo et al. 2003: 50). Epistemologically, a theoretical bridge can be established between the concept of scientific belief and techniques of attitude measurement. In effect, we have argued that beliefs are propositions, whereas attitudes are clusters of interrelated beliefs, that is, networks of propositions. Consequently, each item of a Likert scale may be considered as a proposition and, thus, an evaluative belief. The ensemble of items forming a scale can be theoretically interpreted as a socially shared cluster of evaluative beliefs. Hence, scaling techniques are a strategy for empirical research on scientific beliefs. Theoretically, it is important to subordinate quantitative attitude measurement to qualitative analysis of scientific discourse. The qualitative analysis is indeed crucial to infer and conceptualize beliefs and attitudes that can be quantitative measured. In other words, how many scales and what items are necessary in order to model the complexity of an epistemic framework is something that can only be determined a posteriori of the qualitative discourse analysis. Qualitative analysis is the methodological basis for quantitative modeling of scientific beliefs. Additionally, multivariate analysis techniques are required to model the organization of a system of scientific beliefs; in particular, we emphasized the importance of exploratory factorial analysis (AF) and multiple linear regression. On the one hand, AF is useful for exploring the underlying structure of the items that form a scale and, therefore, to transform multiple observable variables in a reduced number of latent variables. Methodologically, AF permits to construct a factorial index whereas, theoretically, it allows for conceptualizing constructs identified in the analysis. On the other hand, multiple linear regression may help us understand the relation between different attitudes and clusters of beliefs and, thus, model the principle organizing a system of scientific beliefs.

4 How and Why to Model the Complexity of Thought Systems

63

The complementary articulation between factorial analysis and multiple linear regression is a methodological strategy for understanding and explaining the magnitude and density of a web of scientific beliefs. In other words, multivariate techniques allow us to quantify the complex fabric of reticulated clusters of scientific beliefs. For this reason, statistical analysis helps to account for the organizational pattern that connects clusters of beliefs (attitudes) and forms an epistemic framework. Therefore, the quantitative component of the methodological strategy entails a statistical operationalization of both the construct of ‘epistemic framework’, proposed by Jean Piaget and Rolando Garcia, and the construct of ‘paradigm’, elaborated by Edgar Morin.

4.4.6 The System of Scientific Beliefs in Complexity Science Having summarized an epistemological model of complex thought [EMCT], the goal of this section is to briefly illustrate the empirical application of this model within a research on thought systems. This model has been tested in a qualitative and quantitative research project on the system of scientific beliefs of complexity science (Rodríguez Zoya 2013). The aim of this research was to understand paradigms and epistemic frameworks organizing practices of knowledge construction in complexity science. The qualitative portion was based on 53 in-depth interviews with researchers from complex systems and social simulations from seven countries in Latin America and Europe. Quantitative portion was based on the findings of qualitative analysis of scientific discourse. A battery of 18 Likert scales and a bank of 404 items were constructed in order to empirically measure the scientific beliefs of complexity science. Quantitative data was produced through a Global Survey of Complex Systems and Social Simulation applied to a sample of 232 researches from 28 countries. Modeling the system of scientific beliefs was based on qualitative discourse analysis, factorial analysis and multiple linear regression techniques. Statistical operationalization of the epistemological model of complex thought [EMCT] was conducted through the development of 59 statistical indexes of scientific beliefs and the construction of 22 regression models. In doing so, we were able to reconstruct the structure of clusters of beliefs forming different epistemic frameworks of complexity science and paradigms articulating its relationships. The full findings of this analysis have been reported elsewhere (Rodríguez Zoya 2013). The survey used ten scales with 111 items from the bank of 404 items due to questionnaire extension. Table 4.1 shows the reliability analysis for the ten Likert scales used to model the system of scientific beliefs of complexity science. Factorial analysis of Likert scales allows us to identify twenty two relevant constructs from a theoretical perspective. For each of these constructs another reliability analysis was conducted. Support on factorial analysis we have created scientific

64

L. G. Rodríguez Zoya

Table 4.1 Reliability analysis of Likert scales Nº Name of the scale Original itemsa 1 Cognitive strategies of complex 12 thought

Final itemsb 12

Inter-item Alfa correlationc coefficientd .197 .746

2

Simplified cognitive operations

12

12

.208

.759

3

Scientific objectives

12

8

.291

.767

4 5

Role of values Scientific responsibility

16 12

10 9

.346 .295

.841 .790

6

Conceptions about reality

12

6

.294

.714

7

Complexity and subjectivity

8

8

.353

.813

8

Ontology of complex systems

7

7

.402

.825

9

Scale on Conception of complexity

14

12

.335

.858

10

Strategies for modeling complex phenomena Total ítems

6 111

This scale was remove due to validity problems 84

a Number

of items used for data collection in the Survey number of items retained after statistical analysis and reliability test c It is widely recommended that the average inter-item correlation for a set of items be between 0.15 and 0.50 (Cupani 2008: 256) d (α) The Alpha coefficient (Cronbach’s alpha) varies from zero to one. Concerning its theoretical interpretation, some authors suggest a magnitude higher than .60. Nunnally (1970) suggests a magnitude of .70 as a minimum acceptable level for theoretical research. An Alfa of at least .85 is recommended if the scales are going to be used to take decisions regarding individuals, such as in psychological tests b Final

beliefs indexes. Figure 4.1 shows a graphical representation of a factorial analysis of scale nº4 regarding the conception of the role of values in scientific research.6 The most interesting aspect of the analysis consists in the modeling scientific beliefs thorough the construction of multiple linear regression models. The twenty two regression models were submitted to a strict analysis to verify the fulfillment of fundamental assumptions of regression analysis (linear relationship, multivariate normality, no or little multicollinearity, no auto-correlation, homoscedasticity). These models allow us to reconstruct the organizational complexity of a system of scientific beliefs. Table 4.2 shows a synthesis of the twenty two models. Conceptual maps were used to graphically represent statistical relations analyzed by regression models. Modeling networks of scientific beliefs using conceptual maps allow us to visually summarize the findings of two, three or even more regression models in a unified diagram, providing a synthetic view of the structure of scien6 Item-total

correlations are given in brackets for each entry. The correlation between an item and its factor represents the degree to which each item measures the same construct as the factor. Correlations with low values (less than 0.2 or 0.3) indicate that the item is not very well correlated with the factor, thus expressing another construct.

4 How and Why to Model the Complexity of Thought Systems

65

Fig. 4.1 Factorial analysis of Likert scale nº4: “Conception on the role of values”

tific beliefs. In order to produce an easily interpretable map, chromatic and shape distinctions were implemented through visual means (see Table 4.3). We offer a brief example of a visual model of an epistemic framework based on the analysis of three clusters of beliefs: axiological beliefs, ontological beliefs and anthropological beliefs. The empirical analysis reveals that conceptions about the role of value in scientific research (axiological beliefs) are simultaneously determined by the conception of reality (ontological beliefs) and the place of the subject in the construction of knowledge (anthropological beliefs). A conceptual map was produced to illustrate this analysis by articulating three regression models (see Fig. 4.2). We offer a brief example of a visual model of an epistemic framework based on the analysis of three clusters of beliefs: axiological beliefs, ontological beliefs and anthropological beliefs. The empirical analysis reveals that conceptions about the role of value in scientific research (axiological beliefs) are simultaneously determined by the conception of reality (ontological beliefs) and the place of the subject in the construction of knowledge (anthropological beliefs). A conceptual map was produced to illustrate this analysis by articulating three regression models (see Fig. 4.27 ).

7 The

nodes of the conceptual maps represent the variables of the regression models. The links between nodes show Beta weights, partial, semi-partial and zero-order correlations.

66

L. G. Rodríguez Zoya

Table 4.2 Synthesis of the 22 multiple regression models on scientific beliefs Theoretical construct explained by multiple linear regression model

Xn a Rb

R2c

Adjust R2d

DurbinWatsone

Lowest Highest tolerancef VIFg

1

Conception on Complex Systems (Model A)

2

.629

.395

.387

2.028

.770

1.299

2

Conception on Complex Systems (Model B)

5

.675

.456

.437

2.041

.644

1.553

3

Constructivist conception on Complex Systems

5

.662

.438

.418

2.103

.638

1.568

4

Realist conception on Complex Systems

5

.541

.292

.267

2.118

.651

1.535

5

Conception on the role of values Axiological neutrality on models and data (Model A)

2

.657

.432

.424

2.181

.760

1.315

2

.584

.341

.332

1.936

.760

1.315

7

Axiological neutrality on models and data (Model B)

5

.662

.438

.419

2.035

.647

1.546

8

Constructive role of values (Model A)

2

.589

.347

.338

2.171

.761

1.315

9

Constructive role of values (Model B)

5

.605

.367

.344

2.161

.651

1.535

10

Conception on the goal of science (Model A)

5

0.417

0.174

0.145

1.528

.651

1.535

6

(continued)

4 How and Why to Model the Complexity of Thought Systems

67

Table 4.2 (continued) Theoretical construct explained by multiple linear regression model

Xn a Rb

R2c

Adjust R2d

DurbinWatsone

Lowest Highest tolerancef VIFg

11

Conception on the goal of science (Model B)

3

0.725

0.525

0.517

1.795

.783

1.277

12

Social role of science Epistemic role of science Simple models and epistemic attributes Simple models and cognitive strategies

10

0.717

0.513

0.478

1.638

.484

2.066

10

0.569

0.323

0.274

1.718

.476

2.103

4

.574

.330

.309

1.846

.779

1.283

3

.504

.254

.236

1.932

.848

1.179

Simple models: cognitive strategies and epistemic attributes Simple models and conceptions on science, subject and reality

7

.725

.526

.499

1.797

3

.291

.085

.062

1.843

.793

1.260

Participatory social simulation models and epistemic attributes Participatory social simulation models and cognitive strategies

5

.627

.393

.369

1.981

.915

1.093

5

.484

.235

.205

2.007

.789

1.267

13 14

15

16

17

18

19

751

1.332

(continued)

68

L. G. Rodríguez Zoya

Table 4.2 (continued)

20

21

22

Theoretical construct explained by multiple linear regression model

Xn a Rb

Participatory social simulation models: cognitive strategies and epistemic attributes Participatory social simulation models and conceptions on science, subject and reality

10

Conceptions on complex models

5

a Number

3

R2c

.674

450

.603

Adjust R2d

DurbinWatsone

Lowest Highest tolerancef VIFg

.454

.409

1.983

.759

1.318

.202

.183

2.058

.793

1.260

.364

.338

2.173

.859

1.164

of independent variables (X) included in the regression model multiple correlation coefficient (R1.23 …) measures the correlation between one dependent (prediction) variable (Y) and a set of independent (explanatory) variables (X1 , X2 …Xn ) c The coefficient of determination (R2 …)—pronounced R squared—is the percentage of the vari1.23 ance of the dependent variable which is explained by the variation of independent variables. R squared represents the ratio of the explained variation to the total variation d R squared tends to increase its magnitude when further variables are added to the model. In order to take this issue into account, adjusted R squared (R2a ) is used e Durbin-Watson statistics is used to test the assumption of homoscedasticity, that is, the same variance: the variance around the regression line is the same across all values of the independent variable. When the assumption of homoscedasticity is violated there is heteroscedasticity. DurbinWatson varies from zero to four. Ideally, values between 1.5 and 2.5 express homoscedasticity f The tolerance coefficient is an indicator of the independence between one independent variable (X1) with respect to others independent variables included in the model (X2…Xn). Tolerance varies from zero to one. “A value near 1.0 expresses complete absence of multicollinearity: a variable Xi does not have any correlation with the rest of independent variables” (Cea D’Ancona 2002: 52). The higher the tolerance coefficient (value near 1), the greater the independence between X and the rest of the independent variables. Tolerance coefficients are calculated for each independent variable. Only the lowest tolerance in the model is shown in the column g Variance Inflation Factor (VIF). VIF quantifies how much the variance of the estimated regression coefficients is increased because of collinearity. The higher is the VIF, the greater multicollinearity, that is, higher correlation between independent variables. “A VIF of 1.0 indicates no relation between independent variables. Values higher than 10.0 express severe multicollinearity” (Cea D’Ancona, 2002: 52). VIFs are calculated for each independent variable. Only the highest VIF in the model is shown in the column b The

4 How and Why to Model the Complexity of Thought Systems

69

Table 4.3 Chromatic and shape distinctions for modeling scientific beliefs Dependent variable

R2

Independent variables (X1 , X2 , X3 , Xn )

The red rectangle node is exclusively reserved for identifying the dependent variable (outcome variable)

The blue circumference is used to express the value of adjusted R squared

Rectangle with curved edges is used to represent independent variables

Fig. 4.2 A model of an epistemic framework

The conceptual map integrates three regression models (Model nº5, 6 y 8 from Table 4.2). Let us make some brief observations. Conceptions of reality and conceptions about the subject explain the 42.4% of variation of axiological conceptions (Model Nº5). Nevertheless, conceptions about the subject play a much more relevant role with regards to conceptions of reality (Beta .587 and .125 respectively). Furthermore, conceptions about the subject and reality explain the 33.2% of variation in axiological Beta is a standardized coefficient (expressed in z-scores) that indicates the individual contribution of each independent variable to the variation of the dependent variable. Specifically, Beta weight expresses how many standard deviation units the dependent variable increases or decrease when the independent variable increases one standard-deviation and other independent variables are held constant. Partial correlation measures the degree of associations between two variables after controlling the effect of one or more independent variables. Semi-partial correlation represents the unique contribution of an independent variable to the model. It indicates how much R square increases with the inclusion of that variable in the regression equation. Zero-order correlation is the bivariate correlation between independent and dependent variables.

70

L. G. Rodríguez Zoya

neutrality (Model Nº6) and the 33.8% of conceptions concerning the inclusion of values in research (Model Nº8). Anthropological beliefs contribute, in a similar way, to explain both axiological conceptions, even though the relationship is stronger concerning the subject and inclusion of values.8 Differently put, conceptions of reality have more explicative weight in the formation of axiological neutrality beliefs (beta .153) than in those beliefs including values (beta .079).

4.5 Conclusions This paper has sketched a possible answer to the question of how and why to model the complexity of thought systems. Thinking and knowledge are not a passive reflection of a pre-constituted reality. On the contrary, there is a recursive loop through which thinking constructs reality while, at the same time, is constructed by it. For this reason, two concomitant challenges emerge: the problematization of reality and the problematization of the thought. These two challenges are intertwined in the study of complex problems with which contemporary societies must cope. We have coined the category of “complex problems” in order to highlight the fact that: – A complex problem is a socio-cognitive construction that emerges through a process of problematization where value judgments are involved. – A complex problem articulates ethico-political, practical and epistemic dimensions. – Complex problems require an interdisciplinary research strategy due to the fact that their constitutive processes and elements stem from different disciplines. – A complex problem is a meta-system encompassing the articulation of an observed system (the problematized experience) and an observing system (the intersection of many viewpoints on such an experience). The empirical research on a complex problem simultaneously requires a reflexive approach on how we think the complexity of such problem. In short, the objectification and reflexivity of complexity are two inseparable dimensions. Thus, we have suggested the importance of articulating a set of methods for modeling and simulating complex systems as well as the mega-cognitive strategy of complex thought. An epistemological model of complex thought (EMCT) has been proposed as a research strategy for modeling a meta-viewpoint on a thought system. In order to methodologically operationalize such a model, we have elucidated the theoretical relationship between the notions of paradigm, epistemic framework, beliefs and attitudes. We have shown the relevance of qualitative and quantitative methods for the production and analysis of empirical evidence concerning the different aspects of a thought system, particularly, discourse analysis, psychometric techniques and multivariate analysis. 8 Let

us observe Beta coefficients. The relationship between the node subject and value inclusion shows a beta weight of .547, whereas the relationship between the subject and axiological neutrality has a beta of .493.

4 How and Why to Model the Complexity of Thought Systems

71

From this standpoint, empirical research has been conducted for the sake of modeling the system of scientific beliefs within complexity science and social simulation. Multidimensional aspects of scientific beliefs have been empirically modeled through ten Likert scales and twenty-two regression models. A brief summary and a short empirical illustration of this task have been outlined, whereas a full report has been delivered elsewhere (Rodríguez Zoya 2013). The model in question constitutes a reflexive and self-critical meta-viewpoint from which complexity science can observe its own way of thinking and problematizing the paradigms guiding its knowledge-construction strategies. Constructing a better future is a political as well as scientific challenge that involves transforming complex problems. In order to cope with the complexity of these problems we should learn to make sense of our own thinking processes. Likewise, science may help us to problematize our way of thinking but, in order to do so; it should first be able to think itself. In the foregoing, we have offered some theoretical, epistemological, ethical and methodological tools for the development of a complex thought as well as a reflexive complexity science.

References Bertalanffy, L. V. (1976). Teoría general de sistemas. Fundamentos, desarrollos, aplicaciones. Buenos Aires: Fondo de Cultura Económica. Best, J. B. (2002). Psicología cognoscitiva. México D.F.: Thomson. Bruning, R. H., Schraw, G. J., Norby, M. N., & Ronning, R. R. (2005). Psicología cognitiva y de la institución. Madrid: Pearson. Bunge, M. (2009). A la caza de la realidad. Barcelona: Gedisa. Cabeza, R. (1987). Temas de psciología cognitiva. Buenos Aires: Tekné. Castorina, J. A. (2007). Construcción conceptual y representaciones sociales. Buenos Aires: Miño y Dávila. Cea D’Ancona, Á. (2002). Análisis multivariable. Teoría y práctica en la investigación social. Madrid: Síntesis. Cupani, M. (2008). Análisis psicométrico con SPSS. In S. Tornimbeni, E. Pérez, & F. Olaz (Eds.), Introducción a la psicometría (pp. 245–267). Buenos Aires: Paidós. Duveen, G., & Lloyd, B. (1990). Social representations and the development of knowledge. Nueva York: Cambridge University Press. Edmonds, B. (1999). Syntactic measures of complexity. Manchester: Manchester Metropolitan University. Foerster, H. V. (1996). Las semillas de la cibernética (Colección terapia familiar). Barcelona: Gedisa. Foucault, M. (1998). Las palabras y las cosas. Una arqueología de las ciencias humanas. México: Siglo XXI. Foucault, M. (1999). Polémica, política y problematizaciones. In M. Foucault (Ed.), Obras esenciales (pp. 991–998). Barcelona: Paidós. Gagné, E. D. (1991). La psicología cognitiva del aprendizaje escolar. Madrid: Aprendizaje Visor. García, R. (1997). La epistemología genética y la ciencia contemporánea. Barcelona: Gedisa. García, R. (2000). El conocimiento en construcción. De las formulaciones de Jean Piaget a la teoría de los sistemas complejos. Barcelona: Gedisa.

72

L. G. Rodríguez Zoya

García, R. (2006). Sistemas complejos. Conceptos, método y fundamentación epistemológica de la investigación interdisciplinaria. Barcelona: Gedisa. Gell-Mann, M. (1995). What is complexity? Complexity, 1(1), 16–19. Gómez, R. (2008). Otto Neurath: Lenguaje, ciencia y valores. La incidencia de lo político. Los Ángeles: California State University, Estados Unidos. Grize, J.-B. (1993). Logique naturelle et représentations sociales. Textes sur les Représentations Sociales, 2(3), 1–9. Grize, J.-B. (1996). Logique naturelle te communications. Paris: PUF. Grize, J.-B. (2012). Logique naturelle et représentations sociales. In D. Jodelet (Ed.), Les représentations sociales (pp. 170–186). Paris: Puf. Hanson, R. (1958). Patrones de descubrimiento. Observación y explicación. Madrid, España: Alianza. Hernandez Forte, V. (2005). Mapas conceptuales. La gestión del conocimiento en la didáctica. DF, México: Alfaomega. Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, inference, and consciousness. Cambridge: Harvard University Press. Johnson-Laird, P. N. (1987). Modelos mentales en ciencia cognitiva. In D. A. Norman (Ed.), Perspectivas de las ciencias cognitivas (pp. 179–231). Barcelona: Paidós. Koyré, A. (1999). Del mundo cerrado al universo infinito. Madrid: Siglo XXI. Latour, B. (1999). La Esperanza de Pandora. Ensayos sobre la Realidad de los Estudios de la Ciencia. Barcelona: Gedisa. Le Moigne, J.-L. (1990). La Modélisation des systèmes complexes. Paris: Dunod. Likert, R. (1932). A technique for the measurement of attitudes. Archives of Psychology, 140, 44–53. Longino, H. E. (1990). Science as social knowledge: Values and objectivity in scientific inquiry. Princeton: Princeton University Press. Maruyama, M. (1980). Mindscapes and science theories. Current Anthropology, 21(5), 589–608. Maturana, H., & Varela, F. (1972). Autopoietic system. Santiago de Chile: Facultad de Ciencias. Maturana, H., & Varela, F. (2003). De máquinas y seres vivos: Autopoiesis, la organización de lo vivo. Buenos Aires: Lumen. Minsky, M. (1965). Matter, mind and models. Paper presented at the International Federation of Information Processing Congress (Vol 1. pp. 45–49). Morales Vallejo, P., Urosa Sanz, B., & Blanco Blanco, A. (2003). Construcción de Escalas de Actitudes Tipo Likert. Cuadernos de Estadísticas Nº 26. Madrid: La Muralla. Morin, E. (1973). El paradigma perdido. Ensayo de bioantropología. Barcelona: Kairós. Morin, E. (1977). El Método I. La naturaleza de la naturaleza. Madrid: Cátedra. Morin, E. (1984). Ciencia con Conciencia. Barcelona: Anthropos. Editorial del Hombre. Morin, E. (1986). El Método III. El conocimiento del conocimiento. Madrid: Cátedra. Morin, E. (1990). Introducción al Pensamiento Complejo. Barcelona: Gedisa. Morin, E. (1998). El Método IV. Las ideas. Madrid: Cátedra. Morin, E. (1999). La cabeza bien puesta. Repensar la reforma. Reformar el pensamiento. Buenos Aires: Nueva Visión. Morin, E. (2001). El Método V. La humanidad de la humanidad. La identidad humana. Madrid: Cátedra. Moscovici, S. (1979). El psicoanálisis, su imágen y su público. Buenos Aires: Huemul. Neurath, O. (1983). Pseudorationalism of falsification. In O. Neurath (Ed.), Philosophical papers 1913–1946 (pp. 121–131). Boston, United States of America: Reidel. Nicolescu, B., Bianchi, F., Morin, E., & Motta, R. D. (1994). Carta a la transdisciplinariedad. http://www.pensamientocomplejo.com.ar/docs/files/aavv_carta_a_la_interdisciplinariedad.pdf. Accessed 10/9 2008. Nunnally, J. C. (1970). Introducción a la medición psciológica. Buenos Aires: Paidós. Piaget, J. (1973). Psicología y epistemología. Barcelona: Ariel. Piaget, J. (1978). Introducción a la epistemología genética. 1. El pensamiento matemático (Vol. 1). Buenos Aires: Paidós.

4 How and Why to Model the Complexity of Thought Systems

73

Piaget, J. (1979). Tratado de lógica y conocimiento científico. I. Naturaleza y métodos de la epistemología (Vol. I). Buenos Aires: Paidós. Piaget, J., & García, R. (2008). Psicogénesis e historia de la ciencia. México DF: Siglo XXI. Prigogine, I., & Nicolis, G. (1997). La estructura de lo complejo. Madrid: Alianza. Putnam, H. (1994). Las mil caras del realismo. Barcelona: Paidós. Putnam, H. (2004). El desplome de la dicotomía hecho-valor y otros ensayos. Barcelona: Paidós. Quillian, M. R. (1968). Semantic memory. In M. L. Minksy (Ed.), Semantic information processing (pp. 227–270). Cambridge, MA: MIT Press. Ricoeur, P. (1996). Si mismo como otro (1º ed.). México: Siglo XXI. Rodríguez, R. J. (2010). Herramientas informáticas para la representación del conocimiento. Subjetividad y Procesos Cognitivos. Revista del Instituto de Altos Estudios en Psicología y Ciencias Sociales (IAEPCIS). Universidad de Ciencias Empresariales y Sociales (UCES), 14 Herramientas informáticas y análisis del discurso(2 Otoño 2010), 217–232. Rodríguez Zoya, L. (2012). Sistemas complejos y conocimiento emancipador en América Latina. Notas acerca del rol social y político de un programa de investigación científica de larga duración. Revista Pacarina del Sur. Revista Crítica de Pensamiento Latinoamericano, Octubre–Diciembre. Rodríguez Zoya, L. (2013). El modelo epistemológico del pensamiento complejo. Análisis crítico de la construcción de conocimiento en sistemas complejos. Universidad de Buenos Aires y Universidad de Toulouse, Toulouse. Simon, H. (1973). La arquitectura de la complejidad. In H. Simon (Ed.), Las ciencias de lo artificial (pp. 125–169). Barcelona: España. Tulving, E. (1972). Episodic and semantic memory. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 381–402). New York: Academic Press. van Dijk, T. A. (1999). Ideología. Una aproximación multidisciplinaria. Sevilla: Gedisa. Varsavsky, O. (1975). Marco Histórico Constructivo para estilos sociales, proyectos nacionales y sus estrategias. Buenos Aires: Centro Editor de América Latina. Vilar, S. (1997). La nueva racionalidad. Comprender la complejidad con métodos transdisciplinarios (Colección Nueva Ciencia). Barcelona, España: Kairós. Voloshinov, V. (1976). El signo ideológico y la filosofía del lenguaje. Buenos Aires: Nueva Visión. Vygotski, L. (1995). Pensamiento y Lenguaje. Barcelona: Paidós. Wallerstein, I. (1996). Abrir las ciencias sociales. Informe de la Comisión Gulbenkian para la reestructuración de las ciencias sociales. México: Siglo XXI. Weaver, W. (1948). Science and complexity. American Scientist, 36, 536–544. Whitehead, A. N. (1944). Modos de pensamiento. Buenos Aires: Losada. Woolgar, S. (1991). Ciencia: abriendo la caja negra. Barcelona, España: Anthropos.

Part II

Language Change

Chapter 5

Linguistic Variation and Change: Approach from the Perspective of Complex Adaptive Systems Àngels Massip-Bonet

Abstract We provide general theoretical introduction to the topic of linguistic change from the perspective of dynamic systems and applying the concepts developed by the ECCO (Evolution, Complexity and Cognition) group’s paradigm as a “conceptual framework, based on an ontology of action”. We apply a look to linguistic change from complexity paradigm and we link the ECCO concepts with different areas of linguistics that deal with linguistic change.

5.1 Language System, Dynamic System A language system has a clearly dynamic behaviour: it makes sense that the activity (language) of a complex adaptive system such as the human brain should reflect the properties of such systems. In the same way that the brain is organized in neural networks, language is organized in networks as well. The way we do science should also more closely approach the way in which our brains function: in a transdisciplinary manner (to put it metaphorically) and organised in networks. Language offers the same characteristics that typify complex real-world bodies: it is open, boundary-free and replete with unstable parts; it has infinite cardinality with respect to its components, relations and configurations; it is dynamic, metastable (metastability is a temporarily stable state: equilibrium is broken when there are disturbances in a system, at which point the system shifts to another metastable equilibrium) and path-dependent; it is nonlinear, sensitive to initial conditions, exponentially amplifiable and in regions chaotic; and it is emergent, non-additive, nonmodularisable and organizationally intricate (Andrason 2015). When speaking of language, we do not refer only to the language system being treated as an object of study per se, but also to the communicative exchange, to À. Massip-Bonet (B) Department of Catalan Philology and General Linguistics, Projecte Scripta (FFI2016-80482P), Sociocomplexity—Complexity, Communication and Sociolinguistics Group, CUSC—Research Centre for Sociolinguistics and Communication, UBICS—Universitat de Barcelona Institute of Complex Systems, Universitat de Barcelona, Barcelona, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_5

77

78

À. Massip-Bonet

language as a shaper of cognition, to language as an expression of our nature (MassipBonet 2007: 236). And this conception is absolutely necessary if we wish to embrace as expansively as possible what linguistic change entails. In this paper, I want to offer some observations on linguistic change by looking at it through the perspective of what happens in dynamic systems. To develop these ideas on change, I have reviewed the philosophy underpinning the transdisciplinary work of the research group Evolution, Complexity and Cognition (ECCO). The aim of ECCO is to construct a coherent worldview to counter the fragmentation in science and in society today. The conceptual framework adopted by ECCO is based on an ontology of action, in which the fundamental constituents of reality are seen as actions and the agents that produce them (Heylighen 2008). Humans, when they adapt their linguistic resources to new contexts, change the language. This means that a language system and its uses are mutually constitutive. This perspective of dynamic systems gives us a way to focus on linguistic issues. I retake the concepts of Heylighen and I try to give ideas of application in the fields of linguistics that are most familiar to me or in which I have done some research. It is clear that each linguist, depending on their area of study, will be able to draw their own conclusions.

5.2 Linguistic Change Change is an essential part of the functioning of complex systems.1 Linguistic change occurs as a result of the constant emergence of variational phenomena resulting from the interactions between individuals and their surroundings. When one of these productions is generalized, we say that the change has been consolidated. Often, however, pre-existing or alternative forms to the consolidated form are kept in some varieties of the language.2 We make here a parallel with genetics, in which the stimuli of change can be intrinsic (from genes), and extrinsic, which are produced by the interactions with neighbouring cells. If we look at change in the words of any language, we would also have internal stimuli (which are already given from the etymon of each word and from its evolution according to the laws of each language throughout history) and some external stimuli (all contexts in which any word has been used over time, contact with other languages and/or varieties in a given territory). 1 As an example of the constant dynamics of complex systems, I like to mention what happens with

our eyes. The eyes do not stop moving as we look at any immobile object. In order to be able to see it in its static quality, our visual activity is in continuous dynamics (Kandel et al. 1991: 674), the eye must move because the image does not disappear from our field of vision. 2 Let’s see, for example, the catalan word ca of the Balearic dialect which was general in Catalan and alternated with gos (documented from the 15th century), ca has remained in the oral language as a habitual form only in Balearic and in northern Catalan (ALDC, map 1476). In English, in northern England and Scotland, they use bairn (Scot.beyrn) instead of child (Old English bearn, of Germanic origin; related to the verb bear).

5 Linguistic Variation and Change: Approach from the Perspective …

79

In this chapter, I will not carry out a specific analysis of factors of linguistic change in a specific language. Rather, I will only put forward an approach to the elements involved in the change in languages, making use of a conceptual adaption of the ECCO paradigm to linguistic change.

5.3 Systems Theory and Complexity General systems theory, viewed in part as a forerunner of complexity theory, treated systems as static, well-defined structures, which were objectively given. These assumptions, however, do not work for complex adaptive systems, such as societies, minds, markets or languages. In these systems, structures tend to be fuzzy, variable and to an important degree subjective (Heylighen et al. 2007). In this context, we must speak of self-organization and evolution. Self-organization is the spontaneous process whereby systems emerge and evolve, becoming ever more complex, more adaptive and more synergetic (Heylighen 2008). The most successful application of general systems to basic social science is that of Kuhn (1975, as we will see in 5.4.2). Self-organization (see Sect. 5.4.6 on organization) is the mutual adaptation and co-evolution of the system‘s initially autonomous components, that is, the agents (defining agent as a persistent producer of actions). Self-organization and evolution are seen as two aspects of the same process of spontaneous adaptation (viewed holistically). Through interactions, agents develop a network of increasingly synergetic relations that coordinate their activities. The system evolves continuously (with evolution being based on variation and selection internal to the system) and thus becomes more complex, more adaptive and more synergetic (see 5.4.2 below). Complexity emerges in systems through incremental changes even when the changes do not arise from a deliberate movement towards a goal. Change is produced continuously through quantitative accumulation that leads to periodic qualitative revolutions. Change is continuous and ineluctable (see 5.5.)

5.4 Ontology of Action: Dynamic Systems In his glossary, Heylighen defines the concepts of what he has called an “ontology of action” and groups them into six sections: (1) agents and evolution, (2) cybernetics, (3) cognition, (4) interactions, (5) system, and (6) organisation. Drawing on some of the terms defined by Heylighen, I will comment on them and put them in relation to linguistic change.

80

À. Massip-Bonet

5.4.1 Agents and Evolution Action: A change in the situation, from an initial state (cause) to a subsequent state (effect). An action is intrinsically relational, connects two situations and cannot exist independently of the phenomena that it links together. Any linguistic performance can be considered an action. State: The state of the world at a certain time is defined by the set of all actions that could be performed at that moment (Turchin 1993). Agent: An agent is an autonomous and persistent producer of actions. Generally, actions are aimed at achieving one result more than another. Through language, agents act in a directed manner, but not normally on the language itself. Only sporadically, if some aspect of a language has extra-linguistic or identity-related connotations of prestige (or lack of prestige) may agents act by using or ceasing to use some forms more than others. Who are the agents in languages taken as CAS? Either speakers (Beckner et al. 2009; Holland and Croft), or alternatively items in the system and speakers (according to Lee et al. 2009; Mufwene 2001). Event: An interesting distinction for linguistics, I find, is the difference between action and event. An action is voluntary, but an event is an action not produced by an agent. This distinction is very useful when we talk about linguistic change and about social variation and, more specifically, about language policy. An event would correspond to external facts that may influence the language and action to specific facts such as prohibitions, laws against the use of a language, censorship, etc. The state of a language is given by the history of this language, the set of linguistic actions that agents have done throughout history. Goal: One of the end states or attractors to which an agent’s actions is likely to lead. Goals in this sense are most of the time implicit: the agent does not have any awareness of what that end state is (Heylighen 2008). In language policy it is, of course, explicit. An attractor is also defined as a spontaneously evolved constraint. Variation: The on-going change in the state of world caused by following actions (Heylighen 2008). Variation is always to some degree blind, since no agent can foresee all the consequences of its actions. Sometimes, there are deliberate innovations, but these must be distinguished from the unconscious operation that is produced, for example, in basic learning and in the mental processing tendencies that have given shape to certain phonological, syntactic and semantic regularities. Natural selection: The selective retention of a given state, because no further actions occur that change that state. Such a state corresponds to what has been called an attractor of the dynamics, or a state with high fitness. And it is in this sense that we can speak of natural selection of languages. Often the loss of a species in the natural world (by natural selection) has been equated

5 Linguistic Variation and Change: Approach from the Perspective …

81

to the loss of a language. Most cases of language death, however, are not due to the conjunction of natural phenomena (or even to events) but are due to events and even to actions of human repression of one group over another, to human actions of genocide and linguicide, to human actions of wanting to subordinate (and even enslave) other human groups. Evolution: The long-term, directed change in the state of world towards higher overall fitness which results from the interplay of the (partly random, partly directed) variation of states and the eventual selection of states with higher fitness. Evolution can be seen as a search for fitness based on trial-and-error, where variation produces the trials, and selection eliminates the errors (Heylighen 1992). Evolution in biology (whether the target of selection is the gene or the individual as a whole) must be viewed as two parallel processes: adaptive change and the origin of diversity (Rovelli 2016: 219). Not all the characteristics of a system evolve at the same speed or advance at the same time. In general, a system evolves towards greater complexity and changes are irreversible. The origin of adaptive complexity in language would be analogous to the origin of adaptive complexity in biology (Christiansen and Chater 2008: 505). Thus, the adaptive complexity of language, much like that of biology, would emerge from linguistic variation “filtered” randomly through selection pressures that concern learning and processing (Massip-Bonet 2013: 47). But chance is just one of the causes. In my opinion, one could speak, as in the biological reality, of four factors: (1) Inheritance; (2) Evolution; (3) Environment; (4) Chance. And genes, in biological evolution, are only the prisms through which chance is filtered and refracted (Mukherjee 2017: 257). In the same way in languages, the intrinsic stimuli (see Sect. 5.2 above) would be the prism through which chance or other causes are filtered and refracted. These other causes could be, between others: easiness of articulation-which in some varieties is expressed in a way that in others is not3 ; the greater frequency of certain words -which represents a shorter route for these words regarding the neural networks of our nervous system. It is necessary to distinguish between evolution and history. History refers to the process of analysing the system as a product of irreversible evolutionary changes and emergent properties at each state of the system (Tim Ingold 2002; Tarride 1995). In the description of dynamic systems, there is always a historical element. Dynamic systems are not stable at any given moment (if anything, they are metastable—see Sect. 5.1 above). In biology, nothing makes sense without evolution (contrary to what happens in other fields, e.g., there is no need to understand the origin of computers to know how to use one or gain benefit from it). And so is it in linguistics. Language considered as a CAS leads us to understand the overall behaviour as it emerges out of local interactions among a large number of agents. Quite often this 3 For example assimilation tendencies, which are often considered that facilitate pronunciation, can

be much more frequent in some varieties than in others and may vary in several ways. For instance, in Mallorca they assimilate the c to the t (actor) ct > tt, or ps > ts (capsa). That does not happen in the other catalan varieties in this particular lexical item, but it can happen in others (rector [rett’o]).

82

À. Massip-Bonet

overall behaviour (and the emergent dynamics) is complex, but it is not determined by some previous design nor is it subject to a centralised control mechanism. It is hard to predict what will be the overall emergent properties of the system, based only on knowledge of the parts that make up the system. As we know, change is an essential element in the behaviour of complex systems. Systems are adaptive in the sense that they have a capacity to evolve in response to a changing environment. From local interactions, complex global models with new properties can emerge. Besides that, language evolves over time, and evolution always requires a transdisciplinary approach. This article arises from this effort for transdisciplinarity.

5.4.2 Cybernetics (See Introduction) Cybernetics is a transdisciplinary approach for exploring regulatory systems—their structures, constraints, and possibilities. It concerns scientific study of how humans, animals and machines control and communicate with each other (see also Introduction to this book). Cybernetics, which has been defined by some as the “science of complexity” (Ashby 1976; Simon 1990), came to be the leading discipline focused on complex systems. And that is why a long list of fields of study have influenced or have been influenced by cybernetics (game theory, systems theory (a mathematical counterpart to cybernetics), control systems, electrical network theory, mechanical engineering, logic modeling, sociology, architecture, organizational theory, evolutionary biology, neuroscience, anthropology, and psychology. Linguistics should be another field where to apply cybernetics concepts (see Introduction of this book): Cybernetics is applicable when a system being analyzed incorporates a closed signaling loop—originally referred to as a “circular causal” relationship—that is, where action by the system generates some change in its environment and that change is reflected in the system in some manner (feedback) that activate a system change. We see clearly that linguistic systems are systems of this kind. Ashby (1976) organizes the main subjects relating to machines, living organisms, control and information, and he sets out concepts and methods for the treatment of complex systems in which the concepts of transformation, variety, feedback and the black box provide powerful tools to tackle complexity. Bertalanffy (1976) develops the theory of open systems and gives shape to the notion that when we define a system we are also necessarily defining the environment with which it exchanges material. Bertalanffy was an organicist describing open systems and living organisms, wholes that exchange matter and energy with their environment. The distinction between open and closed systems adds new elements of understanding, especially in relation to the capacity of certain systems for self-organization (such as linguistic system or, conversation (Martorell 2013)). If we focus on closed systems, the second principle of thermodynamics—entropy—demonstrates the inexorable march of systems towards the equiprobability of their components’ existence,

5 Linguistic Variation and Change: Approach from the Perspective …

83

the disappearance of organizations, the end of the universe. This theory, however, did not account for the generation of organization that is being produced at every moment (negentropy): See Prigogine’s thermodynamics, ecological systems modelling, business administration, etc. In cybernetics, open systems are wholes interacting with inputs, throughputs and outputs of energy and information. Through negative feedback, systems remain in dynamic equilibrium. The essential concepts in the cybernetics model include: codification (communication triggers), negative entropy (organisation of energy from the environment to maintain the system) and equifinality (different development paths can lead to the same destination). All these concepts I consider that are capital in linguistics. The basic model of cybernetics features a controlled system, feedback and three components: detector (information), selector (preferences), effector (action). Individual behaviour can be analysed using this formulation. For example, in the interpersonal, the detector would be communication, the selector transaction and the effector organization. Interpersonal components can be combined into social composites. Kuhn (1975) employs a conceptual structure that views the basic disciplines of social science not as sociology, economics and political science, but as communication, transaction and organization. Language is the basic tool in order to accomplish individual, interpersonal and social systems running.

5.4.3 Standardization Process of Languages In my view, in this area of cybernetics, it is interesting to establish a parallelism between how, on one hand, a change spreads through a society—always in keeping with a gradual process both within the language, in lexis, and outside it, in different social strata and geographical regions—and how, on the other hand, standardisation processes occur (Massip-Bonet 1994). (1) Step 1: Selection process Co¸seriu (1992) defined selection in linguistic change as an alternative use of an old form and a new tradition, so that any change is less a change than the selection of one form and the exclusion of others. In standardization processes there is always a step of selection, first of the variety (or varieties in polycentric kinds of standardization) and then of the elements (from a pool of historical forms). Selection is always the result of dynamic tensions that determine the triumph of a solution based on solutions adopted in other parts of the system, that is why it is difficult to predict the linguistic form that will become widespread.

84

À. Massip-Bonet

(2) Step 2: Generalization process There are two kinds of generalization: extensive and intensive. Extensive generalization occurs among all speakers in a group, while intensive generalization appears in all of the words that contain the affected phoneme or other paradigmatic class. With respect to the standard, the concept of extensive generalization is quite important. One can say that linguistic change has become generalized when speakers use the selected forms spontaneously, either always or sometimes, alternating with the old forms. We can say that a standard has become generalized when a form is accepted and used without any debate by the users of a language. The distinction between diversion and disturbance is highly interesting when treating the standardization topic and other related sociolinguistic topics. Diversion: Changes in the situation of an agent, over which the agent has no control (though the agent can try to control the effects). Disturbance: A negative deviation (diversion). Disturbances can originate in the environment or arise from the malfunctioning of an agent.

5.4.4 Cognition Cognition: Acquisition, processing, storage and use of information and knowledge (see knowledge, below) to support intelligent decision-making. A perspective that must be taken into account when we speak of linguistics and linguistic change is that of collective (or social) cognition (Heylighen 2015). For languages it is a very relevant perspective, as they have innate and acquired cognitive foundations and for changes to become a part of the system, they should always be generalized in society. The social brain4 implies a network of cortical and subcortical regions that are especially sensitive to social situations. Each member of a community, with their cognitive characteristics and brain (and their social brain), participates in social and linguistic interaction and contributes to the conformation of the society they belong to. The contributions of the studies on the social brain are changing the way to understand the reasons and ways of many social phenomena, and they are absolutely necessary to explain the contemporary society in all its aspects, including the linguistic one social brain role in the language setting was clear even before this concept (social brain) was discovered. Social cognition requires skills such as recognition of affection, memory and facial recognition, an adequate interpretation of affection and prosody, and knowledge of the minds of others (as theory of mind makes us apparent). Variety of action: The number of actions that an agent can potentially execute. The larger an agent’s variety of action, the larger the variety of diversions that the 4 Neural

network that allows the perception of social signals, the formation of social memory and social experience.

5 Linguistic Variation and Change: Approach from the Perspective …

85

agent can deal with, because different types of diversions typically require different types of actions. This is a generalization of Ashby’s (1958, 1976) well-known “law of requisite variety” (Heylighen 1992). The distinction between information and knowledge is meaningful: information is any change of condition that reduces the agent’s uncertainty, while knowledge is the ability derived from experience, education and/or communication to anticipate the consequences of a given state or action. Knowledge differs from information in that it produces general predictions that are applicable to different situations, while information—strictly speaking—only applies to the present situation. Knowledge: The ability, typically derived from experience, education or communication, to anticipate the consequences of a given state or action. Intelligence: Given certain information, the degree to which an agent is able to make good decisions, i.e. selections of actions that maximally accumulate utility in the long term. Intelligence has two components: (1) knowledge (see previous paragraph) and (2) fluid intelligence (the ability to internally explore many different combinations of possible events and actions in order to find the one that, according to the existing knowledge, will produce the greatest utility). In this section, it should be mentioned that the functioning of the brain helps us to frame the understanding of how we build meaning (and semantic networks). The meaning of a concept depends on the way in which it is connected to other concepts. There are many types of semantic relations (between words, between meaning and significance, between words of the same origin and lexical family, in specific terminologies, etc.) and all this in addition to the complexity of the meanings themselves. Depending on the type of relation that is studied, we will present a specific network structure that will allow us to see the associations between the concepts. The model of brain functioning is key to helping us to understand how all these relationships are produced. That is why nowadays neuroscientists are working with artificial neural networks. Neural networks are artificial intelligence algorithms that imitate the way bioelectric networks in the brain work (formed by neurons and their synapses). They simulate the architecture and the behavior of human brain and nervous system. The memory also works like a network. The human brain links the information that comes from the outside by associating it with other information already existing in the memory. Once the brain detects association tracks, it looks for stored data that matches the association tracks of the new information. And finally, it eliminates items it does not consider useful and stores in memory what it has considered could be useful in the future. Knowledge of any type is stored in the connection and in the correlations between the processing units, not in the units themselves. These units would be equivalent to neurons or networks of neurons and the connections would be the synapses. Semantic linguistic knowledge is also found in relations, not in individualized words.

86

À. Massip-Bonet

In second language acquisition, it is also very useful to consider the concepts of knowledge and intelligence as different abilities to be achieved in order to be competent in a learned language (besides obviously the cognitive features of the agents).

5.4.5 Interactions Interaction: Reciprocal effect of two (or more) agents (say, A and B) on each other: the action performed by A creates a condition that triggers another action (reaction) from B. This second action in turn affects the condition of A, stimulating it to react in turn, and so on. Synergy: Any gain caused by an interaction—an interaction with a positive sum. The opposite of synergy is friction, and the relation between agents involved in a synergetic or positive-sum interaction is cooperation. The relationship between agents involved in an interaction with friction or a negative sum is conflict. All of these terms are in current use in sociolinguistic terminology. We can see that they are generally applicable to any interaction in complex systems. These concepts can also be applied to the conversation, so that from the interactive linguistic participation of two or more actors, synergy may result -and therefore, creation of knowledge, understanding, cooperation. Or friction, in which linguistic interaction serves only to conclude the conflict. In any case, linguistic interaction (even if not only oral language) is almost the only way to reach cooperation. Synergy in systems theory is the process that runs between antithetic forces of nature that take to a development of strongly complex structures (Parra Luna 1992: 439).

5.4.6 Systems System: A group of agents held together by a shared constraint or network of bonds. The agents in the system can be seen as the system’s components. To the extent that the agents in a system share a goal, the system functions as a higher-order agent. Supersystem: A system whose components are themselves systems, called “subsystems”. Complex adaptive system: A system consisting of many interacting agents, where their interactions are not rigidly fixed, pre-programmed or controlled, but continuously adapt to changes in the system and in its environment.

5 Linguistic Variation and Change: Approach from the Perspective …

87

The problems that presently confront individuals, organisations and society at large, all concern complex, evolving systems, such as the global ecosystem, society, the economy, and our own internal system of thoughts and emotions (see 5.3 in this chapter). Distributed cognition: The acquisition, storage and use of information and knowledge distributed over different agents in a system, so as to support their collective intelligence (Heylighen et al. 2004). Language is one of the examples of such supporting mechanisms for the exchange of information. Many emerging phenomena that we observe in complex adaptive systems can be understood as resulting from shifts in balance and stability, when the system moves between equilibria. These shifts can result from gradual or sudden change and are typically called critical transitions when the system reaches ‘tipping points’. When complex systems reach (or are near) attractors (see 5.4.1 above), they demonstrate stability and variability. When a system or one of its parts shifts from one relatively stable attractor to another one in a process of phase transition, the transition point may be marked by increasing behavioural variability. When systems pass through a phase transition, they self-organize and the organization that emerges may be new and differ qualitatively from the various organizations of the systems or their parts (see Massip-Bonet 2013). The natural state of a language system can be defined as a dynamic adaptation to a specific context. Constraint: A limitation on the variety of action for some agent(s). An attractor (see 5.4.1 Goal, too) is a spontaneously evolved constraint. Bond or connection: A stabilised interaction between two agents exerting a constraint on their further action. Environment: Everything that is considered to be external to a given agent or system, but that still interacts with it. This means that the environment provides the initial conditions or input that triggers an interaction, while accepting the output or change produced by that action. In linguistics we should talk of environment of the speakers and environment of words inside the linguistic system itself (that we commonly name context). Medium: The medium is often the environment shared by the interacting agents, but it can also be internal to the agents.

5.4.7 Organization Organization: A stabilised network of interactions between agents that functions to ensure the coordination of their action and specifies the particular roles and interactions between the system’s agents. It can be imposed from the outside or emerge from self-organization. Coordination: Coordination between the agents pools their resources, material as well as informational, so that the group as a whole can act more effectively and

88

À. Massip-Bonet

intelligently than each agent individually. This is the origin of collective intelligence. Moreover, the system further increases its knowledge and intelligence by interacting with its environment as it learns from these interactions so as to become ever more effective in anticipating phenomena and choosing appropriate actions. Thus, the network of relationships between the agents starts to behave like a neural network, capable of increasingly sophisticated cognitive processes. Such self-organizing systems are distributed in their organization: it is in general difficult to distinguish separate components or subsystems performing separate functions; the components cooperate as a whole (Heylighen 2008). Mediator: A regulatory structure external to the agents that promotes coordination between them. It may emerge from self-organization or be imposed by an inside or outside agent. Stigmergy: a form of indirect coordination via the medium, where the trace left by an action in the medium stimulates the performance of a subsequent action, thus building further on the work that has already been done (Heylighen 1992; Parunak 2006). Applied to language, this concept is useful in sociolinguistics: according to the environment and the language in which a person approaches another (selection influenced by prestige or for other reasons), a kind of uses or another kind are generated between those who intervene in the communicative action.

5.5 Proximal and Distal Causes Causes of change are mechanisms generating change in systems. And the narratives that enable human agents to express such changes meaningfully. Proximal causes affect the phenotype (morphology) and behaviour, while distal causes affect the genotype and its history. Experimentation typically helps in the determination of proximal causes, while distal, or ultimate, causes are determined by inference based on historical narratives (Rovelli 200: 135). Nearly all ecological problems involve proximal and distal causes (and this is true of linguistics as well), because ecology (like linguistics) deals with highly complex systems. General relativity teaches us that space is not an inert box, but rather something dynamic: a kind of mobile snail shell in which we are contained—one which can be compressed and twisted. I would say that, metaphorically, space changes with social, political, economic or other parameters change. The other consequence of relativity theory is this: just as the idea of a continuous space that contains things disappears, so does the idea of an elementary and primal ‘time’ flowing regardless of things. This means that change, in time and space, is ubiquitous.

5 Linguistic Variation and Change: Approach from the Perspective …

89

5.6 The Observer The attention to the observer starts hand to hand with the birth of philosophy in ancient Greece, goes on with the conception of subject and objectivity, the development of mechanicist science and positivism in the Modern world and arrives to the relativity revolution, the constructivism and the complex thinking today. Complexics5 includes the observer at the very heart of its theory. Vallée (1990: 240) uses three different ways to refer to complexity: the “observer’s feeling”, the “observer’s judgment”, and the “name” of a subject-object relationship (an objective expression and a subjective one, which the relationship does not put into confrontation). He views reality as a mental construction. Morin (1986) says that the method for acting on complex systems cannot be the scientific method (or what is today called scientific, a term whose meaning is undergoing a process of evolution): he calls for complexity to be tackled through a process of action and reflection. What tools or instruments of reflection and action do we have? Education, culture, science, technology, democracy, justice. Bachelard (1985: 130) notes that every phenomenon of reality is a net of relationships. Simple ideas are working hypotheses, working concepts that must be revisited to be given the most fitting etymological role. New cybernetics (21st century) views information as constructed and reconstructed by an individual interacting with the environment. This provides an epistemological foundation of science, by viewing it as observer-dependent.

5.7 Final Considerations We could visualise the processes that are producing change in a language at any given moment as a set of gears in which the cogwheels are turning, some according to regular laws of evolution and others according to irregular evolving processes. At the same time, each cogwheel has different speeds of motion. As for the Romance languages, for example, one of these cogwheels would include Latin-based lexical elements, others would include elements that have come from other languages, and still others would include elements of speciality languages. Seen in this light, it becomes clearer that each cogwheel moves at a different speed. The set of gears keeps turning through time and there are emergences produced from its interaction with the environment and among the elements themselves (even among the elements that we have pictured as included in different cogwheels). For example, patrimonial words do not follow the same rules of phonetic evolution as cultisms (for example, coming from the etymon RAD˘ıU we have in catalan the patrimonial word raig ‘ray’ and the semicultism radi ‘radius’); words imported from 5 See Bastardas-Boada (2016): Complexics

and to study complexity.

instead of Complexity when we refer to our way to look

90

À. Massip-Bonet

other languages do not always follow the rules of the change of their own either (for exemple, cat. merci [‘mεrsi] from French merci, very usual nowadays in oriental Catalan, while mercès is a more ancient form for ‘Thank You’ and it is a form according the evolution rules of Cat.). Patrimonial words often have adaptations of the phonemes that are typical of the inventory of their own language (such as the case of cat. maco, where the Spanish sound [×] (de majo), which is not in the phonological system of Catalan, was adapted into [k]). In conclusion, linguistic change arises as a consequence of the constant emergence of variational phenomena resulting from the interactions of individuals among one another and with the environment and also of the interactions between elements in the linguistic system. After considering change and the time-vector in linguistics, I propose semantically adapting the term diachrony and altering its intension, so that diachrony would refer to the integrational study of linguistic variation—the true essence of linguistics—which always has a time dimension. We speak of dynamic synchrony and that is diachrony. I see no possibility of doing diachrony without integrating synchrony and, of course, it is not possible to do synchrony without diachrony, due to the fact that it is already integrated in synchrony. The next challenge for us is to apply this framework to a specific language and to use it when we treat other linguistic topics.

References ALDC, Veny, J. i Pons, L. (des de 2001). Atles Lingüístic del Domini Català. IEC. Andrason, A. (2015). A complex system of complex predicates: Tense, taxis, aspect and mood in basse mandinka from a grammaticalization and cognitive perspective. A doctoral thesis supervised by Dr. Marinna Visser, Stellenbosch University. Ashby, W. R. (1958). Requisite variety and its implications for the control of complex systems. Cybernetica, 1(2), 83–99. Available at http://pcp.vub.ac.be/Books/AshbyReqVar.pdf, republished on the web by F. Heylighen—Principia Cybernetica Project Ashby, W. R. (1976). Introducción a la cibernética. Buenos Aires: Editorial Nueva Visión. Bachelard, G. (1985). O novo espírito científico. Rio de Janeiro: Tempo Brasileiro. Bastardas-Boada, A. (2016). Complexics as a meta-transdisciplinary field. http://www. pensamientocomplejo.org/blog/index.php/2017/02/lecturas-emergentes-del-congreso-mundialde-pensamiento-complejo/. Beckner et al. (2009, December). Language is a complex adaptive system: Position paper. Language Learning, 59(1), 1–26. Bertalanffy, L. V. (1976). Teoría General de los Sistemas. México: Fondo de Cultura Economica. Christiansen, M. H., & Chater, N. (2008). Language as shaped by the brain. Behavioral and Brain Sciences, 31, 489–558. Co¸seriu, E. (1992). Le changement linguistique n’existe pas. Communication et Cognition: An Interdisciplinary Quarterly Journal, 25, 121–135. Croft, W. (2000). Explaining language change: An evolutionary approach. London: Longman. Crystal, D. (2000). Language Death. Cambridge. Edelman, G. (1987). Neural Darwinism. New York: Basic Books. id 89 The Remembered Present. New York: Basic Books.

5 Linguistic Variation and Change: Approach from the Perspective …

91

Heylighen, F. (1992). Principles of systems cybernetics: An evolutionary perspective. In R. Trappl (Ed.), Cybernetics and Systems ’92 (pp. 3–10). Singapore: World Science. Heylighen, F. (2008). Glossary of ECCO concepts. http://ecco.vub.ac.be/?q=node/99. Heylighen, F. (2015). Cognitive Systems. A Cibernetic Perspective of the new Science of the Mind. Lecture Notes 2014–15. ECCO, Vrije Universiteit Brussels. Heylighen, F., Cilliers, P., & Gershenson, C. (2007). Complexity and philosophy. Complexity, science, and society. Oxford: Radcliffe Publishing. Retrieved January, 16, 2008. Ingold, T. (2002). On the distinction between evolution and history. Social Evolution & History, 1(1). Kandel, E. R., Schwartz, J. H., & Jessel, T. M. (1991). Principles of Neural Science. Connecticut: Appelton&Lange. Klimasauskas, C. C. (1988). Neural Works. Sewickley, PA: Neural-Ware Inc. Kuhn, A. (1975). Unified social science: A system-based introduction. Illinois: Dorsey Press. Lee, N., Mikesell, L., Joacquin, A. D., Mates, A. W., Schumann, J. H. (2009). The interactional instinct: The evolution and acquisition of language. Oxford: Oxford University Press Mark, D. H., & Frank, A. U. (Eds.) (1991). Cognitive and linguistic aspects of geographic space. In NATO ASI Series (Series D: Behavioural and Social Sciences) (Vol. 63). Dordrecht: Springer. https://link.springer.com/chapter/10.1007/978-94-011-2606-9_12. Martorell, X. (2013). Conversation as emergent function. In Complexity perspectives on language, communication and society (pp. 75–84).Berlin, Heidelberg: Springer. Massip-Bonet, A. (2007). Canvi I Variació: Noves metàfores, Noves Prospeccions in Actes del Congrés de l’ AILLC (Universitat de Girona 8-13 de setembre de 2003) (pp. 235–243). Barcelona: Publicacions de l’Abadia de Montserrat. Massip-Bonet, A. (1994). Observations sur la variation linguistique: des exemples dans la langue catalane médiévale. Communication and cognition: An Interdisciplinary Quarterly Journal, 27, 337–349. University of Ghent. ISSN: 0378-0880. Massip-Bonet, A. (2013). Language as a complex adaptive system. In A. Massip-Bonet, A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society. Berlin: Springer. Morin, E. (1986). El método: la naturaleza de la naturaleza. Madrid: Càtedra. Mufwene, S. S. (2001). The ecology of language evolution. Cambridge University Press. Mufwene, & Salikoko, S. (2013). The emergence of complexity in language. In A. Massip-Bonet, A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 197–218). Heidelberg: Springer. Mufwene, S. S. (2017). Language vitality: The weak theoretical underpinnings of what can be an exciting research area. Language, 93(4), e202–e223 Mufwene, S. S., Coupé, C., & Pellegrino, F. (2017). Complexity in Language. Developmental and Evolutionary Perspectives. Cambridge: Cambridge Univerity Press. Mukherjee, S. (2017). El Gen. La Campana: Una història íntima. Parra Luna, F. (1992). Elementos para una teoría formal del sistema social. Madrid: Ed. Complutense. Parunak, H. V. D. (2006). A survey of environments and mechanisms for human-human stigmergy. In D. Wenyns, H. V. D. Parunak, F. Michel (Eds.) Enivronments of Multi-Agent Systems II, (Vol. 3830 of LNCS, pp. 163–186). Springer. Rovelli, C. (2016). Set lliçons breus de física. Barcelona: Anagrama. Sampson, G., Gil, D., & Trudgill, P. (Eds.). (2009). Language complexity as an evolving variable (Vol. 13). Oxford University Press. Simon, H. A. (1973). The organization of complex systems, In H. H. Pattee (Ed.), Hierarchy Theory, G. Braziller, New York, NY, pp. 3–27. Compiled in Simon, H. A., Models of Discovery, Boston: Reidel, 1977, pp. 245–264. Simon, H. A. (1990). Invariants of human behaviour. Annual Review of Psychology, 41, 6 Tarride, M. T. (1995). Complejidad y sistemas complejos in História, Ciéncias, Saúde, Manguinhos, II(1), 46–66.

92

À. Massip-Bonet

Turchin,V. F. (1993). The cybernetic ontology of action. Kybernets, 22(2), 10–30. Vallée, R. (1990). Sur la complexité d’un systéme relativement à un observateur Revue Internationale de Systémique, 4(2), 239–43. Wiener, N. (1989). The Human Use of Human Beings: Cybernetics and Society. USA: (HoughtonMifflin).

Chapter 6

Some theoretical Prerequisites for the Integrated Study of Linguistic “Macrochange” Enrique Bernárdez

Abstract The term macrochange (following the example of macroevolution as used in evolutionary theory) is here proposed to characterize changes affecting several areas of a language at the same time with a unitary cause, inside or outside language proper. A general theoretical model for this type of change will be proposed. As a case study, a set of changes in Icelandic will be analysed that lead to the creation of a widespread system for the marking of the agents’ or experiencers’ responsibility, evidentiality, etc., and that affect the lexicon, morphology, syntax and discourse. This macrochange will be defined in the terms of the integrated view of language and language change proposed here, which makes it necessary to adopt and develop the methods and concepts of complexity theory as applied to language. From this perspective, ‘language’ has to be taken to include not only structures and usage, but the whole ‘ecological niche’ where it exists and is in use by human beings in specific cultural and historical situations; it is to be viewed as a complex natural phenomenon.

6.1 Introduction. On the Complexity of Linguistic Change 6.1.1 Complexity, Language, and Change Language is a complex phenomenon; no one would reject this tenet nowadaysalthough many linguists still carry out their research through a radical simplification of the object ‘language’, trying to minimize complexity. At any point in time, a language can be characterized as being in a stable or, rather, metastable state. A multiplicity of attractors keeps the ‘system’ balanced. If

Partial support for work on this topic was provided by Research Project FFI2013-41366-P. MINECO. Lingüística cognitiva y lenguaje disfuncional. E. Bernárdez (B) Department of Linguistics, Universidad Complutense, Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_6

93

94

E. Bernárdez

the balance is disturbed, the metastable states yield their place to new, related, equally metastable spaces. In Massip-Bonet’s words (2013: 48): When a system or one of its parts changes from one relatively stable attractor state to another in a process of phase transition, the transition point may be marked by increasing variability of behaviour (…). When passing through a phase transition, systems engage in self-organization and the organization that emerges may be new and qualitatively different from the various prior organizations of the systems or their parts (…). The natural state of a language system can be defined as a dynamic adaptation to a specific context.

The same author (p. 40) writes that “[w]hen speaking of a language system, we do not refer only to a language, but also to the communicative exchange, to language as a shaper of cognition, to language as an expression of our nature…”. This point is of the greatest importance because, as we shall see, when dealing with change, and most especially with what in this paper is termed macrochange, it is necessary to go beyond the borders usually assigned to language and consider its intimate relations with the ‘non-linguistic world’ which, moreover, can be said to take primacy over the linguistic structures proper. Bastardas-Boada (2013: 18ff.) emphasizes a similar view. Similar views are everywhere to be found, suffice it to give two additional references: Bartmi´nski and Zinken (2009) and Krawczak (2007), whose views on language and meaning are close to those of the author of this chapter.

6.1.2 Language as a Complex Natural Phenomenon Although the metaphors commonly used to talk about language characterize it as an ‘object’, it clearly is no object except in an extremely abstract way. Weigand (210: 539) makes this point clear: (…) language belongs to human beings. It has taken some time to acknowledge this obvious fact and to recognize that there is no independent object language (…). Language as a natural phenomenon is comprehensible as the ability to speak, and this cannot be described independently of other communicative abilities. The nature of language depends on the nature of human beings.

Language is a biological-cultural object, thus sharing a double type of complexity: the complexity inherent in all biological systems and the complexity typical of socio-cultural organization (cf. Severi 2004). The biological element of language is obviously linked to cognition, which has been the object of much research in recent times. It is not necessary to enter into this point here, therefore: the complexity of the brain’s functioning is reflected in language. (Williams 2005). This also means, by the way, that recent developments in distributed, synergic or collective cognition also increase the ‘openness’ of language in its biological sense, and its complexity, as a consequence of the interaction of a number of participants in the cognitive ‘production and processing’ of language itself (Bernárdez 2007a, b, 2008; Garrod and Pickering 2004; Weigand 2010; Wilson 2004). This point needs to be kept in mind: in the analysis of linguistic change it is necessary to take into

6 Some theoretical Prerequisites for the Integrated Study …

95

account the collective role of the speakers, in such a way that language must not only, simplistically, be understood in the terms of individual speakers. As Haspelmath (2004: 24) writes, (…) it is impossible to understand language change phenomena if we see them as divorced from the speakers. If we talk about a morpheme traveling along a pathway, we should be aware that this is a very abstract metaphor that may invite all kinds of unwarranted inferences. We need to be careful with metaphors, and we should make more efforts to go down to the micro-level of individual speakers and derive the observed constraints on structural changes from known constraints on speakers’ linguistic behavior.

6.2 Linguistic Change 6.2.1 The Usual View Linguistic change is most frequently seen in individual terms, i.e., single changes affecting a single—or a very limited—part of the linguistic system are studied at a time. The system therefore ‘changes’ element by element or in small parts of a subsystem at a time. Thus, for instance, the First Germanic Consonant Shift is a change involving the Indoeuropean stop consonants as a whole, and only them; but of course these consonants only form a subsystem of the Ide. phonological system as such. It is fairly clear, too, that not all the changes took place simultaneously, that is, that some consonants were affected first and some only rather late. The same happened during the Second, High German Consonant Shift, where the gradual character of the changes affecting the individual consonants is still visible in their dialectal, geographical distribution.

6.2.2 Macrochange 6.2.2.1

Definition of Macrochange

It is only extremely infrequently that changes affecting several subsystems are analysed as such. I propose the term macrochange for this kind of linguistic change, following the usage of the term macroevolution in the biological sciences for evolution at or above the species level (see Kutschera and Niklas 2004). In the case of language, it can be defined as a change affecting a number of the (sub)systems of a language and due to one or a small coherent set of changes inside or outside the language system as such. In many cases this type of change was traditionally and justly criticized and rejected; for instance, the old views on the alleged adoption of the Bantu noun classes system by non-Bantu languages of Western Africa, which “by some momentous and amazingly rapid process of fusion (…) acquired, almost overnight, entire chunks of

96

E. Bernárdez

Bantu morphological structure” (Welmers 1973: 2). Some quite systematic, generalized changes affecting large areas of the language system can be identified, however.

6.2.2.2

Contact, Diffusion and (Macro)Changes

Such changes are usually the result of situations of strong language contact. For instance, Schieffelin (2002) shows that in a short period encompassing just one generation, the Basavi Kaluli language of New Guinea, under the strong pressure exerted by the missionary efforts of “Australian Christian fundamentalists” developed a whole set of changes affecting several distinct systems of the language, such as the introduction of words, grammatical structures, and discourse genres for marking and keeping various types of European-based time. The introduction of European-style institutionally organized activities in which participation was regimented and monitored (…) in the early 1970s gave rise to new ways of dividing days, weeks, months, and years based on linguistic innovations in Kaluli and Tok Pisin. Individuals who aligned themselves with mission organizations referred to these new economies of time, using particular expressions to differentiate themselves and their activities from those who were not similarly positioned. Simultaneously, new genres such as literacy lessons and sermons delineated time in terms of oppositional dichotomies that were temporally less specific but nonetheless linked to notions of social differentiation based on affiliation with a Christian community and/or identification with a nation-state. (Schieffelin 2002: 5).

We could identify here a generalized process of individuation of language use, as a complement or partial substitution of the earlier exclusively interpersonal conversation constituting text genres or types. This would go together with a new view of time—including time-sequence—which has to do with the Christian ideology of past, present, and future as steps in the human being’s way towards god. An interpretation like this would not, however, be readily acceptable in the usual language-historical terms. The procedure would be to analyse a set of changes affecting (a) the time-related vocabulary (semantic change), (b) the reanalysis of verbal forms resulting in the creation of a sequential time-related subsystem of the tenses (morphological change), and (c) the creation of new discourse genres (pragmatic change). Even if they may have a lot in common, they would be ‘best’ seen in an independent fashion—as the mechanisms involved are considered to be different in every case: pragmatic change is not semantic change is not morphological change. In general, a strict separation is kept between what was traditionally termed ‘internal and external change’, or ‘internal versus contact-induced change’. This is not the only acceptable view nowadays, however: “the dichotomy typically presented as ‘contact-induced’ or ‘external’ versus ‘normal’ or ‘internal’ change needs to be significantly revised, if not dissolved… (Ansaldo 2004: 485). This author takes his point of departure in Enfield’s book on ‘linguistic epidemiology’ (2003). A very similar view is held by Heine and Kuteva (2003). And from the point of view of a theory of complexity, there is simply no place for such a strict dichotomy: “when human beings adapt their language resources to new contexts, they change a language” (Massip-Bonet 2013: 48).

6 Some theoretical Prerequisites for the Integrated Study …

97

Cases like that of the Kaluli language do not seem as exceptional as one might think. Sometimes reference has been made to languages of different genetic origins which coalesce in a stronger or weaker way. From the first example of such Sprachbund described in detail, that of the Balkan languages, to B. L. Whorf’s proposal of a Standard Average European, nowadays the object of much interesting research (see Haspelmath 1998, 2001); the Beringian mesh as defined by Fortescue (1998), the Amazonian linguistic area, with interesting cases of development of time-based tenses in languages which previously had none (Aikhenvald 2003). This situation can lead to language structures having to be described simultaneously in both genetic and areal terms (Fagua Rincón and Seifart 2010). Similar instances can be identified in the linguistic history of the Australian languages and many others (Aikhenvald and Dixon 2001). A change can affect whole categories or (sub)systems, as in the case of Bosavi Kaluli or the Amazonian languages discussed by Aikhenvald (2003). The Old Germanic languages were affected by an all-out change leading to a completely new verbal system expressing time, time relations and time sequences. The ultimate reason lies probably in the ideological and cultural upheaval caused by the Roman Empire and posterior Christianization. The linguistic confluence of the Western European languages, termed STANDARD AVERAGE EUROPEAN (SAE) by Benjamin L. Whorf, is the object of serious study nowadays (Haspelmath 1998, 2001), and a common history of the European languages has been proposed (Bernárdez 2005).

6.2.2.3

Macrochange Inside One Language

It is not only in cases of language contact that macrochanges can take place. In certain conditions, basically of a socio-cultural character, changes can take place in several systems and/or subsystems of a single language at approximately the same time. My hypothesis is that such changes can be triggered by a specific need in that language’s speakers’ cultural realm that can be driven itself by ecological changes, new life conditions, and the like. That linguistic changes can be due to such types of non-linguistic change is fairly obvious and examples abound. One of the dialectal differences separating Eastern and Western Greenlandic includes the opposite meanings of some direction and space terms: The nomenclature for cardinal directions is the same on both sides of Greenland, but rather than being absolute, it is based on geographical landmarks such as the riverine system (upriver versus downriver), prevailing winds, and one’s position on the coast (up versus down). Since it is likely for rivers to flow and for winds to blow in opposite directions on the two sides of Greenland, this means that what is north for a western Eskimo is south for an easterner, and vice versa. (Heine 1997: 51)

For instance, avunga means ‘to the north’ (from the point of view of the speaker in Western Greenland), ‘to the south’ in Eastern Greenland, and so forth (Bjørnum 2003). Ecological, geographic conditions can have a direct, clear impact on a language as in this case, that by no means can be considered as an isolated case.

98

E. Bernárdez

If the ecological conditions change, change can be triggered, too, as witnessed by the recent need, also experienced in Greenlandic, to develop terms for realities born out of the Global Climate Change (Ejsing 2004). Socio-political change can also trigger change, e.g. those undergone by the Polish words for ‘homeland’ or ‘German’ (Bartmi´nski and Zinken 2009, Chaps. 13 and 14).

6.2.2.4

Macrochange and Complexity Theory

The situation outlined in the previous section is just to be expected, if we keep in mind the interrelation of all aspects of the human being that are directly but also indirectly related to language. A view that is widely shared nowadays, suffice it to refer to Enfield (2014), Enfield and Sidnell (2014), Weigand (2010), Fusaroli and Tylén (2012) or the papers in Enfield (ed., 2002), Massip-Bonet and Bastardas-Boada (eds., 2013) or Sharifian (ed., 2015). On the other hand, the permeability of the systems and subsystems of language is made evident in recent publications as Barðdal and Chelliah (eds., 2009), where semantics, pragmatics and discourse are analysed as triggers of change in the casesystems of a number of languages, or Schouwstra and de Swart (2014: 435) who show that word order, usually understood as a basically syntactic phenomenon, can best be explained in semantic terms: “the semantic origins of word order should not be overlooked in the debate on language evolution, and semantic organisational principles should be seen as the precursors of syntactic rules”. Change, therefore, need not take place in the same subsystem: a change in a metastable system can lead to, or be identified in a different system, which is then reorganized. The situation outlined is the norm in complex open systems, and very significantly also—and especially—biological systems, where changes ultimately triggered by the environment can affect several systems and a change within one system may extend to others and percolate throughout the whole organism—or species. In a natural complex system as language is, it can equally be expected for changes to affect one or more systems at the same time and percolate through the whole language. It can be said that macrochange, in the preliminary form sketched above, is a consequence of the nature of language as a complex, natural system. Goldstein and Janssen (2005: 428) write the following: If each person is to be modelled as a conceptual network, then a social group is to be modelled as a network of networks. (…) Communicating is not simply transmitting individual concepts. Communication involves aligning the conceptual systems of agents. One implication of this alignment process is that as concepts migrate across people, they will be systematically altered to fit their owners’ conceptual network.

These words can serve as a background for much of the theoretical approach of this chapter.

6 Some theoretical Prerequisites for the Integrated Study …

99

6.3 A Case Study: Some Changes in Icelandic 6.3.1 Some Idiosyncratic Features of Icelandic Contemporary Icelandic shows a number of linguistic features that are not shared with the other Nordic and/or Germanic languages, or only restrictedly so. Most of these features have developed in Modern and Contemporary Icelandic, ex novo or on the basis of some similar, but not identical, constructions of the old language of the sagas. All of them seem to intend to show—frequently with extreme precision—details on the source of our knowledge about a fact, its exact spatial and temporal location, especially in relation to others, the degree of responsibility of agents for their actions, as well as the origin of ideas, opinions, etc., together with the great attention paid to types and degrees of knowledge. As it is not possible to analyse all such details in any detail, a few rather general comments will have to suffice. Moreover, research on this macrochange is still at a preliminary level, and the conclusions are then to be taken as hypothetical. The phenomena investigated are the following.

6.3.1.1

Precise Expression of Source of Knowledge

Icelandic makes a distinctive use of verbs as sjá (see) and heyra (hear) to mark whether the information is first-hand (sjá) or has been provided by a secondary source (heyra) (Bernárdez 2013). The grammaticalization of heyra as a marker of secondhand evidentiality is so advanced that this verb can be used to refer to information acquired through reading. This construction is also present in Old Icelandic and in other Nordic languages, such as Danish. Peculiar to Icelandic is its being a part of the wider domain analysed in these pages.

6.3.1.2

Precise Expression of Temporal Relations Holding Between Two Events

The development of a number of new ‘periphrastic conjugations’ in Icelandic allows the precise identification of the temporal relations holding between two events; cfr.: (1) Ég var að borða morgunmatinn þegar hún kom I was eating the-breakfast when she came1 (i.e., eating A simultaneous with her arrival B; emphasis on the temporal coincidence) (2) Ég var búinn að borða morgunmatinn þegar hún kom I had finished to eat the-breakfast when she came (A finished before B; emphasis on the end of A) 1 The translations are obviously not idiomatic in English: their only purpose is to make possible the

literal understanding of the Icelandic examples.

100

E. Bernárdez

(3) Ég hafði borðað morgunmatinn þegar hún kom I had eaten the-breakfast when whe came (A finished before B; no emphasis) (4) Ég borðaði morgunmatinn þegar hún kom I ate the-breakfast when she came (A follows B; in another interpretation, both are simultaneous as in (1), without emphasis on temporal coincidence). Neither in Old Icelandic nor in the other Germanic or Nordic language does such a full set of grammaticalized constructions exist which may allow for economic and precise marking of the time relations, i.e. of knowing what comes first and what is second. The same meanings can be conveyed by different means, as in the English glosses above, but they do not form a full-fledged paradigm as is the case of Icelandic.

6.3.1.3

Precise Expression of Spatial Relations

Icelandic very frequently reinforces the expression of location or movement by adding the precise identification of e.g. the (socially institutionalized, not only merely geographical) cardinal directions: (5) Hann fór vestur til Ameríku he travelled east to America. If any doubt could exist about the destination, the cardinal direction helps to solve it.

6.3.1.4

Precise Expression of the Degree of Responsibility of the Agent

There is an ongoing discussion on the ‘non nominative subjects’ in Icelandic, both in sentences as (6–7) where there is no noun in the nominative case which may be seen as the subject of the verb, as both nouns stand in the accusative (6) or as in (7), where the only noun is in the dative case, and in sentences as (9–10) where a noun in the accusative (8) or the dative (9–10) is taken to be the subject of the verb (on the basis of a set of basically syntactic conditions). (6) Þig (acc) hefur dreymt Svein (acc) to-you has dreamt Sveinn (7) Hér lýkur sögunni (dat) here closes to-the-story (8) Mig (acc) langar að vita það To-me wishes to know that (9) Mér sýnist, að… to-me seems, that… (10) Það þykkir mér að… it seems to-me (= I think) that…

6 Some theoretical Prerequisites for the Integrated Study …

101

I cannot enter into the discussion on the status of such ‘non-nominative subjects’ (see Thráinsson 2007). What seems clear is that in cases like these a process of agent demotion (or deagentivisation, Bernárdez 1997) is identifiable. The result of these constructions is that the person involved is not portrayed as directly responsible for the action or event. This type of construction has existed since Old Norse times (Barðdal and Eythórsson 2003) but it is nowadays still productive (Barðdal 2003). A very special case is a construction I have termed ‘unconscious, irresponsible’, that marks the participant as neither responsible or in control of the event or action, nor even conscious of it. Kress (1982) described it as having a ‘magical agent’. The following examples (11–13) show that an agent is non voluntarily ‘driven’ to look or think something, or to move out of a place; it is built as a kind of impersonal passive with the auxiliary verða (become) plus the neuter form of the past participle; the participant appears in the dative (Bernárdez 2007a): (11) Mér varð litið í kringum mig I looked [to-me became looked] around me (12) Mér varð hugsað til þessara orða I thought [to-me became thought] these words (13) …verður ömmu minni geingið útúr kotinu að líta á kálið… (Laxness) … my grandma went out [to-grandma mine became gone out] of the cottage to look at the vegetables…

6.3.1.5

Precise Expression of Epistemological States: Knowing, Thinking, etc.

Similarly to what has been shown in (6.3.1.3). Icelandic tries to be extremely precise when expressing the degree of control and direct involvement of the participants in their own thinking, knowing, imagining, suspecting, etc. A number of verbs covering these areas of epistemic meaning exist, together with constructions which show the state (of thinking etc.) as being the result of the participant’s conscious, controlled efforts, or of a completely alien agent, event or process. For instance, dreyma (dream) with a nominative subject, as in (14) expresses a controlled process, due to the volitional mental activity of the participant (‘day-dreaming’); if the participant is in the accusative (15), it means that a dream took place while s/he was sleeping, thus no control or volition exists: (14) ég dreymdi að… I dreamt that… (I was fully awake) (15) mig dreymdi að… to-me dreamt that… (I was sleeping). The number and degree of grammaticalization of these structures, as well as their high frequency of use, is something new and idiosyncratic in contemporary Icelandic, although it has deep roots in both Germanic and Old Norse.

102

6.3.1.6

E. Bernárdez

Precise Expression of Forms and Types of Possession

Icelandic expresses possession in three different ways (apart from possessives, genitives, and the like). The verb hafa (have) ‘to be in the possession of something but without unrestricted rights to it’; eiga (own) ‘to have unrestricted rights of possession to something; have a special, intimate relation with someone or something’; and vera með (‘to be with’), to have something integrated in one self, especially although not necessarily the participant’s body. Obviously, this compulsory distinction helps make the relations between objects and owners, or between persons, as clear as possible.

6.3.2 The Unitary Character of These Constructions and Their Ecological Causes All these constructions have something important in common: they make as clear as possible a number of relations, events, etc. They affect such systems as the lexicon, syntax, and morphology. The frequency of their use points to a sensible pragmatic effect. The following is a tentative attempt at a possible explanation: The ecological system where the Icelanders used to live imposed a number of problems, among them: (a) Isolation during the winter months, with the consequence of serious ignorance of other people’s actions, plans, etc. (b) Lack of visibility in the darkness of winter, during blizzards, etc., leading to the impossibility of direct perceptual access to anything that may happen or is happening. (c) Personal relations, including family or property feuds with members of neighbouring families, a situation made even worse by isolation: you cannot know what such other persons may be planning to do against you, for instance. (d) Invisible but ‘real’ presence and activity of the so called ‘hidden people’, i.e. elves, goblins, and the like: only propitiatory rituals, still partly alive in some rural areas, could help people against possible evil doings. Of course, lack of visibility plays a crucial, although ancillary role: the main point is the belief in the ‘real’ existence of those invisible beings. These ecological factors result in a number of socially-sanctioned activities intended to minimize the ensuing risks. For instance, the belief in elves (d) triggers the ritual of living food for them on the window-sills; it is a purely cultural response. On the other hand, (a, b, c) have an impact on language, not only on cultural rituals and activities: to avoid the danger posed by such negative conditions, it is necessary to be on the alert all the time and at any time. To be able to anticipate what can happen due to the weather conditions (b) or to the possible activities of others, independently of the degree of relation, confidence, friendship or enmity (a, c), it is convenient to be as well informed as possible on as many circumstances as could be expected. For instance, whether the source of our knowledge on the state of the weather in a particular place is trustworthy or not, what is the precise meaning of someone’s words passed on by someone else, etc. One must therefore distrust everything and everybody and gather all the complementary information that may be available.

6 Some theoretical Prerequisites for the Integrated Study …

103

6.4 Conclusions An open, complex system like language, can be seen as composed of a number of parts, usually called ‘systems’ or ‘subsystems’. These however exist only inasmuch as they are integrated in a more complex object that can be termed ‘language’. Language, in turn, is integrated in a still larger entity which encompasses human cognition (understood in the terms sketched above) and human culture + history. The metastable states present in the different parts of the whole and their integrant parts can be altered by changes in any of the other parts; these changes can affect a single component or element, a large set of elements, or a whole ‘system of systems’, as in macrochange, similarly to the evolutionary processes that take place in macroevolution. A case has been shown which may be a good example of such integration and macrochange: the systems developed in contemporary Icelandic in order to avoid ambiguity or vagueness in information which may be difficult to obtain due to the (traditional) ecological conditions in which the speakers of the language had to live. Environment, speakers and their cognitive and cultural systems, as well as language, its functioning and structures, act therefore in an integrated way, as a single complex system. Of course, a lot of work remains to be done in this field, also in the deeper and more complete analysis of the Icelandic features reviewed here.

References Aikhenvald, A. Y. (2003). Mechanisms of change in areal diffusion: New morphology and language contact. Journal of Linguistics, 39, 1–29. Aikhenvald, A. Y., & Dixon, R. M. W. (Eds.). (2001). Areal diffusion and genetic inheritance. Oxford: Oxford U.P. Barðdal, J. (2003). Case and argument structure of novel verbs of communication in Icelandic. In L. O. Delsing, C. Falk, G. Josefsson, & H. Á. Sigurðsson (Eds.), Grammatik i Fokus (Vol. 2, pp. 25–35). Lund: Lund Universitet. Barðdal, J., & Chelliah, S. L. (Eds.). (2009). The role of semantic, pragmatic, and discourse factors in the development of case. Amsterdam: Benjamins. Barðdal, J., & Eythórsson, T. (2003). The change that never happened: The story of oblique subjects. Journal of Linguistics, 39, 439–472. Bartmi´nski, J., & Zinken, J. (Eds.). (2009). Aspects of cognitive ethnolinguistics. (Advances in cognitive linguistics). London: Equinox Publishing Ltd. Bastardas-Boada, A. (2013). In A. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 15–34). https://doi.org/10.1007/978-3642-32817-6_3. Bernárdez, E. (1997). A partial synergetic model of deagentivisation. Journal of Quantitative Linguistics, 4, 53–66. Bernárdez, E. (2005). Toward a common history of the Germanic and European languages in the Middle Ages. REVISTA SELIM Journal of the Spanish Society for Medieval English Language and Literature, 12, 5–32. Bernárdez, E. (2007a). The unconscious, irresponsible construction in modern Icelandic. In C. S. Butler, R. Hidalgo Downing & J. Lavid (Eds.), Functional perspectives on grammar and discourse (pp. 149–164). Amsterdam: Benjamins. https://doi.org/10.1075/slcs.85.09ber.

104

E. Bernárdez

Bernárdez, E. (2007b). Synergy in the construction of meaning. In M. Fabiszak (Ed.), Language and meaning: Cognitive and functional perspectives (pp. 15–37). Frankfurt/Main: Peter Lang. Bernárdez, E. (2008). Collective cognition and individual activity: Variation, language, and culture. In R. Dirven, R. Frank, E. Bernárdez, & T. Ziemke (Eds.), Body, language, and mind (Vol. 2, pp. 137–176). Berlin: Mouton de Gruyter. Bernárdez, E. (2013). Evidentiality and the Epistemic Use of the Icelandic Verbs Sjá and Heyra. A cultural linguistic view. In A. Głaz, D. S. Danaher, & P. Łozowski (Eds.), The linguistic worldview. Ethnolinguistics, cognition, and culture (pp. 415–441). London: Versita. Bjørnum, S. (2003). Grønlandsk Grammatik. Nuuk: Atuagkat. Ejsing, J. (2004, November 20). Inuitter mangler ord for nye naturfænomener. Berlingske Tidende. Retrived from https://www.b.dk/nationalt/inuitter-mangler-ord-for-nye-naturfaenomener. Enfield, N. J. (Ed.). (2002). Ethnosyntax. Explorations in grammar and culture. Oxford: Oxford U.P. Enfield, N. J. (2003). Linguistic epidemiology: Semantics and grammar of language contact in mainland Southeast Asia. London: Routledge Curzon. Enfield, N. J. (2014). Natural causes of language. Frames, biases, and cultural transmission. Berlin: Language Science Press. Enfield, N. J., & Sidnell, J. (2014). Language presupposes an enchronic infrastructure for social interaction. In D. Dor, C. Knight, & J. Lewis (Eds.), The social origins of language (pp. 92–104). Oxford: Oxford University Press. Fagua Rincón, D., & Seifart, F. (2010). Aspectos morfosintácticos del ocaina: rasgos genéticos (familia witoto) e influencias areales. Mundo Amazónico, 1, 215–241. Fortescue, M. (1998). Language relations across the Bering Strait: Reappraising the archaeological and linguistic evidence. London: Cassel. Fusaroli, R., & Tylén, K. (2012). Carving language for social coordination. A dynamic approach. Interaction Studies, 13(1), 103–124. Garrod, S., & Pickering, M. J. (2004). Why is conversation so easy? TRENDS in Cognitive Science, 8(1), 8–11. Goldstone, R. L., & Janssen, M. A. (2005). Computational models of collective behaviour. TRENDS in Cognitive Sciences, 9(9), 424–430. Haspelmath, M. (1998). How young is standard average European? Language Sciences, 20(3), 271–287. Haspelmath, M. (2001). The European linguistic area: Standard Average European. In M. Haspelmath, E. König, W. Oesterreicher, & W. Raible (Eds.), Language typology and language universals (pp. 1492–1510). Berlin: de Gruyter. Haspelmath, M. (2004). On directionality in language change with particular reference to grammaticalization. In O. Fischer, M. Norde, & H. Perridon (Eds.), Up and down the cline—the nature of grammaticalization (pp. 17–44). Amsterdam: Benjamins. Heine, B. (1997). Cognitive foundations of grammar. New York: Oxford U.P. Heine, B., & Kuteva, T. (2003). On contact-induced grammaticalization. Studies in Language, 27(3), 529–572. Krawczak, K. (2007). Meaning as an epiphenomenon of cognition, social interaction and intercognition. In M. Fabiszak (Ed.), Language and meaning. Cognitive and functional perspectives (pp. 187–198). Frankfurt/Main: Peter Lang. Kress, B. (1982). Isländische Grammatik. Leipzig: Enzyklopädie. Kutschera, U., & Niklas, K. J. (2004). The modern theory of biological evolution: An expanded synthesis. Naturwissenschaften, 91, 255–276. Massip-Bonet, À. (2013). Language as a complex adaptive process. Towards an Integrative Linguistics. In Á. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity Perspectives on Language, Communication and Society (pp. 35–60). Berlin: Springer. Massip-Bonet, À., & Bastardas-Boada, A. (Eds.). (2013). Complexity Perspectives on Language, Communication and Society. Berlin: Springer.

6 Some theoretical Prerequisites for the Integrated Study …

105

Schieffelin, B. (2002). Marking time: The dichotomizing discourse of multiple temporalities. Current Anthropology, 43, 5–17. Schouwstra, M., & de Swart, H. (2014). The semantic origins of word order. Cognition, 131, 431–436. Severi, C. (2004). Capturing Imagination: A cognitive approach to cultural complexity. The Journal of the Royal Anthropological Institute, 10(4), 815–838. Sharifian, F. (Ed.). (2015). The Routledge Handbook of Language and Culture. London: Routledge. Thráinsson, H. (2007). The Syntax of Icelandic. Cambridge: Cambridge U.P. Weigand, E. (2010). Linguists and their speakers. Language Sciences, 32, 536–544. Welmers, W. E. (1973). African Language Structures. Berkeley etc.: University of California Press. Williams, G. (Ed.). (2005). Language, brain, culture. Linguistics and the Human Sciences, 1(2), 147. Wilson, R. A. (2004). Boundaries of the Mind. The individual in the fragile sciences. Cambridge: Cambridge U.P.

Chapter 7

The Impact of Social Reputation in Language Evolution Gemma Bel-Enguix

Abstract Language evolution has been widely tackled by means of mathematical and computational models. To approach the problem, several proposals have been introduced based on simulation, some of them taking the perspective of language emerging, by means of different variants of the naming game (Steels in The synthetic modeling of language origins, 1:1–34, 2000). The present work is placed into this framework, addressing the impact of social reputation in the emergence and evolution of languages from an agent-based point of view. Some simulations are performed with populations of 100, 300 and 500 agents, taking into account different structures of the society based on the distribution of reputation. The ultimate goal of the paper is to understand whether the way human communities are organized has a prominent role in how the language emerges and the change in communication systems over the time.

7.1 Introduction This work aims to be a preliminary contribution to the study of coevolution of language and social structures from the perspective of complex systems and evolutionary biology, by means of simulation and mathematical analysis of results. The paper focuses on the role that social factors—especially reputation—play in the arising and diffusion of innovation in language acquisition and interaction. During the last decade, studies of language emergence and evolution have undergone a deep transformation thanks to new interdisciplinary approaches introduced from artificial intelligence, physics and biology. From the new perspective, language evolution has been tackled by means of mathematical and computational models. Currently, a mainstream line of research in diachrony is based on the understanding

G. Bel-Enguix (B) Grupo de Ingeniería Lingüística, Instituto de Ingeniería, Universidad Nacional Autónoma de México, Ciudad de México, Mexico e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_7

107

108

G. Bel-Enguix

of language as a complex adaptive system (Steels 2000) and an evolutionary system (Croft 2000; Brighton et al. 2005). The aforementioned authors stress the necessity of computational support and simulation to fully understand the mechanisms underlying the dynamics of language evolution. Several aspects of this research area, like language change, interaction, convergence and splitting, can take advantage of these simple models. This fact is crucial in a branch of linguistics in which, at some stages, the main problem is the lack of data. Among the proposals that have been introduced, one of the simplest experiments based on simulation is the one by Baronchelli et al. (2006). Baronchelli’s model is conceived as a variant of the naming game (Steels 1997) and implies a number of agents that have to agree in naming an object with no pre-established protocol. Two chief features of Baronchellis’s model are that: (a) in the beginning, the agents have no linguistic knowledge at all, and (b) when two agents agree in a word, they delete everything else they have stored. Such characteristics clearly imply some cognitive limitations in the design but, conversely, this can be seen as an advantage because of the simplicity of the interaction algorithms. The results obtained with the experiment show that language change and evolution follow some patterns that can be formulated with the rules of mathematics and physics. After Baronchelli’s work, Brigatti (2008) introduced the concept of reputation, pointing out the possibility of analyzing how and how much such parameter affects language evolution. We understand reputation as a social status that provides credibility and capacity to communicate with other individuals. This work approaches some of the concepts that Baronchelli introduced in the paper, like the number of (max) words, or the time for the convergence. We are also inspired by the idea of reputation that Brigatti used, even though its implementation does not work in the same way in our model. Taking these components, we study the computational behavior of agents under some initial configurations of the society in order to create a language.

7.2 Bases of the Model and Configuration of the Experiment Our experiment deals with the process of linguistic convergence in societies whose members have to negotiate in order to name an object. This is a version of the naming game by Steels (1997). The agents of the community have very restricted cognitive capacities, and when they learn a name for an object, they forget everything else they had in the memory related to that object. The notation is as follows: N is the number of agents of the game, T is the time measured in number of interactions, W is the number of words, Wd stands for the number of different words and S for the success of the interaction. The aspects that are studied with the experiment are:

7 The Impact of Social Reputation in Language Evolution

109

• Tconv: the number of interactions needed for the agents to agree in naming an object with the same word. • Wmax: the maximum number of words in the game. At the end of the computation, the number of words W  N. • Wdmax: The maximum number of different words. At the end of the game, the number of Wd  1. • Tmax: the time T where Wmax occurs. • SR: the success rate. The Protocol 1, which is applied to the simple games without reputation, is the following: • The agents, that are empty in the beginning, have to agree in naming an object. • In each step, speaker (S) and hearer (E) are selected randomly. • The speaker randomly takes a word from its inventory and sends it to the hearer. If the inventory is empty, it invents a word. • If the hearer has the word the speaker sends, then the interaction is satisfactory. Both agents delete all and keep only the item they share. • If the hearer does not have the word in its inventory the communication is unsuccessful. Both agents keep the word. • The game ends when each agent has only one word which is the same, this is, when the whole system shares a meaning for an object. However, our main interest in this paper is studying what the impact of social structures in language evolution is. To approach the problem, we consider that these societies have a component called Reputation. As defined before, reputation affects the capacity of an agent to interact with others, and this leads to modify the protocol of communication of a society. Protocol 2 is the basic algorithm for communication between agents in a community that has social groups, or different reputations. It is as follows: • The agents, that have nothing in the beginning, have to agree in naming an object. • In each step, speaker (S) and hearer (E) are selected randomly. • The speaker, with reputation RS, transmits the selected word to the hearer, characterized by the reputation RE. • If the hearer’s inventory contains such a word, the communication is a success. The two agents update their inventories to keep only the word involved in the interaction. The speaker’s reputation increases by one. • Otherwise: – If RS > RE, the hearer adds the new word to its inventory and the speaker does nothing. The speaker’s reputation decreases by one. Success is 0.5. – If RS < RE, the communication is a failure. The speaker’s reputation decreases by one. Success is 0. • The game ends when each agent has only one word which is the same, this is, when the whole system shares a meaning for an object.

110

G. Bel-Enguix

As it can be seen in the protocol, communication between two agents Speaker (S) and Hearer (E) is allowed even if they have two different values of Reputation (R). If the communication is successful—if S and H share the word they exchange - the parameter of reputation does not play any role, but if communication fails because a word W sent by S in not known by E, then reputation becomes a key parameter in the development of the linguistic evolution of the society. Different variants of the protocol have been developed for the same communicative situation. In Bel-Enguix (2010) the model only allows communication if RS ≥ RE. The result is that, with some configurations of the society, being the group H a minority, convergence does not happen. The variant proposed in this research is weaker, allowing communication if agents have some in common and, therefore, making possible to reach an agreement in 100% of the cases. Within this framework, we have implemented simulations that allow us to study with some attention the impact of reputation in language change. More precisely, the work tests the behavior of the agents starting from different distributions of initial reputation, this is, different configurations of the society, at the beginning of the computation. The same experiments have been performed from four different initial states, which are the following: 1. There is no reputation at all in the society and it is not going to be created. 2. Every agent has the same reputation at the beginning of the game. 3. Reputation is configured following a symmetric distribution with different coefficients. 4. There is an asymmetric distribution of the reputation among the agents. The four different initial states deserve some comments in order to settle the principles of the experiment.

7.2.1 Reputation Does not Exist We include this case to compare the results of the game with and without reputation. We want to know whether the existence, or not, of reputation improves the convergence time and the other parameters of the game. To do that, we have to compare the results obtained with Protocol 1 to the ones of the other scenarios (Protocol 2).

7.2.2 Every Agent Has the Same Reputation For the experiment, we consider societies that we could label as egalitarian, where every agent has the same reputation at the beginning of the game, but this parameter is going to change during the development of the computation. Even though the initial value of Reputation  0 in this simulation, Protocol 2 is the one applied, because R exists.

7 The Impact of Social Reputation in Language Evolution

111

7.2.3 Symmetric Distribution of the Reputation In this model, reputation is distributed among agents following a Gaussian-shaped distribution, where is the difference of reputation between two contiguous agents. In the model, two agents of every society have the same reputation in the initial configuration, in a way that this value R is distributed between individuals, in an unequal society that lacks powerful social groups.

7.2.4 Asymmetric Distribution of the Reputation Finally, we consider populations that are divided into two different social groups, named H (High reputation) and L (Low reputation), each one with a given reputation (RH and RL). The difference of reputation between H and L (RH and RL) is denoted by δ. There are many possible initial configurations of asymmetrically distributed populations. We consider societies distributed in three different ways, giving rise to three scenarios: • Scenario 1: The population is divided into two social groups, where the smallest one, around 20%, has a higher reputation that the rest. This can be a very plausible situation in the distribution of a society, where a small group has the power and some privileges over the rest. • Scenario 2: There exist two classes in a society with different reputation distributed in two equal groups, taking each one of them the 50% of the population. • Scenario 3: The society is divided into two classes, being the group with lower reputation the 20%. This may illustrate societies with segregated minorities.

7.3 Configuration of the Experiment and Results We are performing tests with communities of 100, 300 and 500 agents. The results are obtained taking an average over 100 runs. As explained above, five situations are considered: 1. There exists a society without any reputation (No R). 2. Every agent has the same reputation at the beginning of the game (ER). R  0 3. Members of the community have symmetrically distributed reputation with σ  5 (abbreviated as Symσ5). 4. Asymmetrically distributed reputation with δ  10, being RH  5, RL  − 5 and H  20 (abbreviated AsymH20δ10). 5. Asymmetrically distributed reputation with δ  10, being RH  5, RL  − 5 and H  50 (abbreviated AsymH50δ10).

112

G. Bel-Enguix

6. Asymmetrically distributed reputation with δ  10, being RH  5, RL  − 5 and H  80 (abbreviated AsymH80δ10). The experiment has the following outcomes (Tables 7.1, 7.2 and 7.3): As for the convergence time, Tconv, a comparative graph is shown in Fig. 7.1. The most prominent result for this parameter is that the existence of Reputation highly increases the convergence time. For populations with equal R or distributions between only two groups, the convergence time is similar, with an advantage for the model where a small part of the society has higher R. However, the distribution or different values of R along the population makes the agreement very hard. Figure 7.2 shows the results for Wmax, the maximum number of words that are in the system. The obtained data follows the rule Wmax ≈ N (4 + (N − 100)/100). However, for any number of agents, if reputation exists and it has a non-equal distri-

Table 7.1 Results of the experiment with 100 agents. No R stands for societies without reputation 100 agents Tconv

SR

Wmax

Tmax

Wdmax

t/tmax

No R ER  0 Symσ5

2420 4041 6376

0.42 0.45 0.61

398 384 311

635 1117 833

49 97 67

3.81 3.61 7.65

AsymH20δ10

3992

0.49

321

920

83

4.33

AsymH50δ10

4235

0.48

353

1031

81

4.10

AsymH80δ10

4019

0.47

375

1105

88

3.63

ER  0 represents societies with equal reputation between their members In this case R  0 but the protocol of communication takes into account the parameter, so R varies Symσ5 stands for societies in which two members have the same initial reputation. Initial indices of R have a difference of 5 AsymH20δ10 is the population with two different groups The 20% has the Higher Reputation and the rest the Lower Reputation Difference between HR and LR is 10: HR  5 and LR  − 5 The same distribution applies for AsymH50δ10 and AsymH80δ10, except the fact that, for the former, the population with HR is 50% and for the latter is 80% Table 7.2 The results are shown in this table for a population of 300 agents 300 agents Tconv

SR

Wmax

Tmax

Wdmax

t/tmax

No R

10,076

0.34

ER  0

17,131

0.42

1871

3098

149

3.25

1700

5329

296

Symσ5

32,101

3.21

0.66

1323

3660

204

8.77

AsymH20δ10

13,922

0.82

1468

3759

195

3.70

AsymH50δ10

14,347

0.81

1477

3978

198

3.60

AsymH80δ10

13,731

0.81

1558

4197

194

3.27

7 The Impact of Social Reputation in Language Evolution

113

Table 7.3 Table of results for a population of 500 agents 500 agents Tconv

SR

Wmax

No R

19,990

0.31

3894

Tmax 6537

250

Wdmax

t/tmax 3.05

ER  0

32,059

0.40

3461

11,076

495

2.89

Symσ5

70,990

0.68

2661

7657

342

9.27

AsymH20δ10

27,297

0.81

2961

7752

326

3.53

AsymH50δ10

28,621

0.81

2983

8230

331

3.47

AsymH80δ10

28,208

0.80

3184

8776

324

3.21

Tconv

Fig. 7.1 Comparative results of the convergence time

bution, the value is under this measure. Notoriously, populations without reputation need to reach a higher number of words in order to converge. In contrast with Wmax, the parameter Wdmax (Fig. 7.3), that shows the maximum number of different words, reaches much lower ratings, as it can be seen in Fig. 7.3. The values of Wdmax, clearly go from N/2 to N, being for NoR ≈ N/2 and for ER  0 ≈ N in every set of agents. The results obtained for populations with nonequal reputation are between N/2 and N, with very similar results with different configurations of the groups H and L. The feature Tmax, as seen in Fig. 7.4, follows a general distribution similar to the one of Wdmax, where the worst performance is found in ER  0 and the best in Syms5 and asymmetrically distributed populations with H  20%. In general, the results show how the number of agents is not especially important in the performance of the system, but the numbers increase in a proportional relation according to the size of the population. There is another data that has to be taken into account, the value Tconv/Tmax, which explains how sharp the graphic of the convergence is. He higher the number Tconv/Tmax is, the sooner the system gets the maximum number of words compared

114

G. Bel-Enguix 4500 4000 3500 3000 2500 2000 1500 1000 500 0

Wmax

100 Agents 300 Agents 500 Agents

Fig. 7.2 Comparative results for the maximum number of words that the system reaches

600

Wdmax

500 400 300 200 100

100 Agents 300 Agents 500 Agents

0

Fig. 7.3 Number of different words

with the total time of computation, although the differences are not extremely representative. Additionally, this value is higher with fewer agents, which means that, against intuition, in this situation, the time of negotiation to reach a final agreement after Tmax is harder. Figure 7.5 illustrates this. However, the numbers of Symσ5 are unexpected because they are clearly higher than the others, while its proportion between the number of agents and the value Tconv/Tmax is inverse. In such configurations, once the maximum number of words has been reached, the relative performance is better with a smaller number of agents. There is probably a threshold in the number of agents in which negotiation to reach agreement becomes much more complicated.

7 The Impact of Social Reputation in Language Evolution

115

Tmax

Fig. 7.4 Time when the system reaches the maximum number of words

T/Tmax

Fig. 7.5 Time when the system reaches the maximum number of words

7.4 Conclusions The paper presents a preliminary approach, built over the model introduced by Baronchelli (2006), whose agents have been designed with some cognitive weakness. For example, they are not able to retain the information in the memory once they learn a word, which makes them totally monolingual and unable to know two different words for an object. However, keeping the simplicity has been one of the main requirements of the experiment, which wants to show if the behavior of a community with social structures is different from one without them in what refers to language agreements. It seems clear that social rules have an important impact in the linguistic behavior of individuals, as this simulation shows. It is true that the design of the protocols that

116

G. Bel-Enguix

regulate interaction between agents determines the final results. Notwithstanding, this also happens in real societies, which have very different rules for participation in dialogues and turn-taking. The protocol designed for this research is simple, and allows two people to talk if they have something in common, no matter which social group they belong to. The parallelism between this type of experiments and real life is risky. However, our outcome gives an idea of the main fact, that reputation and social groups are very important in the generation and spreading of language novelties. We have compared several initial configurations, including egalitarian societies and groups with a small group that dominates the others. The results seem to show that, in terms of spreading of words and agreement, equality is not the best. The worst results are for the situations in which every agent has the same reputation and for the ones in which the parameter is very distributed among the population. Taking this research as a starting point, a new step would be to decide if there is a coevolution between language and social structures. Further studies should show whether or not such patterns of coevolution exist. Other issues to investigate would be the possibility that social factors have a bias toward language acquisition, and the impact of the social environment in the rate of linguistic innovation of an individual. Finally, developing this line of research can lead to some sophisticated theories and simulations that could be helpful to predict future configurations of language.

References Baronchelli, A., Felici, M., Caglioti, E., Loreto, V., & Steels, L. (2006). Sharp transition towards shared vocabularies in multi-agent systems. Journal of Stat. Mech. Bel-Enguix, G. (2010). A Multi-agent Model for Simulating the Impact of Social Structure in Linguistic convergence, Proceedings of ICAART 2010, 2nd International Conference on Agents and Artificial Intelligence, II: 367–372. Brigatti, E. (2008). Consequence of reputation in an open-ended naming game. Physical Review E, 78, 1–1111. Brighton, H., Smith, K., & Kirby, S. (2005). Language as an evolutionary system. Physics of Life Reviews, 2, 177–226. Croft, W. (2000). Explaining language change. Harlow (Essex), Longman: An evolutionary approach. Smith, K., Kirby, S., & Brighton, H. (2003). Iterated learning: a framework for the emergence of language. Artificial Life, 9(4), 371–386. Steels, L. (1997). The synthetic modeling of language origins. Evolution of Communication, 1, 1–34. Steels, L. (2000). Language as a complex adaptive system. LNCS, 1917, 17–26.

Part III

Sociolinguistics

Chapter 8

‘Restricted’ and ‘General’ Complexity Perspectives on Social Bilingualisation and Language Shift Processes Albert Bastardas-Boada

Abstract Historical processes exert an influence on the current state and evolution of situations of language contact, brought to bear from different domains: the economic and the political, the ideological and group identities, geo-demographics, and the habits of inter-group use, among others. Clearly, this kind of phenomenon requires study from a complexical and holistic perspective in order to accommodate the variety of factors that belong to different levels and that interrelate with one another in the evolving dynamic of human languaging. The need in my view is for the restricted and general complexity approaches to come to a meeting of the minds, and take steps toward a mutual integration based on the acceptance of the shortcomings of each approach, achieving progress through a non-contradictory complementarity of perspectives. It must be conceded that the practical and methodological applications of basic complexical ideas need to be developed much farther in order to apply them to specific research. At the same time, the limits of complex adaptive systems as computational strategies must be accepted in the pursuit of a better understanding of the dynamic and evolutionary processes typical of human beings. New tools for the conception, apprehension and treatment of the data will need to be devised to complement existing ones and to enable us to make headway toward practices that better fit complexical perspectives. It seems obvious that human complexics must be seen as multi-methodological, insofar as necessary combining quantitative-computation methodologies and more qualitative methodologies aimed at understanding the historical mental and emotional world of people.

A. Bastardas-Boada (B) Department of Catalan Philology and General Linguistics, Sociocomplexity—Complexity, Communication and Sociolinguistics Group, CUSC—Research Centre for Sociolinguistics and Communication, UBICS—Universitat de Barcelona Institute of Complex Systems, Universitat de Barcelona, Barcelona, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_8

119

120

A. Bastardas-Boada

8.1 A Personal Experience on Complexity Until some time ago, my personal approach to ‘complexity’ had not followed the same roads travelled by quantitative physicists, mathematicians, biologists and computational scientists exploiting the new possibilities of computer science. I was, and still am, more an adherent of the ‘complexité’ inspired primarily by Edgar Morin and supplemented by Elias’s (1982, 2000) contributions to figurational and processual sociology. This perspective was also consistent with the perspective of authors in other fields, such as biology (Maturana and Varela 2004) and theoretical physics itself (Prigogine and Stengers 1979, 1992; Bohm 1980; Capra 2002). I have laid out the foundations of my synthesis of what we might call ‘figurational complexity’ or ‘complex figurational sociology’ in a few works that have recently appeared in print, so I will forgo detailed explanation here (Bastardas-Boada 2013a, b, 2014, 2016a). In my case, Morin’s ‘complexité’ and its strong push for an ‘ecologisation’ of thought—an approach also advanced by scholars such as Gregory Bateson (1972) and Aracil (1982, 1983)—confirmed for me the merits of building on the basis of a holistic vision of reality, one that is nonetheless conscious of the parts, in order to grasp sociolinguistic events and phenomena more effectively. Indeed, this notion of ‘ecologisation’ was not new to sociolinguistics. It had been proposed earlier by Haugen (1972) and again later by Mackey (1979, 1980, 1994) but it did stand in need of further elaboration. Biological ecology helps us with its theoretical propositions and models (Margalef 1991; Allen and Hoekstra 2014), yet human ‘languaging’ (Maturana 2002) is clearly not a species and there was still a need to look beyond the initial analogies. This is what led me to postulate a ‘sociocognitive’ ecology for cases of language contact (Bastardas-Boada 1996, 2016a), an ecology based not on a simple transposition of ideas and concepts from biological ecology, but rather on the propositions of Edgar Morin, David Bohm, Fritjof Capra, Norbert Elias, The Gulbenkian Comission (Wallerstein 1996), and other scholars working in the sociocultural sciences. This is the intellectual climate in which I sought, primarily, to draw on the various contributions of the authors mentioned above, in order to build a complexical and dynamic perspective that could offer an account of the factors affecting human language behaviour and its historical evolution. My main guiding principles were: (a) the centrality of the brain/mind, (b) self-organisation, (c) emergence, (d) circular, retroactive and recursive causality (vs. linear causality), (e) the ecosystemic and holographic nature of reality, which implies not only that the part is in the whole, but that the whole is also ‘in’ the part, and (f) that existence is processual and dynamic (cf. Bastardas-Boada 1999, 2013a, 2016c). These principles underlied my proposal to adopt an ecological framework and bring sociocomplexity into the study of language contact. The result, in practical form, was a dynamic, multi-layered ‘orchestral’ picture that can embrace the distinct domains underpinning human language activity and its interrelationships in order to gain a much better grasp of the factors affecting language behaviour and its historical evolution (see Bastardas-Boada 1996, 2016a). Succinctly put in the typical parlance of sociology, my work has taken as central

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

121

human language behaviour in situations of personal and/or social cultural contact and then examined its interrelationships with the sociocultural factors that might co-influence its configuration and development. It must be conceded that the use of the terms ‘complex’ and ‘complexity’ in the vast majority of publications appearing in English—the most widespread language of science—corresponds much more to ‘restricted’ complexity than to the more ‘general’ perspective, as Morin (2005) calls them. For instance, the activity of researchers at the Santa Fe Institute (Gell-Mann 1994; Holland 1996, etc.) has been immense and extremely interesting. At present, this approach is also seeing a generous crop of developments in Europe. This can be seen, for example, at the several European Conferences on Complex Systems (ECCS) and at other symposia. The contributions in Spain are also prominent and continue to spread among various universities and researchers as for example San Miguel et al. (2012) or Solé and Bascompte (2006). By contrast, much of Morin’s work has not yet been translated into English or translations have appeared only slowly and have not reached a wide audience, and Elias’s approaches remain at the fringes of mainstream sociology, even though they are gaining wider recognition with each passing day. The development of the mathematical and computational line poses a challenge and an obligation for us to enter into mutual dialogue. Today it is a matter of profound urgency for us to examine what it has to offer, its contributions and its advantages, and to explore its limits as well, if we are to make headway in sociological knowledge, particularly in the field of sociolinguistics. Recently, some authors have already taken this task in hand and are able to offer their reflections to us. This is the case, for example, with Castellani and Hafferty (2009), Malaina (2012), Roggero (2013), Ruiz Ballesteros (2013), Solana Ruiz (2013), and Byrne and Callaghan (2014). I am especially delighted by the publication of the last one, because it has appeared in English—and can therefore reach a broader audience—and because it deeply takes both traditions into consideration, integrating and evaluating them, and it points to the limitations of ‘restricted’ complexity for the comprehension of human facts.

8.2 Models and Agents Models developed from the viewpoint of ‘restricted’ complexity —Castelló Llobet (2010), Castelló et al. (2011, 2013), Loureiro-Porto and M. San Miguel (2017) for example—typically use cellular automata programs in a computer to depict agents governed by simple rules of conduct they apply in accordance with any other types of agents with which they come into contact. At the same time, these other types of agents will apply their own rules. After a given number of iterations, the result at the level of language will be the greater or lesser use of one or another of the languages present. For instance, if one of the groups of agents is more bilingual than another group and it is more predisposed to use its second language to speak with members of the other group than to use its first language, we can see on screen how such a

122

A. Bastardas-Boada

situation will evolve. Depending on the number of agents present in each group, the program will enable us to observe—and calculate—the extent of each language use among the individuals and, as a consequence, the possible evolution of the situation as the behaviour resulting from the application of the rules becomes widespread. Although we value the potential of these kinds of contributions to our understanding of phenomena such as bilingualisation or language maintenance and shift, it is also clear that a really complexical approach to these phenomena does not stop here, even if we accept the validity of such contributions particularly when we are working with data from real cases, like Beltran et al. (2009, 2011).1 An examination of these kinds of phenomena from the viewpoint of general complexity also needs to be able to explain how and why particular rules have become established in agents, how the agents have been able to develop the necessary competences, whether it is possible for the agents to react and change their rules when they realise the effects of evolving mechanisms, or whether a group can change the rules if it sees that the outcome is that the group is not becoming as bilingual as another group and that this is harming it, for instance, at the economic level. If we accepted that the perspective of complexity should be confined to this sort of modelling and simulation, we would clearly be contradicting its paradigmatic principles, which are based on not “reducing the complex to the simple, but on translating the complex into theory” (Morin 1994: 35). Using the terminology of restricted complexity, the ‘agents’ in human societies are much more complex and changeable than the elements in the theory of physics, although it is true that human agents, in given situations and phases, may exhibit repetitive and regular behaviours that are frequently unconscious (Bastardas-Boada 1995). This phenomenon can be observed, for example, when the rules of language use have been established among the individuals of two language groups and the rules tend to be maintained in a routine and automatic manner, so that the individuals apply them with practically no conscious effort. However, this does not mean that, if some contextual or ideological change occurs, the speakers will not review and change the rules if they so choose. While human beings clearly depend on their contexts for the construction of the cognitive and linguistic faculties of their brains, they also possess autonomy of thought and control over their behaviours, even if this autonomy always exists in co-relation with the social pressures and developments of each society. Cellular automata or agent-based models (Wolfram 1983, 2002; Axelrod 1997) may partly capture the movement of agents and language use outcomes that emerge from their interactions at a given stage, but they are overly one-dimensional. They simplify too much and, at least in their current form, they do not incorporate other levels that play an important role. Even if we accept the mechanisms of inter-individual self-organisation, agents in reality face pressures originating in the economic and political domains as well as in the emotional domain, and these pressures can pro-

1 The

model is built on the basis of a community using two languages, one dominant and one subordinate. Individuals are characterised as monolingual speakers of the dominant code, as bilingual with a preference for the dominant code, or as bilingual with a preference for the subordinate code.

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

123

voke a partial or complete overhaul of the rules, if the agents so deem.2 A perspective steeped in general complexity calls for a theory of cognitively and emotionally active agents embedded within a sociocultural ecosystem that they have co-constructed and that, at the same time, retroactively influences them.

8.3 The Construction of a Theoretical Vision The task of building, in a coordinated and integrated manner, a general complexity perspective such as the one depicted here requires progress on both the theoretical and the methodological levels. Indeed, at present, there are advances being made in both domains, although they appear to lack integration and mutual communication. On the level of theory, general complexity—which we also can call complexics—needs to provide a set of principles, concepts and conceptual landscapes that can be applied transversally to distinct areas of knowledge and phenomena of reality, enabling us to gain a much firmer grasp of the complex aspects of their existence than we currently have. One of the profound changes that we need to address from the epistemological perspective of complexics is the tendency to disconnect the elements of reality once we have given a distinct name to each of them. Apparently, the act of assigning different names tends to lead us to think of these elements as existing independently, not interrelatedly, when, in reality, what is most typical is precisely their interdependence and interwovenness. If we turn our thoughts to ‘society’, for example, we imagine an entity not only different from the agents—human beings—who comprise it and give it existence, but also an entity that is separate in space. Society, we say, is ‘on top of us’. On this matter, Norbert Elias, is clear: “We talk of the person and his environment, a child and his family, the individual and society, the subject and objects without always realising that the person also forms a part of his ‘environment’, the child is a part of his family, the individual is a part of society, the subject is part of the objects” (Elias 1982: 14). In the phenomenon of language, this confusion can also arise. As we have already developed the concept of ‘language’, we may think that language exists in and of itself as an isolated and independent entity, when to the contrary we must conceive of it as a phenomenon closely tied to the human beings who give life to it and/or change it (or let it cease to exist). And this is where we have the debate on the locus of language—or of ‘languaging’. Where do forms of languaging reside: in the individual or in society? As we can see, this is a spurious debate. ‘Society’ is not something outside the individuals who are its members. Rather, they cause it to ‘emerge’. It is always a society-of-individuals. For Elias, the patterns of human culture are an emerging 2 Byrne

and Callaghan take the same view that I do: “Agent-based models in particular remain trapped, when used in isolation, within a micro-emergent understanding of the social. The social is not merely micro-emergent and any account of it which ignores the reality of what we must call conventionally ‘social structure’ is always going to be incomplete” (2014: 257).

124

A. Bastardas-Boada

property of social processes, the unplanned result of interwoven plans and of the emotional and rational impulses of individual people: “From this interdependence of people arises an order sui generis, an order more compelling and stronger than the will and reason of the individual people composing it” (Elias 2000: 366). Indeed, the forms of human languaging are assuredly a singular phenomenon, because they live in and among people, requiring important conceptual changes to the representations that we have hitherto maintained. One approach is to think of them analogically as if they were a dance: “While different people can dance the same dance figuration, there is no dance as such without dancers” (Dunning and Hugues 2013: 53). Thus, we can study the different language ‘dances’ created by humans, but we must not lose sight of the fact that they are the socio-communicative actions of diverse groups of people. Forms of languaging are independent of any particular individual, but not of individuals as such.

8.4 Seeking to Understand the Complexity of Language Contact and Bilingualisation Even in their simplest form—for example, in the case of only two languages—the structure and evolution of the phenomena of language contact are not straightforward. In fact, they show a significant degree of complexity. The number of individuals engaged in contact may be large and there are also many different domains and inter-influences that occur in the lives of human beings. It may well hold greater interest, therefore, to apply a complexical and dynamic perspective that can enable us to see all the factors and their interrelationships and understand their interwoven evolution. This was clearly evident to Weinreich: “It is in a broad psychological and sociocultural setting that language contact can best be understood. (…) On an interdisciplinary basis research into language contact achieves increased depth and validity” (1968: 4). One of the fundamental distinctions that we must take into account from the outset is the type of society in which contact happens. Is the society still rural and poorly developed technologically and economically? Or does it, conversely, have a social structure characterised by a high degree of urbanisation and industrial development? In the first set of cases, there may be greater importance in factors that are more local in nature, concerning physical proximity and face-to-face contact, while the second set of cases will also feature the forceful interventions of a society’s various institutional organisations, ranging from those in the political sphere—the official administration, the educational system, the healthcare system, and so forth—to those that are more closely bound up with the economic sphere and the media. Within these organisations, is there large-scale face-to-face contact or only contact through institutional channels, or do both types of contact occur at the same time? It is also of special importance to stress that language contact must be understood as a historical and, therefore, temporal phenomenon, with earlier events playing a

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

125

major role in how the phenomenon evolves. In other words, we need to pay attention not only to the synchronic elements, but also to the diachronic ones, because the latter may determine the future development of the phenomenon (Elias 1982). For instance, in the initial phase of contact between groups, one of the essential factors concerns the language competences that individuals have previously developed, as well as their mutual cognitive and emotional representations. If, for example, one of the groups possesses considerable knowledge of the language of the other group, while the latter has not acquired similar knowledge of the former’s language, this difference will be a highly significant variable in the process as it develops. This type of situation, for instance, can give rise to a case in which one group—generally in a subordinate situation—has received instruction in the dominant language of the State in which it resides thanks to the compulsory educational system, while another group has not received instruction in the language of the first group. This is typically what happened in States such as Spain that are made up of diverse language groups, but have had only one official language. In this sort of context, when contact produced initially by political means turns into a different situation in which populations move beyond their traditional language areas and come into daily face-to-face contact, how the emerging interaction is organised will tend to favour the use of the more commonly shared language by both language groups (Hamers and Blanc 2000). This will tend to be the language that has become the exclusive language of instruction in the official educational system. The selection of this language in personal interactions, therefore, will be viewed as practical, convenient and ‘normal’. At this point, the process will be acted upon by the social mechanisms of continuity and automaticity—Bourdieu’s habitus (1980)—which encourage the development of routinised and subconscious behaviours that can eventually come to be seen as obvious and beyond question (Nisbet 1977). This is the typical case of contact between the majority group of a State and its subordinate minoritised groups, which, as we shall see, can embark on a negative course in the use of their customary language forms, particularly if the abandonment of these forms is also encouraged through the spread of a negative discourse and representations aimed at this result. In this context, the vast majority of conversations between the subordinate bilingualised group and the other, monolingual group will tend to take place in the language of the latter. In this first phase of their encounter, the bilingualised group will tend to maintain a distribution of functions, given that interactions between members of the group will continue to make use of the own group’s language, while conversations outside the group, and often official written activities as well, will make use of the other language, the one dominant in the State. However, if social interpenetration is great and the presence of people from the group speaking the dominant language is common in the conversations of the bilingualised group, members of the latter will come under pressure to use the dominant language even among themselves, in order not to marginalise monolingual individuals from the conversation, at least until they have developed sufficient comprehension of the original language of the bilingualised group.

126

A. Bastardas-Boada

Starting to use the language originally acquired through the school system in everyday interactions of a private nature will represent a change among bilingualised individuals, increasing their colloquial competence in the language. Daily conversational practice will increase the automaticity of their speech and they may move from a more formal, written knowledge to a spoken and colloquial competence, which could make them feel more comfortable and fluent in their second language. Gradually through feed-back (Wiener 1948) and recursively (Morin 1977), the effects will have an influence on behaviours, which will produce more effects, effects which are favourable in this case to the use of the dominant official language. With intergenerational change, the circumstances may give rise to a situation in which the minority group abandons its original language.

8.5 The Interwoven Evolution of Situations The contact between two language groups will never be static. It will change as a result of the effects of the encounter between the two groups and because of other factors that can arise out of the circumstances of life among the groups in contact. Even without the presence of official or institutionalised communications (Corbeil 1983), the groups’ interrelation in and of itself can produce an increase in the oral skills of the smaller demolinguistic group and thereby contribute to greater interaction. This, in turn, can encourage the growth of pairings between individuals of mixed ethniclinguistic origin by means of a recursive mechanism. In developed societies, these individuals will, in all likelihood, tend to use the dominant official language with one another and, depending on the case, this may or may not also be the language spoken by parents to their offspring. If it is, the children will typically only have the dominant language as their mother tongue. However, depending on the level of ethnic awareness or the usefulness that may correspond to the native language of the bilingual parent in such a pairing, he or she may choose to speak this language with the children, enabling them to become socialised as bilingual within the family (Bastardas-Boada 2016b). This kind of sociolinguistic organisation, however, typically requires the other parent to develop at least receptive skills in the other language. This is because uncomfortable situations can arise if they do not do so. For example, they may not be able to understand conversations between the children and the other parent in their own home. If most couples of mixed origin choose to speak the language of the dominant group with one another and with their children, an interruption will occur in the family transmission of the first language of the bilingualised group, setting in motion a significant process of language shift. Indeed, this has been one of the mechanisms responsible for the loss of speakers in the Welsh case (Williams 2005), for example, and among immigrants (Boix-Fuster 2009). If the number of marriages between individuals of mixed origin is high, the number of individuals possessing the subordinate

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

127

language will decline and these individuals will themselves become a pressure factor driving the social use of the dominant language among speakers who still have the initial language of the bilingualised group. In some cases, the intergenerational abandonment of the group’s own language in favour of the language used by the majority or dominant group within the State occurs not only because of mixed marriages, but also because of parents’ decisions to use the dominant language with their children rather than the language of their origin. If parents have been sufficiently bilingualised through their exposure to the educational system and/or the use of their second language with speakers of this language within their society, they will potentially be in a position to use it with their children. This occurs especially if the parents live in situations in which the asymmetry of power between the groups is very high and if they have internalised negative mental representations of their own code and, conversely, possess representations of the other code that are favourable, e.g., that it is useful for socioeconomic mobility. In these kinds of situations—such as, for example, in Galicia (Lorenzo Suárez 2003) or in the autonomous community of Valencia (Querol 1990; Conill 2003; Montoya Abat and Mas i Miralles 2012) in Spain, or in communities of indigenous speakers in Mexico (Terborg and García-Landa 2011, 2013)—the language of origin can eventually be viewed as an obstacle to economic advancement and to individuals’ social prestige. When this is the case, usage of the language of origin can be abandoned during intergenerational transmission in order to prevent the harmful conditions that parents have lived through from being suffered by their children as well. A similar situation often occurs in immigrant groups who arrive in countries where a different language is used. With intergenerational succession, immigrant groups can lose interest in maintaining their language of origin and become monolingual in the language of their host country. This kind of evolution, which is a priori more linear and predictable, can be depicted more readily by models, but it is perhaps more difficult to build into such models the possibility of changes that are, in principle, unexpected; changes that the agents themselves may decide to adopt at a given historical moment. Moreover, given that the world never stands still, new contextual factors can come into play and modify the projected evolution of any given case. Computational modelling appears suitable for taking into account any adaptive changes that happen to come out of the model itself, that is, the changes that correspond to the conditions and rules initially programmed into the software. What the software cannot predict, however, is the appearance of new meanings in a situation, much less the introduction of external events that can have influence on it. As a result, the predictive power of such models will necessarily be limited. An illustration of this type of case comes from our real-world Catalan laboratory, where we have seen how agents who, in principle, should have followed the anticipated rules of monolingual Spanish usage with autochthonous Catalan speakers who are bilingual in Catalan and Spanish, have evolved over time toward their own bilingualisation not only at the level of comprehension, but even at the level of expression, particularly if they came as young people. With people they met when they were able to speak only Spanish, they have continued to use Spanish. With oth-

128

A. Bastardas-Boada

ers with whom they have struck up a relationship since developing the ability to use Catalan for social contact, they regularly speak Catalan, even in the presence of other Spanish-speaking interlocutors. With the intergenerational replacement of the population, the number of people of immigrant origin who are able to speak Catalan and use Catalan with native speakers has risen where, of course, the necessary demolinguistic conditions have existed. How the situation has evolved corresponds not only to the strict application of the initial rules of the encounter between the two groups, but also to the socioeconomic contexts in which the encounter occurred. Given that Catalan remained the most utilised language in informal interactions among its own language group even during the Franco dictatorship, and also that Catalan speakers were in control of a large portion of the private sector, there was probably some rethinking of the sociolinguistic rules in play at the outset, even though the language policy under Francoism sought to impose the opposite outcome. This does not mean that the political domain is of no importance in determining how situations of language contact evolve. The rules followed by individuals in the use of their languages are not solely the result of what occurs at the level of interactions. Taking the example of Catalonia, as noted earlier, the individuals who came into contact with one another had different language skills at the start. While first-language Spanish-speakers arriving in the Catalan-speaking area had absolutely no knowledge of the language, most of the individuals of Catalan origin had, in fact, received schooling in Spanish and had studied the language. Clearly, the encounter did not occur on equal terms. The two sets of interlocutors could use Spanish with one another, but not Catalan, which has historically been the language of the host area. The rules of interaction that were established between them, therefore, did not arise solely from the properties of the simple encounter between the groups. They were also affected by past and present historical and political factors. Alongside the evolution described in the previous paragraph, which reflected the social bilingualisation of native Spanish-speakers, Catalonia also witnessed thousands of Spanish-speakers who did not develop the ability to speak Catalan, especially after the large-scale migrations in the period 1960–75. Subsequently, many new neighbourhoods sprang up in which the vast majority of residents were Spanish-speakers by origin. The opportunity to have close contact between the two populations was limited, and influence at the demolinguistic level was severely undermined. Given that only Spanish was supported at the political level, bilingualisation in Catalan was meagre (Alarcón and Garzón 2011). Because humans are cognitive and emotional agents, the political level is also interconnected with the ideological level. At the end of the Franco dictatorship (the dictator died in 1975), the Catalan population mobilised to express its desire for political democracy and autonomy. Joining in these demands were many who had arrived in Catalonia in the previous decades. While the recent newcomers were sometimes not yet fluent speakers of Catalan, they were sympathetic toward the language and lent their support to its official recognition. In this setting, additional people with Spanish-speaking backgrounds sought to change their rules of language behaviour in order to speak Catalan with autochthonous Catalan speakers, even though it was hard at times. Nor was the change in behaviour much encouraged by individuals of

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

129

Catalan origin, many of whom continued to apply the rule of adapting to the first language of the interlocutor, typically robbing Spanish-speakers of the opportunity to practice Catalan in conversation. Agents’ ability to change the rules, however, reaches a high point when, in interactions between the two language groups, a native Catalan-speaker uses Spanish with an interlocutor of immigrant origin and the latter replies in Catalan. This is because the two wish to display their empathy and desire to adapt to each other, making clear that the variables of emotion and identity must also be taken into account. In this ideological-political and interpretative domain, the self-representations of language groups also play a major role. These stem from the socio-political and economic history of each group. Within the language area in which Catalan is used in its several variants, we find significant differences that can help to shed light in this respect. One of the complex aspects of Catalan/Spanish contact in Catalonia is to understand why the repression and prohibition of the public use of Catalan during most of the first three-quarters of the twentieth century produced disparate language behaviours and ideologies in Catalonia and other areas in the Catalan-speaking lands, such as the autonomous community of Valencia. At present, when the majority of the population of Catalonia is pushing for full restoration of the use of Catalan and is confronting the difficulties regularly interposed by the Spanish government, the authorities in the autonomous community of Valencia have been, until very recently, less active in defending the use of their language, and the process of intergenerational abandonment continues apace (Vila i Moreno 2011; Boix-Fuster and Farràs 2012). It is not easy to explain the reasons for these contrasts. We would have to compare the historical evolution of these two language regions within the Catalanspeaking territories. One of the differential elements is the earlier industrialisation of Catalonia, which led to the creation of an autochthonous bourgeoisie and a positive self-image with respect to other areas of Spain, which lagged behind in this respect. The autonomous community of Valencia had a more agricultural economy that was less developed. Today, however, the Valencia region has an advanced economy and developed agriculture. Yet the people’s image of their identity, in large part, does not correspond to that of Catalonia. While numerous people in Catalonia report feeling strictly Catalan or more Catalan than Spanish, the opposite is true in the autonomous community of Valencia. That is, a substantial number of individuals feel more Spanish than Valencian or both in equal terms (Coller 2006.) It is in this aspect of the hierarchical organisation of identities where we can find an explanation for their differing language behaviours. The customary language of debates in the Parliament of Catalonia is predominantly Catalan, while in the autonomous community of Valencia, in general, has been Spanish. Language behaviour of this sort has a significant relationship to the identity-related representations that human groups possess. When making a choice of identity between the State and the community of origin, a positive group self-representation supports the intergenerational maintenance of the language. Conversely, if the group’s own identity is considered to be subordinate to the State, the language will be viewed as dispensable and the group will opt for the State’s dominant official language. Once again, we see how the elements that may

130

A. Bastardas-Boada

have an effect on the selection of language behaviours are complexly intertwined, making it difficult to reduce them to precise, stable rules that remain unchanged over time. What this comparison between Catalonia and the Valencian community again shows is how important it is to introduce the historical element when examining language behaviours in situations of contact and to study such situations on a caseby-case basis. Perhaps in many processes of bilingualisation and language shift, the elements are alike or very similar, but path dependency also exerts an influence and it can be crucial for the final outcomes. Early industrialisation, for example, can boost the positive representation of a given language group so that it is then able to confront a situation of political subordination with greater chances of success than a group embarking on economic development later, when the process of bilingualisation is already well underway and widespread among its speakers. In the latter case, the group in question can have a perception of its own inferiority with respect to the elites of the dominant group within the State, and this can lead the group to embrace the dominant group as a yardstick and, therefore, to attempt to assimilate. This will have a strong impact on language behaviours, which will then tend to favour use of the State’s dominant language instead of the group’s own language. As we have seen, the socioeconomic level also appears to play a major role in the adoption or non-adoption of the dominant language. In the expansion process of Spanish in Spain, we can see clearly how the upper layers of different language groups are the first to adopt Spanish in family usage (Boix-Fuster and Torrens 2012). This is because they want to move closer to the centres of power and to distinguish themselves from lower social classes (Bourdieu 1984). A common consequence of this movement is the emulation of their behaviour by other socioeconomic segments, especially the middle classes, which seek economic advancement and associate the use of the dominant language—and the abandonment of their group’s own language—with the social prestige of the yardstick group. On the value scale, the language of origin then becomes associated in terms of its social signification with groups lower down the social ladder and more backward in economic development. The temporal dynamic is essentially urban and centre-periphery in nature, from regional to sub-regional capitals and ultimately to municipalities, always starting with the upper classes. In these contexts, intergenerational replacement will act as the major mechanism of language shift. Parents in search of better socioeconomic prospects for their children will tend to use the dominant language with them and not the language of origin. As a result, the dominant language will become native for the subordinate group. With the help of the compulsory educational system and the media, the process will accelerate and reach the vast majority of the population, who will see the adoption of the new language and the abandonment of their own language as the road to economic progress and social respect.

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

131

8.6 Language Shift and Its Reversal Gradually, this evolution leads to the extinction of many languages even within their own historic lands and it can be halted only with a change in the socio-political conditions in which contact occurs and, particularly, with a change in the cognitive and emotional representations of their speakers (Fishman 1991). The efforts made to reverse language shift in Spain in the past thirty-five years show some promise, but they also point to the limits of this kind of complex process. With the advent of democracy, Spanish lost the exclusivity of its use in the educational system and the other languages also acquired a range of (co)official uses in their own territories, expanding their administrative uses and their functions in the public arena, but not all have done so to the same degree and at the same pace (Siguan 1993; Turell 2001). Although the prospective legal framework was the same in all cases, the measures adopted by the governments of the various autonomous communities have differed. They have reflected the prevailing ideas and attitudes of each community. Thus, while governments in Catalonia and the Basque Country have tended to be largely in the hands of parties backing the restoration of their own languages, this has not always been so in the Balearic Islands, Valencia or Galicia, which nonetheless differ in degree from one another. In communities whose own identity is less strong or where there is greater division among their citizens, the historical momentum of language shift is so powerful that, even though they have now declared official the previously subordinate language, the mechanism of its intergenerational abandonment continues apace and many families of autochthonous origin choose to use Spanish with their children instead of their language of origin. It is as if the ideas inherited from the time of the dictatorship, which ran counter to the maintenance and public use of languages other than Spanish, were still in force in people’s mental representations and continued to act on their behaviours. In the ecology of pressures (Terborg and García-Landa 2013) that they perceive, the elements supporting their adoption of the dominant language of the State are strong and those that might back the intergenerational maintenance of the language of origin are losing out (Lieberson 1970; Gal 1979). In all likelihood, differences in social meanings associated with each language also carry weight here. For example, they may associate Spanish with greater political, economic and cultural power, while attaching meanings to the other languages that relate to the rural, the lower class or lower literary prestige. As I have indicated, it can be easier to overcome these difficulties and reverse processes of language shift and/or gain new speakers for the subordinate language if, historically, the group has achieved economic development and maintains a positive cognitive self-image with respect to the majority group of reference. In the case of Catalonia and the Basque Country (Alarcón and Garzón 2011, Azurmendi and Martínez de Luna 2011), the identities of the people themselves have been viewed by a majority of their citizens as important and not subordinate to the Spanish identity of the State. The opposite tends to be the case in the autonomous communities of Valencia and Galicia, where the group’s own identity seems to have less weight.

132

A. Bastardas-Boada

Adopting a complexity perspective to conduct a comparative study of cases in which a language has regained its official status and public use, we can see certain phenomena that are of interest for sociolinguistic theory. In the case of Catalonia, for instance, we clearly observe the different pace of changes in the political and administrative sphere and within society (Strubell and Boix-Fuster 2011). Although the language undergoing recovery increases its official uses and is introduced as the language of instruction in the school system, the communicative habits established in society as a whole do not change at the same speed. For some time longer, these habits preserve norms that became predominant as a result of sociolinguistic self-organisation. Confirming the distinction drawn by Corbeil (1983) between ‘institutionalised’ and ‘individualised’ communications and also by Ryan (1979), we can see how these two levels co-exist but are distinct, and how social agents acting on one of the two levels can pursue different patterns of behaviour. The temporal asymmetry between what occurs at the institutionalised level and what happens at the level of individuals in their daily lives can also, conversely, explain the maintenance of languages other than the official language of the State for long periods in spite of government policies clearly aimed at encouraging their disuse, such as we saw in Spain during the Franco dictatorship. We can have periods that usher in the formal bilingualisation of a population by institutionalised means, but nevertheless see the maintenance of the groups’ own varieties in everyday social uses, much as in the diglossic distribution of uses in the German-speaking region of Switzerland. However, in the long run, changes may arise in everyday social uses because of the influence exerted by uses at the level of institutionalised communications. It remains to be seen how quickly this may happen and how it will be distributed among different social classes and/or groups of different origin.

8.7 Closing Thoughts: A Necessary Integration As we have seen, historical processes exert an influence on the current state and evolution of situations of language contact. This influence is brought to bear from different domains: the economic and the political, the ideological and group identities, geo-demographics, and the habits of inter-group use, among others. Clearly, this kind of phenomenon requires study from a complexical and holistic perspective in order to accommodate the variety of factors that belong to different levels and that interrelate with one another in the evolving dynamic of human languaging. The general complexity or complexical perspective allows for and encourages this integrated vision to account for what occurs autonomously at the level of agents’ interactions and situations and, at the same time, for how all of these factors are ecodynamically interwoven and inter-influence what occurs in the political, ideological, economic and technological contexts in which individuals live, contexts which they themselves co-develop. It is clear that the appearance and/or consolidation of these new theoretical perspectives must necessarily have ramifications at the more practical level of method-

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

133

ology. New tools for the conception, apprehension and treatment of the data of experience will need to be devised to complement existing ones and to enable us to make headway toward practices that better fit complexical perspectives. In the case of computational complexics, one characteristic of this kind of modelling is that it uses few parameters. This clashes with the aspiration of complexity theory to build a comprehensive ecology out of the elements involved: “Several models have been proposed to account for different mechanisms of social interaction in the dynamics of social consensus. The idea is to capture the essence of different social behaviours by simple interaction rules: following the idea of universality classes, in collective emergent phenomena details might not matter” (Castelló Llobet 2010: 24). Morin (2005: 4) takes a rather more critical view: “Restricted complexity has enabled important advances to be made in formalisation, in the possibilities of models, which in turn stimulates the potential for interdisciplinary efforts. But one is still within the epistemology of classical science. (…) In some sense, complexity is acknowledged, but it is decomplexified”.3 To gain an adequate view of the whole and to understand the how and why of the process pursued by the agents in reaching the states that guide their decisions, as Xavier Castelló has similarly put it, it will probably be necessary to use computational research together with other types of research that are closer to the changing cognitive and emotional activity of the agents. The need in my view is for the two lines to come to a meeting of the minds, and take steps toward a mutual integration based on the acceptance of the shortcomings of each approach, achieving progress through a non-contradictory complementarity of perspectives. It must be conceded that the practical and methodological applications of basic complexical ideas need to be developed much farther in order to apply them to specific research. At the same time, the limits of complex adaptive systems as computational strategies must be accepted in the pursuit of a better understanding of the dynamic and evolutionary processes typical of human beings. It will certainly be useful for sociolinguists, for example, to gain familiarity with the contributions of quantitative-oriented physicists from the field of statistical physics modelling, such as Murray Gell-Mann, Maxi San Miguel and Albert Díaz-Guilera, and see their fruitful application in our disciplines and attempt to exploit them in a coherent and integrated manner. However, I think we must also be cognizant of the peculiarities of human phenomena, which are characterised by the existence not only of purpose and regularity in the control of behaviour, but also by the significant degree of agents’ cognitive and interpretative autonomy and by the powerful influence of the emotional dimension. It seems obvious, therefore, that human complexics must be seen as multimethodological, insofar as necessary combining quantitative-computation methodologies and more qualitative methodologies aimed at understanding the historical mental and emotional world of people (cf. Malaina 2012). Thus, the methods and concepts of restricted or computational complexics can help and be used as sup3 That said, nobody can deny the importance of the studies conducted to date from the perspective of

complex systems, or the utility of modelling, which has brought us nearer to the essential elements of processes and to the expression of their interrelationships with the utmost clarity.

134

A. Bastardas-Boada

plementary strategies that are highly useful in studying certain characteristics, the stages and speeds of processes of language contact, but always within the frame of the broader view offered by general complexics. As Byrne and Callaghan say, “[w]e see complexity as providing a framing for the unifying of a whole set of opposites in scientific practice, of quantitative and qualitative research, of analysis and holism as modes of understanding, and of relativism and hard realism as epistemological position” (2014: 255).

References Alarcón, A., & Garzón, L. (2011). Language, migration and social mobility in Catalonia. Leiden: Koninklijke Brill NV. https://doi.org/10.1163/ej.9789004211230.i-160. Allen, T. F. H., & Hoekstra, T. W. (2014). Toward a unified ecology. New York: Columbia University Press. https://doi.org/10.1016/0169-5347(93)90065-w. Aracil, Ll. V. (1982). Papers de sociolingüística. Barcelona: La Magrana. Aracil, Ll. V. (1983). Dir la realitat. Barcelona: Edicions Països Catalans. https://doi.org/10.1017/ s0047404500012148. Axelrod, R. (1997). Complexity of cooperation. Princeton: Princeton University Press. https://doi. org/10.1002/(sici)1099-0526(199801/02)3:3%3c46:aid-cplx6%3e3.0.co;2-k. Azurmendi, M. J., & Martínez de Luna, I. (2011). Success-failure continuum of Euskara in the Basque country. In J. A. Fishman & O. García (Eds.), Handbook of language and ethnic identity (pp. 323–335). Oxford: Oxford University Press. Bastardas-Boada, A. (1995). Language management and language behavior change: Policies and social persistence. Catalan Review, 9(2), 15–38. Bastardas-Boada, A. (1996). Ecologia de les llengües. Medi, contacte i dinàmica sociolingüística. Barcelona: Proa. (English version: From language shift to language revitalization and sustainability: A complexity approach to linguistic ecology. Barcelona: Publicacions de la Universitat de Barcelona, in press). Bastardas-Boada, A. (1999). Lingüística general y teorías de la complejidad ecológica: algunas ideas desde una transdisciplinariedad sugerente. In J. Fernández González (Ed.), Lingüística para el siglo XXI (pp. 287–294). Salamanca: Publ. Universidad de Salamanca. Bastardas-Boada, A. (2013a). General linguistics and communication sciences: Sociocomplexity as an integrative perspective. In M. À. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 151–173). Berlin: Springer. Bastardas-Boada, A. (2013b). Language policy and planning as an interdisciplinary field: Towards a complexity approach. Current Issues in Language Planning, 14(3–4), 363–381. https://doi.org/ 10.1080/14664208.2013.829276. Bastardas-Boada, A. (2014). Some physics, ecology and complexity ideas for a complexfigurational sociolinguistics. History of the Human Sciences, 27(3), 55–75. https://doi.org/10. 1177/2F0952695114534425. Bastardas-Boada, A. (2016a). Complexity in language contact: A socio-cognitive framework. In S. S. Mufwene, C. Coupé, & F. Pellegrino (Eds.), Complexity in language: Developmental and evolutionary perspectives (pp. 218–244). Cambridge, UK: Cambridge University Press. https:// doi.org/10.1017/9781107294264. Bastardas-Boada, A. (2016b). Famílies lingüísticament mixtes a Catalunya: competències, usos i autoorganització evolutiva. Treballs de Sociolingüística Catalana (English version in press: Mixed-language Families in Catalonia: Competences, Uses and Evolving Self-organisation. In A. Bastardas-Boada, E. Boix-Fuster, & R. M. Torrens Guerrini (Eds.), Family Multilingualism

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

135

in Medium-Sized Linguistic Communities. Bern: Peter Lang.), 26, 285–308. https://doi.org/10. 2436/20.2504.01.121. Bastardas-Boada, A. (2016c). Complexics as a meta-transdisciplinary field. Congrès Mondial pour la pensée complexe. Les défis d’un monde globalisé. (Paris, 8–9 décembre). UNESCO. Retrieved from https://www.reseau-canope.fr/fileadmin/user_upload/Projets/pensee_complexe/bastardas_ boada_complexics_meta_transdisciplinary_field.pdf. Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine Books. https://doi.org/10. 7208/chicago/9780226924601.001.0001. Beltran, F. S., Herrando, S., Ferreres, D., Adell, M. A., Estrader, V., & Ruiz-Soler, M. (2009). Forecasting a language shift based on cellular automata. Journal of Artificial Societies and Social Simulation, 12(3), 1–8. Retrieved from http://jasss.soc.surrey.ac.uk/12/3/5.html. Beltran, F. S., Herrando, S., Estrader, V., Ferreres, D., Adell, M. A., & Ruiz-Soler, M. (2011). A language shift simulation based on cellular automata. In E. G. Blanchard & D. Allard (Eds.), Handbook of research on culturally-aware information technology: Perspectives and models (pp. 136–151). Hershey, New York: Information Science Reference. https://doi.org/10.4018/9781-61520-883-8.ch007. Bohm, D. (1980). Wholeness and the implicate order. London: Routledge & Kegan Paul. https:// doi.org/10.4324/9780203995150. Boix-Fuster, E. (2009). Català o castellà amb els fills? La transmissió de la llengua en famílies bilingües a Barcelona. Barcelona: Rourich. Boix-Fuster, E., & Torrens, R. M. (2012). Les llengües al sofà. El plurilingüisme familiar als països de llengua catalana. Lleida: Pagès editors. Boix-Fuster, E., & Farràs, J. (2012). Is Catalan a medium-sized language community too? In F. X. Vila (Ed.), Survival and development of language communities prospects and challenges (pp. 157–178). Clevendon: Multilingual Matters. Bourdieu, P. (1984). Distinction: A social critique of the judgement of taste. Cambridge: Harvard University Press. Byrne, D., & Callaghan, G. (2014). Complexity theory and the social sciences. The state of the art. London: Routledge. Capra, F. (2002). The hidden connections. New York: Doubleday. Castellani, B., & Hafferty, F. W. (2009). Sociology and complexity science. Berlin: Springer. https:// doi.org/10.1007/978-3-540-88462-0. Castelló Llobet, X. (2010). Collective phenomena in social dynamics: Consensus problems, ordering dynamics and language competition (Doctoral dissertation). University of the Balearic Islands. Palma de Mallorca. Retrieved from https://ifisc.uib-csic.es/en/publications/collectivephenomena-in-social-dynamics-consensus/. Castelló, X., Vazquez, F., Eguíluz, V. M., Loureiro-Porto, L., San Miguel, M., Chapel, L., et al. (2011). Viability and resilience in the dynamics of language competition. In G. Deffuant & N. Gilbert (Eds.), Viability and resilience of complex systems (pp. 39–74). Heidelberg: Springer. https://doi.org/10.1007/978-3-642-20423-4_3. Castelló, X., Loureiro-Porto, L., & San Miguel, M. (2013). Agent-based models of language competition. International Journal of the Sociology of Language, 221, 21–51. https://doi.org/10.1515/ ijsl-2013-0022. Coller, X. (2006). Collective identities and failed nationalism. The case of Valencia in Spain. Pôle Sud, 25, 107–136. https://doi.org/10.3406/pole.2006.1357. Conill, J. J. (2003). The situation of Valencian as reported in non-institutional sociolinguistic research (1998–2002). Noves SL. Revista de Sociolingüística (pp. 1–9). Retrieved from http:// www.gencat.cat/llengua/noves/noves/hm03primavera/catalana/a_conill.pdf. Corbeil, J. C. (1983). Élements d’une theorie de la regulation linguistique. In É. Bedard, & J. Maurais (Eds.), La norme linguistique (pp. 281–303). Québec & Paris: Conseil de la Langue Française. Retrieved from http://www.cslf.gouv.qc.ca/bibliotheque-virtuelle/publication-html/? tx_iggcpplus_pi4%5bfile%5d=publications/pubf101/f101p3.html#x. Dunning, E., & Hugues, J. (2013). Norbert Elias and modern sociology. London: Bloomsbury.

136

A. Bastardas-Boada

Elias, N. (1982). Sociología fundamental. [Spanish translation of Was ist Soziologie? Munich, Juventa Verlag, 1970]. Barcelona: Gedisa. Elias, N. (2000). The civilizing process. [English translation of Über den Prozess der Zivilisation. Soziogenetische und psychogenetische Untersuchungen. Basel, Haus zum Falken, 1939]. Oxford: Blackwell. Fishman, J. A. (1991). Reversing language shift. Clevedon: Multilingual Matters. Gal, S. (1979). Language shift: Social determinants of linguistic change in bilingual Austria. San Francisco: Academic Press. Gell-Mann, M. (1994). The Quark and the Jaguar: Adventures in the Simple and the Complex. New York: Henry Holt and Company. Hamers, J. F., & Blanc, M. H. A. (2000). Bilinguality and bilingualism. Cambridge: Cambridge University Press. Haugen, E. (1972). The ecology of language. In A. S. Dil (Ed.), The ecology of language (pp. 325–339). Stanford: Stanford University Press. Holland, J. H. (1996). Hidden order: How adaptation builds complexity. New York: AddisonWesley. Lieberson, S. (1970). Language and ethnic relations in Canada. New York: Wiley. Lorenzo Suárez, A. M. (2003). Studies relating to the sociolinguistic situation of the Galician language (1990–2002). Noves SL. Revista de Sociolingüística, Primavera, 2003, 1–9. Loureiro-Porto, L., & San Miguel, M. (2017). Language choice in a multilingual society: A view from complexity science. In S. S. Mufwene, C. Coupé, & F. Pellegrino (Eds.), Complexity in language. Developmental and evolutionary perspectives (pp. 187–217). Cambridge, UK: Cambridge University Press. Mackey, W. F. (1979). Toward an ecology of language contact. In W. F. Mackey & J. Ornstein (Eds.), Sociolinguistic studies in language contact (pp. 453–460). The Hague: Mouton. Mackey, W. F. (1980). The ecology of language shift. In P. H. Nelde (Ed.), Sprachkontakt und Sprachkonflikt (pp. 35–41). Wiesbaden: Franz Steiner Verlag. Mackey, W. F. (1994). La ecología de las sociedades plurilingües. In A. Bastardas-Boada & E. Boix-Fuster (Eds.), ¿Un estado, una lengua? La organización política de la diversidad lingüística (pp. 25–54). Barcelona: Octaedro. Malaina, Á. (2012). Le paradigme de la complexité et la sociologie. Possibilité et limites d’une sociologie complexe. Paris: L’Harmattan. Margalef, R. (1991). Teoría de los sistemas ecológicos. Barcelona: Publicacions de la Universitat de Barcelona. Maturana, H. (2002). Autopoiesis, structural coupling and cognition. Cybernetics & Human Knowing, 9(3–4), 5–34. Maturana, H., & Varela, F. J. (2004). De máquinas y seres vivos. Autopoiesis: La organización de lo vivo. Buenos Aires: Lumen. Montoya Abat, B. & Mas i Miralles, A. (2012). Language teaching and family linguistic transmission: Two correlative factors in the Valencian Region (Catalan vs. Spanish)? Sociolinguistic Studies, 6(1), 45–63. https://doi.org/10.1558/sols.v6i1.45. Morin, E. (1977). La nature de la nature. Paris: Seuil. Morin, E. (1994). La complexité humaine. Paris: Flammarion. Morin, E. (2005). Restricted complexity, general complexity. Colloquium Intelligence de la complexité: épistémologie et pragmatique. (Translated from French by Carlos Gershenson in 17 June, 2014). https://doi.org/10.1142/9789812707420_0002. Nisbet, R. (1977). The social bond. New York: Random House. Prigogine, I., & Stengers, I. (1979). La nouvelle alliance. Métamorphose de la science. Paris: Gallimard. Prigogine, I., & Stengers, I. (1992). Entre le temps et l’éternité. Paris: Flammarion. Querol, E. (1990). El procés de substitució lingüística: la comarca dels Ports com a exemple. Miscel·lània, 89, 87–196.

8 ‘Restricted’ and ‘General’ Complexity Perspectives on Social …

137

Roggero, P. (2013). Para una sociología según El método. In E. Ruiz Ballesteros & J. L. Solana Ruiz (Eds.), Complejidad y ciencias sociales (pp. 103–123). Sevilla: Universidad Internacional de Andalucía. Ruiz Ballesteros, E. (2013). Hacia la operativización de la complejidad en ciencias sociales. In E. Ruiz Ballesteros & J. L. Solana Ruiz (Eds.), Complejidad y ciencias sociales (pp. 137–172). Sevilla: Universidad Internacional de Andalucía. Ryan, E. B. (1979). Why do low-prestige language varieties persist? In H. Giles, & R. N. St. Clair (Eds.), Language and social psychology (pp. 145–157). Oxford: Blackwell. San Miguel, M., Johnson, J. H., Kertesz, J., Kaski, K., Díaz-Guilera, A., MacKay, R. S., et al. (2012). Challenges in complex systems science. The European Physical Journal. Special Topics, 214, 245–271. https://doi.org/10.1140/epjst/e2012-01694-y. Siguan, M. (1993). Multilingual Spain. Amsterdam: Swets-Zeitlinger. Solana Ruiz, J. L. (2013). El concepto de complejidad y su constelación semántica. In E. Ruiz Ballesteros & J. L. Solana Ruiz (Eds.), Complejidad y ciencias sociales (pp. 19–101). Sevilla: Universidad Internacional de Andalucía. Solé, R. V., & Bascompte, J. (2006). Self-organization in complex systems. Princeton, N.J.: Princeton University Press. Strubell, M. & Boix-Fuster, E. (Eds.) (2011). Democratic Policies for Language Revitalisation: The Case of Catalan. Basingstoke, U.K.: Palgrave Macmillan. Terborg, R., & García-Landa, L. (2011). Muerte y vitalidad de lenguas indígenas y las presiones sobre sus hablantes. México: UNAM. Terborg, R., & García-Landa, L. (2013). The ecology of pressures: Towards a tool to analyze the complex process of language shift and maintenance. In À. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society understanding complex Systems (pp. 219–239). Berlin: Springer. https://doi.org/10.1007/978-3-642-32817-6_14. Turell, M. T. (2001). Multilingualism in Spain. Clevedon, UK: Multilingual Matters. https://doi. org/10.1017/s0047404502270293. Vila i Moreno, F. X. (2011). Language in education policies. In M. Strubell i Trueta, & E. BoixFuster (Eds.), Democratic policies for language revitalisation. The case of Catalan (pp. 119–149). Basingstoke: Palgrave Macmillan. Vila i Moreno, F. X. (2012). Survival and development of language communities. Prospects and challenges. Clevedon, UK: Multilingual Matters. https://doi.org/10.1075/lplp.38.1.08nue. Wallerstein, I. (1996). Ouvrir les sciences sociales. Rapport de la Commission Gulbenkian. Paris: Descartes & Cie. Weinreich, U. (1968). Languages in contact. The Hague: Mouton Publications. Wiener, N. (1948). Cybernetics or control and communication in the animal and the machine. Cambridge, Massachusetts: The MIT Press. Williams, C. (2005). The case of Welsh/Cymraeg in Wales. In D. Ó. Néill (Ed.), Rebuilding the Celtic languages: Reversing language shift in the Celtic countries (pp. 35–115). Talybont: Y Lolfa Cyf. Wolfram, S. (1983). Statistical mechanics of cellular automata. Reviews of Modern Physics, 55(3), 601–644. https://doi.org/10.1103/revmodphys.55.601. Wolfram, S. (2002). A new kind of science. Champaign: Wolfram Media.

Chapter 9

Patterns of Linguistic Diffusion in Space and Time: The Case of Mazatec Jean Léo Léonard, Marco Patriarca, Els Heinsalu, Kiran Sharma and Anirban Chakraborti

Abstract Complexity theory is a major interdisciplinary paradigm which provides a unified framework for natural and social sciences. At an operative level, it is based on a combined application of quantitative and qualitative methods at various phases of research, from observations to modeling and simulation, to improve the interpretation of complex phenomena (Anderson in Science 177:393–396 1972; Ross and Arkin in Proc Natl Acad Sci 106:6433–6434 2009). Among the many applications, ranging from physics to biology and the social sciences, the study of language through the methods of complexity theory has become an attractive and promising field of research. In this contribution we consider the complex and interesting case of the Mazatec dialects, an endangered Otomanguean language spoken in south-east Mexico by about 220,000 speakers (SSDSH 2011–2016; Gudschinsky 1955, 1958).

List of Abbreviations AY CQ DO

Ayautla Chiquihuitlán Santo Domingo

J. L. Léonard (B) Sorbonne Université, STIH, EA 4509 Paris, France e-mail: [email protected] M. Patriarca · E. Heinsalu National Institute of Chemical Physics and Biophysics, Tallinn, Estonia e-mail: [email protected] E. Heinsalu e-mail: [email protected] K. Sharma · A. Chakraborti School of Computational and Integrative Sciences, Jawaharlal Nehru University, New Delhi, India e-mail: [email protected] A. Chakraborti e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_9

139

140

HU IX JA JI LO MG SMt SO TE

J. L. Léonard et al.

Huautla San Pedro Ixcatlán Jalapa (Santa Maria) Jiotes San Lorenzo San Miguel Huautla San Mateo Yoloxochitlán San Miguel Soyaltepec San Jerónimo Tecoatl (abréviations reprise de Kirk 1966).

9.1 Introduction 9.1.1 General Method By now, Language Dynamics has come to represent a relevant branch of complexity theory which investigates the classical problems arising in the study of language through novel approaches. Several methods have been imported directly from various scientific disciplines and used to model language from different points of view and at different levels of detail, which complete each other, providing, an altogether new and informative picture (Wichmann 2008). In the wide spectrum of these methods, one finds dynamical models such as (i) Simple models described by ordinary (stochastic) differential equations, addressing the language dynamics of population sizes at a macro- or mesoscopic scale, as in the ecological modeling a la Lotka-Volterra (Heinsalu et al. 2014), and which are able to tackle delicate issues such as the perceived status of languages (which directly affect the one-to-one language interaction between individuals) and describe other social features; (ii) Models describing both space and time evolution, formulated with e.g. reactiondiffusion equations, which allow meso-scopic level investigation of the most different issues and effects, related to e.g. population dynamics, the spreading in space of linguistic features on the underlying physical, economical and political geography (Patriarca et al. 2009); (iii) Individual-based models, certainly a type of model providing a most detailed level of description at the microscopic level, which are used to perform refined numerical experiments to study languages e.g. along the perspective of language evolution (Steels 2011) or language competition (i.e., the dynamics of language use in multilingual communities, see Sole et al. 2010; San Miguel et al. 2005). The latter topic is deeply linked to social interactions, thus the models used have direct connections with social sciences and social dynamics. In fact, linguistic features can be considered to be cultural traits of a specific

9 Patterns of Linguistic Diffusion in Space and Time …

141

nature and their propagation can be modeled similarly to cultural spread and opinion dynamics processes (Castellano et al. 2009; San Miguel et al. 2005).

9.1.2 Application to Mazatec Dialects The Mazatec dialect continuum is located in south-east Mexico. The approximate population of 220,000 speakers is characterized by a highly heterogeneous culture and a locally diversified economic production landscape. The Mazatec dialects have become a classical topic in dialectology due to the fact that they offer a typical highly complex panorama usually observed when studying cultural landscapes, in particular those characterizing endangered languages (SSDSH 2011–2016; Gudschinsky 1955, 1958, 1959; Kirk 1966; Jamieson 1988, 1996; Léonard et al. 2012; Léonard and Kihm 2014). In this paper an analysis of the Mazatec dialects and in particular their mutual linguistic distances will be carried out, relying on previous as well as more recent databases and data analyses by various field-linguists. Such results will be reanalyzed and visualized using the tools of Complex Network Theory, which provide measurement and visualization of features both homogeneous and heterogeneous. Different types of data will be considered, such as those related to average linguistic Levenshtein distance (henceforth, LD) between dialects (Heeringa et al. 2003; Bolognesi et al. 2002; Beijering et al. 2008) or those extracted by a direct comparison between speakers, i.e., based on the mutual intelligibility of dialects (Kirk 1970). In Sect. 9.2, relying on knowledge of the system (and in particular of the values of its main parameters) gained by work carried out thus far (Kirk’s comparative phonological database for interdialectal surveys and fieldwork), we will take into account external constraints. These will include such concerns as the ecology of settlement settings throughout the threefold layered system of Lowlands, Midlands and Highlands, as well as the more recently superposed social and economic impact of postcolonial agro-industrial systems, such as coffee, cattle breeding and sugarcane (all related e.g. to the agricultural use of the land). In Sect. 9.3, a comparison between the image suggested by the complex network analysis of the various data sets (overall sample of lexical categories versus a noun database, restricted to phonological analysis) and other relevant aspects of the system under study will be carried out. This includes comparison of linguistic networks with underlying road networks, physical geography, and economical geography. We will oppose materiality, such as ecological settings, to constructs, such as dialect areas in order to account for the evolution of a very intricate diasystem. These will feed into a set of proposals for diasystemic geometry as a component of Language Dynamics, further cementing it as a promising field for Complexity Theory.

142

J. L. Léonard et al.

9.2 Language Ecology 9.2.1 Ecological Settings Mazatec has resisted assimilation in the long term, thanks to its demographic weight (more than 200,000 speakers) and to emerging language engineering for literature and education through modern spelling conventions yet remains a very vulnerable language. The data collected in the ALMaz (A Linguistic Atlas of Mazatec; see Léonard et al. 2012) support this pessimistic impression, and also when considering the collapse of the more recent agrarian systems of café crops and cooperatives, the continuing consequences of the Miguel Alemán dam in the 1950s and a constant drive to migrate to urban centers such as Tuxtepec, Tehuacán, Oaxaca, Puebla, México DF, or the USA. The Mazatec area stands in the very center of the Papaloapam Basin, benefiting from a smooth transition between the plain (e.g. Jalapa de Diaz) and the mountains, west of the Miguel Alemán dam. This ecologically strategic position turned out to be fatal to the Mazatec Lowlands, partly drowned by the Miguel Alemán dam in the mid-50s, when the Rio Tonto, a powerful river connected to the Papaloapam mainstream, was controlled for the benefit of beverage and hydroelectric companies. Sugar cane also demands much water for crops. Patterns of cross-regional integration which had quietly evolved since Olmec times (Killion et al. 2001) were disrupted in one of the few regions where native peasants (Mazatec and Chinantec mostly) worked their own microfundio. Maps in Figs. 4.1–4.3 enumerate the Mazatec municipalities from Baja to Alta Mazateca (Lowlands and Highlands), providing an explicit view of the landscape: to the east, a plain half drowned by the dam (the Lowlands), to the west, the high Sierra mountain chain divided in the south by a canyon—the Cuicatlán Canyon, with the small Mazatec town of Chiquihuitlán, made famous in Jamieson’s grammar and dictionary, published by the SIL in the late 80s and mid-90s (Jamieson 1988, 1996). Figure 9.1 provides an orographic and hydrographic map of the Mazatec area. Figure 9.2 shows the distribution of Municipios over the Mazatec area—the shape of the spots on the maps in Figs. 9.1 and 9.2, according to the new numeration of figures in these proofs reveals demographic sizes of each town, whereas Fig. 9.3 points out the municipios visited for the ALMaz since 2010. In this map, only localities already surveyed by Paul Livingston Kirk are mentioned, showing how the ALMaz network is intended to be much larger than in previous dialectological studies, as (Kirk 1966; Gudschinsky 1958, 1959). The Mazatec diasystem (Popolocan, Eastern Otomanguean) can be divided into two main zones: the Highlands and the Lowlands. Other subzones can be further distinguished, such as the Midlands (Jalapa de Diaz, Santo Domingo, San Pedro Ixcatlán) within the Lowlands, the Cuicatlán Canyon (Chiquihuitlán) and the Puebla area (see San Lorenzo data below). In short, main dialect subdivisions read as follows, slightly modified from (Léonard et al. 2016):

9 Patterns of Linguistic Diffusion in Space and Time …

143

Fig. 9.1 The Mazatec dialect network (localities surveyed in Kirk 1966). Maps CELE (Vittorio dell’Aquila 2014)

(1) The Mazatec diasystem: dialects and subdialects Highland complex: Central Highlands (Huautla de Jiménez, Santa Maria Jiotes, San Miguel Huehuetlán).1 Northwestern Highlands. Central Northwestern Highlands (San Pedro Ocopetatillo, San Jeronimo Tecoatl, San Lucas Zoquiapam, Santa Cruz Acatepec, San Antonio Eloxochitlán). Peripheral Northwestern Highlands (San Lorenzo Cuaunecuiltitla, Santa Ana Ateixtlahuaca, San Francisco Huehuetlán). Lowland complex: Eastern Lowlands (San Miguel Soyaltepec). Central Lowlands (San Pedro Ixcatlán). Piedmont or Midlands (Ayautla, San Felipe Jalapa de Diaz, Santo Domingo). Periphery: South-Western Highlands (Mazatlán Villa de Flores). Cuicatlán Canyon (Chiquihuitlán).

1 For more details, see http://www.inali.gob.mx/clin-inali/html/v_mazateco.html for a complete list

of population centers.

144

J. L. Léonard et al.

Fig. 9.2 The Mazatec dialect network (localities surveyed in Kirk 1966). Maps CELE (Vittorio dell’Aquila 2014)

It should be kept in mind that such a classification is not exhaustive and provides only a heuristic framework to observe variation. The spots on the map cluster into significant subareas. Behind the dam stands San Miguel Soyaltepec, a very important center from ancient times, which was probably connected through the plains to the coastal zone of the Papaloapam Basin. As the size of the spots in Fig. 9.3 reveal demographic weight, we can state that it is still the biggest urban center in the Mazatec lands. The town of Acatlán, north of Soyaltepec, is more Spanish speaking than Soyaltepec. Inhabitants of the archipelago inside the artificial lake—within the huge pool created by the dam—use the same variety as in San Miguel Soyaltepec, as do the new settlements, such as Nuevo Pescadito de Abajo Segundo, in the South. A dialect network probably as intricate as that of the North-West Highlands (around San Jerónimo Tecoatl) likely existed before the flooding of the microfundio agrarian society of the Lowlands. Most of these dialects merged into mixed dialects, apparently under the strong influence of the Soyaltepec koinè (we use this term as “local speech standard”, i.e. pointing at an oral, more than a written koinè, though nowadays a Soyaltepec written koinè does exist, strongly supported by local poets and school teachers). This first segment of the Mazatec world makes up the San Miguel Soyaltepec Lowlands segment: a resilient area, with a strong urban constellation going from the newly built Temascal to the industrial town of Tuxtepec, with strong local dialect intercourse and mingling, in a region whose agrarian structure was essentially drowned

9 Patterns of Linguistic Diffusion in Space and Time …

145

Fig. 9.3 Communal aggregates in the Mazatec area. Map CELE (Vittorio dell’Aquila 2014). Official census data (2002)

by a pharaonic dam project sixty years ago. The long-term consequences of this dramatic redistribution of agrarian resources and property, and of the displacement of over 22,000 peasants, are still to be seen. Linguistically, this event partially enhanced acculturation and assimilation to Spanish under the influence of urban centers such as SM Soyaltepec, but most of all, Temascal, Acatlán, and Tuxtepec. The second area, moving from Lowlands to Highlands, covers the western shores of the Miguel Alemán lake, as a twofold stripe, from S. M. Chilchotla and San José Independencia (Midlands) to San Pedro Ixcatlán (Western Lowlands), in the continuity of the plain or the valley, where the important urban center of Jalapa de Diaz is located. This Midland-Lowland region displays a whole range of small urban centers, dominated by sugar-cane and herding (the agrarian couple caña y ganado). Though we should consider Jalapa de Diaz as a subarea of its own, because of its size and its links with other regions, such as the Highlands (Huautla) and the so called Cañada or Canyon (Chiquihuitlán and beyond), we may lump both subareas together as the Western Plain. The Highlands qualify as the third main area, after the subdivisions of the Lowlands into the SM LL and the Western Plain. In turns, it divides into two subareas: central, with Huautla, and the Western Highlands—a dense network of small urban centers such as San Lucas, San Jerónimo Tecoatl, San Lorenzo, San Francisco Huhuetlán and San Pedro.

146

J. L. Léonard et al.

We shall call the fourth complex “the Cañada Connection”, where the most conspicuous urban center is Mazatlán de Flores, on the periphery of the Canyon, and Chiquihuitlán. This is a region of intense language contact: from Chiquihuitlán downhill through the Canyon, Cuicateco, a Mixtecan language, is spoken. Nowadays, the zone seems to have fallen into the hands of the Narcos, and the road to Chiquihuitlán is no longer an easy trip from Jalapa de Diaz, as the ALMaz staff has experienced in recent years. The dialect of a spot such as Santa Maria Tecomavaca on the western plateau, has scarcely been documented up to now. This despite being not so far from neighboring centers such as Mazatlán or Teotitlán del Camino, and thus forming a subarea on its own in the Canyon region. This may be due to the low rate of Mazatec speakers as compared to the central area of the Mazatec world, and its location on the plateau, with a tropism outward of the Mazatec area (towards Teotitlán del Camino, Tehuacán, etc.). Strikingly enough, the variety spoken in this peripheral area has more to do with the Northwestern Highland dialects than with the neighboring Mazatlán area, pointing at strong resettlement dynamics throughout the Mazatec area, far beyond state of the art knowledge of these phenomena. To us, the main reason lies in the way the coffee economy drained people from the poorest regions of the Midland Outer Belt (Santa Maria Chilchotla, San Mateo Yoloxochitlán), towards the Teotitlán del Camino urban center, where coffee used to be sold to merchants. Though, the San Juan de los Cúes/Santa Maria Tecomavaca still makes up an original dialect of its own, as several varieties apparently migrated there, from the early 19th to the end of the 20th Century. The agrarian ecology of these subzones appears in Fig. 9.5. Next, we will deal with sociolinguistic ecology, exposing certain aspects about linguistic vitality.

9.3 Sociolinguistics: Vitality Zones Areas and subareas can also be defined by the sole criterion of the rate of speakers, as in Fig. 9.4, where the symbols have the following meanings: H mh ml L

high rate of Mazatec speakers (over 75%), mid-high value, i.e. 50% of the population speaking Mazatec, mid-low density of speakers, i.e. 25%, low density, i.e. 0–25%.

At first sight we can see that the core of the Mazatec area still uses the language intensively (H index), whereas the periphery does not (L on the Eastern shore of the dam and in the Canyon). Two pockets have medium scores: ml at San Juan de los Cúes and mh at Chiquihuitlán. Still, official statistics and maps can hardly account for the complexity of the Mazatec area: why does the La Toma dialect, so close to San Juan de los Cues, and not far from Mazatlán Villa de Flores, have more to do with TE and San Antonio Eloxochitlán or San Lucas Zoquiapam (Northwestern Highlands) than with MZ, to which it should resemble? Why and how did San Lucas Zoquiapam, which still

9 Patterns of Linguistic Diffusion in Space and Time …

147

displays many signs of having been in the past a powerful center, come to be considered by TE speakers to show no dialect differences, even as important traits, such as endemic kw > p singleton labialization, serve to make it very special to its area? To what extent was it once a town dialect? Why do small hamlets such as Soyaltitla, close to MZ, sound more like HU than the former? How can the SO dialect speakers be so confident that everyone speaks the same variety in the SO region, when resettlement zones, such as Corral de Piedra or (Nuevo) Pescadito de Abajo Secundo swarm with resilient subdialects of the ecological refugees accommodated after the displacement caused by the dam’s construction (see Schwartz 2016)? We remember in 2013, a school teacher enumerating an endless and very accurate list of villages and hamlets where she had observed various degrees of resilient Mazatec-Spanish bilingualism, in resettlement spots scattered over Veracruz, where people do not necessarily declare themselves as Mazatec speakers, or even as bilinguals. Snapshots of varieties we were able to elicit through fieldwork in 2010 and 2013 around the Miguel Alemán dam suggest that most of the drowned varieties spoken before displacement are still alive, though mingled with IX, SO proper and other varieties like San José Independencia (Fig. 9.5).

9.4 Dialect Dynamics: A Study in Miniature The title of this section takes over the subtitle of a seminal paper on Mazatec ethnohistory by Sarah Gudschinsky (1958), in which Gudschinsky claimed that geolinguistics provided reliable clues from which one could infer how Proto-Mazatec evolved into five main dialects, through seven periods:

Fig. 9.4 Urban centers and degrees of vitality of Mazatec (source idem)

148

J. L. Léonard et al.

Fig. 9.5 Ecological and agrarian zones in the Mazatec area (source idem)

(2) Gudschinsky’s 1958 dialect dynamics model (A) homogeneity (B) slip according to alternating *a and *u (C) emergence of a Lowlands dialect, to which Mazatlán (MZ) and San Miguel Huautepec (MG) still belonged in 1958—whereas the former is nowadays a peripheral Highlands dialect, the latter strongly clustering with Huautla (HU) in the Central Highlands area (D) the Valley dialect emerges (Jalapa, i.e. JA) and differs from MG, then the Southern Valley dialects splits from a Northern one, while ‘foreign domination’ (Mixtec) takes hold of the region (E) the Highlands dialect emerges, and attracts MZ to its circle of influence, roughly during the period 1300 to 1456 two kingdoms compete, in the Highlands and the Lowlands respectively (F) Western Highlands, MG and Northern Lowlands dialects differ, and Aztec rule takes hold. A more cautious model without so many details on the Mixtec and Aztec hegemonies was proposed previously by the same author (Gudschinsky 1955) and describes five differentiation periods (or phases):

9 Patterns of Linguistic Diffusion in Space and Time … Table 9.1 Mazatec locations Mazatec locations Label Long. Latitude

149

Name

AY

−96.66757

18.03021

San Bartolome Ayautla

CQ

−96.74472

17.99583

San Juan Chiquihuitlan

DO

−96.64758

17.99583

Santo Domingo

HU IX

−96.84314 −96.51017

18.13109 18.14438

Huautla de Jimenez Ixcatlan (San Pedro)

JA

−96.53542

18.06993

Jalapa de Diaz (San Felipe)

AS/JI

−96.81940

18.10808

Santa Maria (de) La Asuncion

LO MG

−96.91334 −96.79694

18.20556 18.10111

San Lorenzo Cuaunecuiltitla Hualtepec (San Miguel Huautla)

MZ

−96.91310

18.03369

Mazatlan Villa de Flores (San Cristobal)

SO

−96.48330

18.20358

Isola del Viejo Soyatepec

TE

−96.91430

18.16717

San Jeronimo Tecoatl

(3) Gudschinsky’s 1955 dialect dynamics model (I) (II) (IIIa) (IIIb)

Homogeneity, followed by the rise of HU and JA Emergence of a transitional buffer zone between HU & JU (The lowland zone splits in two, with the emerging variety of IX Both HU and IX areas diversify: SMt (San Mateo) emerges in the highlands, whereas SO splits from IX. In the buffer zone, MG also emerges. Flows of lexicon and variables still pass from the Lowlands to the Highlands (IV) Further and more clear-cut differentiation between IX and SO, in the Lowlands (V) Consolidation of the six dialects: sharper frontiers.

In the next sections, where the LD is applied to Kirk’s data on twelve varieties (Kirk 1966) for surveying dialect dynamics, Gudschinsky’s models as summarized in (2) and (3) above are very useful to interpret the results. Application of LD suggests a better overall agreement with Gudschinsky’s model (3)—rather than with (2). The twelve Mazatec dialects considered and the corresponding labels and locations are listed in Table 9.1 (old names of location are in parentheses).

9.5 Levenshtein Distance and Its Application In this section some analysis of linguistic databases will be done and discussed. It is useful first to make some observations about the LD and how it can be applied to a comparison of dialects.

150

J. L. Léonard et al.

In its general definition, the LD is a measure of how different two strings are from each other. It can be used, as discussed below, not only for comparing two strings (or words) but also more complex linguistic systems, such as dialects, through a suitably defined average linguistic distance, constructed in some prescribed way from the set of strings to be compared, associated to the entries of a dialect. Importantly, it can be used not only as an estimate of linguistic distance between two given strings, but also between e.g. two uttered words, also when a writing system is missing in the language, by using as strings to be compared with each other the phonetic transcriptions of the words e.g. in IPA. We start from its definition: the LD L(a, b) measures the level of difference between two strings a and b as the minimum number of operations, represented by insertions, deletions, or editions, needed to turn a into b or vice versa. As an example of calculation, let us consider the phonetic transcriptions of the words meaning “arm” in the AY (San Bartolome Ayautla) and in the AS/JI (Santa Maria de La Asuncion) Mazatec dialects. Then the strings in IPA are a  “thia” (AY) and b  “tùha” (JI), respectively. The corresponding LD between a and b is L(a, b)  2, since two changes, h ↔ ù and i ↔ h, are needed in order to turn one string into the other. The LD L(a, b) represents an extremely useful and popular measure of linguistic distance in text processing, having the merit of being simple both in definition and use. Its simplicity, however, also represents its limit, since it cannot resolve any feature of words unless it is accompanied by a set of suitably designed conditions. In fact, it is independent of the type of the actual operations (whether insertions, deletions, or editions), the number and type of characters changed (e.g. vowels or consonants), and of the order in which they are changed. In this work, for the sake of simplicity, the database analyses will be carried out by using the basic version of Levenshtein Distance. If we label by k the semantic meanings covered by the database to be analyzed and represent the IPA strings of the variants of the same semantic meaning k in different dialects i and j as aki and akj , respectively, their Levenshtein Distance will be denoted as   L i,k j  L aik , a kj . Here the integers i and j label the locations/dialects and can assume the values in the interval (1, N L ), where N L is the total number of locations (dialects). For a given pair of locations i and j the label k is a number in the interval (1, M i,j ), where M i,j is the number of pairs of nouns aki and akj in dialects i and j sharing the same semantic meaning. Levenshtein matrix from arithmetic (non-weighted) averages. An estimate of the linguistic distance between two dialects i and j is provided by the corresponding average Levenshtein distance L i,j , L i, j

Mi, j 1  k  L . Mi, j k1 i, j

9 Patterns of Linguistic Diffusion in Space and Time …

151

This is a simple arithmetic average over M i,j terms, meaning that all the distances to be considered, L ki,j , are assumed to have equivalent statistical weights. The analyses presented below in Sect. 3.1.2 employs this simple type of average distance. If the average LDs L i,j are computed for all pairs of dialects (i, j), their values can be collected in a matrix, that will be referred to in the following as the “Levenshtein matrix”, L  {L i,j }. From the definition of average LD L i,j it follows that the Levenshtein matrix is symmetrical and the diagonal elements are equal to zero, L i,j  0. Therefore there are fewer than N 2L independent elements in L. For example, for the N L  12 Mazatec dialects studied below, the corresponding Levenshtein matrix, visualized in Table 9.4, contains N L (N L − 1)/2  60 independent distances. Notice that for many different reasons M i,j, the number of pairs of nouns in dialects i and j sharing the same semantic meaning, will depend on i and j, i.e., it is in general different for different pairs of dialects (i, j). This situation could make the value of the average LDs, {L i,j }, unreliable. To see why and how this could happen, consider as a simple example the case in which for some reason a given meaning k was investigated only for one pair of dialects, e.g. for dialect i  1 and j  2, but not for the others. Yet, it is possible to compute the corresponding distance L k1,2 between the variants and it can give a positive contribution to the average LD, L 1,2, if it is included in the formula above. However, the contribution of L k1,2 cannot have any statistical meaning in this particular case because it is not possible to compare it with the analogous distances between other dialects. The set of average LDs {L i,j } between dialects is statistically most meaningful if all the L i,j receive contributions for the same set of meanings k. If L k1,2 is added to L 1,2 but no analogous quantity is added to e.g. L 1,3 being unavailable, then it is actually equivalent to assume that L k1,3  0 or at least that L k1,3 < L k1,2 , which may not be true.

9.6 Levenshtein Matrix from Weighted Averages In fact, the Levenshtein matrix L  {L i,j } can also be constructed using a more general definition for the average LDs L i,j between dialects i and j, which are obtained not as an arithmetic but as a weighted average of the LDs between the single strings L i,j k associated to the dialects i and j. Also this type of analysis will be used below in Sect. 3.1.1. There can be many reasons for weighting the terms in the sum with some quantity pk . Usually, the assignment of different weights pk to the distances between strings L i,j k is done on the base of linguistic considerations. The formula for the (weighted) average is now  Mi, j L i, j 

k k1 pk L i, j  Mi, j k  1 pk

152

J. L. Léonard et al.

in which each term k is weighted with a corresponding statistical weight pk. The weights themselves {pk } can have any values—e.g. they can represent rank values. Once the average distances are constructed from the weighted averages of the LDs, the way to use the Levenshtein matrix is the same. Visualization. A useful way to study the data analyzed and represented in a Levenshtein matrix L is to visualize the matrix through an evolving network which changes as a function of a threshold T . To this aim, one first normalizes all the elements of the Levenshtein matrix L, i.e., the Levenshtein distances {L i,j }, by dividing them by the largest average LD found L MAX  max{L i,j }: L i,j → L i,j /L MAX. The normalized LD values {L i,j } will be in the interval (0,1), the value L i,j  0 corresponding to perfectly equivalent dialects i and j (as along the matrix diagonal where the elements represent the LD of a dialect from itself), and the value L i,j  1 corresponding to the pair of dialects with the largest LD. Then one sets a certain threshold T on the LDs. Finally, one plots the network with such a threshold T , in which only nodes and links are present, associated to a mutual LD L i,j lower than the assigned threshold, L i, j < T. Starting from T  0, no dialect node is shown yet, because no dialect is perfectly equal to another dialect. Gradually increasing T leads to the appearance of some dialect nodes and the linguistic network begins to form. When reaching the maximum value T  1, all the dialect nodes will have appeared and will be connected to each other. However, even for T  1 there can still be a lot of information displayed in such a network, because not all the strengths of the connections are equal. It is useful to plot the network links with different thicknesses (and possibly colors) corresponding to the relative value, i.e., the thicker the link, the smaller the corresponding LD. In this way even the image of such a fully connected network obtained for a LD threshold T  1 can convey in a clear and intuitive visualization the actual structure of the dialect linguistic network.

9.6.1 An Overall Sample for LD In this section, dialectological data from Kirk (1966) will be measured according to Levenshtein’s distance. As this algorithm measures and ponders distance between dialects synchronically, most of the results rely upon phonological and morphological patterns. Etyma are not used, contrary to a phylogenetic approach. We will thus consider these results as highlighting ontological distances and complexity between dialects (e.g. the most complex dialect here is LO, in the Poblano area, in the NW outskirts of the Mazatec dialect network)—see Table 9.2.

0 0.28

0.20 0.32 0.21 0.24 0.30 0.52 0.29 0.27 0.24 0.29

AY CQ

DO HU IX JA JI LO MG MZ SO TE

AY

0.30 0.38 0.30 0.33 0.37 0.54 0.34 0.35 0.30 0.34

0.28 0

CQ

0 0.33 0.19 0.11 0.33 0.54 0.27 0.26 0.24 0.28

0.20 0.30

DO

0.33 0 0.32 0.30 0.21 0.53 0.25 0.30 0.24 0.33

0.32 0.38

HU

0.19 0.32 0 0.22 0.31 0.53 0.29 0.27 0.24 0.25

0.21 0.30

IX

0.11 0.30 0.22 0 0.32 0.55 0.28 0.28 0.25 0.28

0.24 0.33

JA

Table 9.2 Matrix of Levenshtein distances for 12 Mazatec dialects—117 cognates JI

0.33 0.21 0.31 0.32 0 0.55 0.33 0.28 0.24 0.28

0.30 0.37

LO

0.54 0.53 0.53 0.55 0.55 0 0.55 0.33 0.50 0.50

0.52 0.54

MG

0.27 0.25 0.29 0.28 0.33 0.55 0 0.25 0.24 0.31

0.29 0.34

MZ

0.26 0.30 0.27 0.28 0.28 0.33 0.25 0 0.22 0.29

0.27 0.35

SO

0.24 0.24 0.24 0.25 0.24 0.50 0.24 0.22 0 0.26

0.24 0.30

TE

0.28 0.33 0.25 0.28 0.28 0.50 0.31 0.29 0.26 0

0.29 0.34

9 Patterns of Linguistic Diffusion in Space and Time … 153

154

J. L. Léonard et al.

Fig. 9.6 Dialect network with Levenshtein distance threshold T  0.20. Map Marco Patriarca and co-authors

Proceeding as explained above, when reaching the threshold value T  0.20, there appears a choreme (a kernel area, see Goebl 1998: 555). The bolder line uniting JA and DO points at a dialect of its own, whereas the finer line, between DO and IX, resorts to a less organic structural relation, though rather strong—i.e. a chain, between this basic choreme [JA-DO] with the more autonomous and powerful Lowlands dialect of San Pedro Icxatlán (Fig. 9.6). With the threshold T  0.22, another choreme shows up, in the Highlands: HU and JI, whereas the inner cohesion within the [IX[DO-JA]] chain is confirmed. This [HU-JI] choreme will soon be connected to the most peripheral dialect, in the Eastern Lowlands (SO), and remains yet unconnected to close neighbors like MG or TE. As we soon shall see, these two choremes now available will soon raise their interconnectivity in the dialect network, enhancing patterns of resilience of a previous feature pool consistent in the valley (Fig. 9.7). With threshold T  0.24, a complex communal aggregate [[MZ-SO], [HU-JI], [[IX[DO-JA]]] emerges. The pattern now points at two clusters [HU-JI], [[IX[DOJA]] and one far distant chain [MZ-SO]. As a matter of fact, all these patterns confirm Gudschinsky’s model (1955), initially elaborated out of lexicostatistics (Fig. 9.8). With T = 0.27, though, the overall picture becomes far clearer, and goes far beyond Gudschinsky’s expectations, in terms of fine-grained representation of the intricacy of the diasystem: now we have a whole complex network with clear-cut communal aggregates: a [TE[SO[IX]] chain, a [HU-JI-MG[SO]] chain, a macrochain connecting in a most intricate way MZ with the [IX-DO-JA] chain, through AY and MG, working as areal pivots in the Midland and the Highlands respectively. The most peripheral varieties are LO in the Northwestern fringe, and CQ, in the Southwestern border of the Mazatec area. Interestingly enough, these spots are not connected yet in this phase, forming what we can call “default areas” or “default

9 Patterns of Linguistic Diffusion in Space and Time …

155

Fig. 9.7 As above, T  0.22. Map Marco Patriarca and co-authors

Fig. 9.8 As above, T  0.24. Map Marco Patriarca and co-authors

spots”, i.e. strongly divergent varieties, which do not correlate tightly enough with the rest of the network to highlight deep geolinguistic structures. Of course, one can cluster these erratic varieties by elevating the threshold of divergence (Fig. 9.9). The threshold T = 0.29 shows how CQ does correlate with already available clusters—namely, with AY. Nevertheless, AY and CQ strongly differ in all respects, as our own fieldwork recently gave us evidence. The reason why CQ converges somewhat with AY is due more to the transitional status of AY, between the Highlands and the Lowlands, rather than to structural heritage, although indeed, these two variants can be seen as geographical neighbors (Fig. 9.10).

156

J. L. Léonard et al.

Fig. 9.9 As above, T  0.27. Map Marco Patriarca and co-authors

Fig. 9.10 As above, T  0.29. Map Marco Patriarca and co-authors

The same could be said of LO, as compared to TE: the former finally connects to the latter in a nearest-neighbor graph, as in Fig. 9.11, although the structural discrepancy is conspicuous. The nearest-neighbor graph is obtained by joining each dialect node only to the one from which it is at the shortest LD, thus providing an approximate but clear indication of the most similar dialect in the network and, in the end, an intuitive idea of the core structure of the dialect network. Indeed, LO proceeds from the same historical matrix as TE: the San Antonio Eloxochitlán dialect—not surveyed by Paul Livingston Kirk, but from where we were able to elicit phonolog-

9 Patterns of Linguistic Diffusion in Space and Time …

157

Fig. 9.11 Nearest neighbor map from LD, 117 cognates (data Kirk 1966). Map Marco Patriarca and co-authors

ical and morphological data in 2011. This nearest-neighbor graph below provides a handy overall picture of the Mazatec dialect network, on the basis of the Levenshtein Distance processing of our 117 cognates: it clearly highlights the far reaching interconnection of Highlands dialects with Lowlands dialects, with macro-chains [TE[IX]], [MZ[SO]] and the intricate cross-areal (i.e. Highlands/Lowlands) cluster [HU-JI-MG[SO]]. Lower range clusters, such as [AY[CQ[DO]]], and choremes, such as [DO-JA] and [HU-JI], as seen previously at stage T = 0.20 and T = 0.22 are also available in this map. Considering Gudschinsky’s model of dialect dynamics (3) above, one can now verify to what extent its predictions were right. As a matter of facts, her claim I (homogeneity, followed by the rise of Hu and JA) is confirmed by phase T = 0.22, which clearly enhances the emergence of two choremes—high and low: [HU-JI] versus [DO-JA]. Gudschinsky’s period II entails the emergence of a transitional buffer zone between HU & JU. This claim is strongly supported, but also enriched by phases T = 0.24 and T = 0.27: not only does HU cluster with JI and MG, but AY also clusters with the IX and JA-DO chain. In turn, all these aggregates connect with Lowlands varieties, pointing at the formation of Highlands varieties as a by-product of Lowlands dialect diversification. The ambivalent structural status of MZ, standing far west in the Highlands, though connecting far into the East with SO, and even to IX, through the buffer area of AY, hypothesized by Gudschinsky in both models (2) and (3), is also strongly confirmed. Gudschinsky’s Periods IIIa-b, implying the split of the Lowlands dialect in two (JA vs. IX) on the one hand (IIIa), and on the other hand the inner split of the Highlands (i.e. IIIb: HU vs. TE, standing for Gudschinsky’s SMt, in this dialect network according to Kirk’s data) are also confirmed by steps T = 0.29 and T = 0.30 respectively, as these slots in the graph become more densely interactive

158

J. L. Léonard et al.

with the rest of the dialect network. Though it must be noted that results here display much more detail on general connectivity than in models in (2) and (3). Last, but not least, period VI, with further and more clear-cut differentiation between IX and SO, in the Lowlands, is also confirmed by far reaching patterns of connectivity of SO with TE, HU, MZ in the highlands and AY in the Midlands. Results from this set of 117 cognates (see Léonard 2016: 77–79 for a complete list of items) are not simply congruent with Gudschinsky’s hypothesis on dialect dynamics, as summed up in (2) and (3): they provide much more information about the hierarchization and intricacy of differentiation within the Mazatec dialect network. Moreover, they enhance the status and interplay of such (dia) systemic categories as choremes, chains, macro-chains and pivots or buffer zones. They also clearly point at a level of diasystemic organization which supersedes the Stammbaum and the chain level of organization: distant ties, either out of retention, or as an endemic effect of a feature pool (Mufwene 2001, 2012, 2013) of traits inherited from the Lowlands dialects, which carried on mingling together long after the splitting of the main Highlands and Lowlands dialects. For example, many morphological facts point at an inherited stock of inflectional mechanisms in the Lowlands dialects and peripheral Northwestern dialects such as LO (in Kirk’s data) and San Antonio Eloxochitán (ALMaz data). The link between TE and IX in Fig. 9.11 confirms this trend—whereas the link between HU and SO or MZ and SO may rely more on mere retention, and to an older layer of structural continuity. The sample processed here covered all lexical classes of the Mazatec lexicon, for a set of 117 cognates, from (Kirk 1966): verbs, nouns, pronouns, adjectives, adverbs, etc. The results do provide a useful overall picture, but we still suspect this sample to be too heterogeneous, and to blur finer grained patterns of differentiation within the lexicon and grammar. Verbs are especially tricky in Mazatec, and bias may be induced by elicitation, for instance when the linguist asks for a verb in neutral aspect (equivalent to present tense), and may get an answer in the incompletive (future tense) or completive (past tense), or the progressive aspect, according to pragmatic factors (e.g. verbs such as ‘die’ can hardly be conjugated in the present, as ‘he dies’, and informants are prone to provide completive or incompletive forms, as ‘he died (recently)’ or ‘he’ll (soon) die’). Nouns in Mazatec are far less inflected than verbs—only inalienable nouns, such as body parts and some kinship terms have fusional inflection (see Pike 1948: 103–106). The subset of nouns in the Kirk data base therefore is more likely to provide abundant and much more reliable forms to implement the LD than a sample of all lexical categories.

9.7 A Restricted Sample for LD: Nouns Although this paper aims at modeling dialect dynamics rather than at providing a description of the language, some data may be useful at this point of the argumentation, in order to get a glimpse word structure in Mazatec, and related processes on which the LD distance may apply: see Table 9.3 for a sample of the data.

9 Patterns of Linguistic Diffusion in Space and Time …

159

Table 9.3 Sample of nouns in Kirk’s Mazatec data

Source Data from Kirk 1966, data processing: CELE, Vittorio dell’Aquila 2014

In set A n(d)˜ıP(i)ju ‘ant’, glottalization and second vowel alternation u/i is the main issue at stake (compare token A/12 with the rest of the set, for glottal and A11, with final -i versus all the others, with -(i)u). In set B Ùuntu ‘worm’, alternation of the first vowel -i/u- makes the difference; whereas in set C nkaho/nkihu ‘cave’, both initial and final vowel alternate. These are optimal conditions to run the LD algorithm. For item D nÙut˜ı/Ùut˜ı ‘maize cob’, the prenasalized stop can be dropped, as shows token D5 Ùut˜ı, and initial vowel alternates oral u and nasal I, whereas the final vowel lowers for token D11 nÙit˜e. The phenomenology of item D lahnka /thiunka/nţhanka ‘wing’ is far more intricate because of the interplay of prefixes la-, thiu-, ntsha-, but the tokens still remain quite easy to take into account for LD processing. As (Kirk 1966) only documented tones for JA, HU and SO, this feature has been skipped off from the data base. Another advantage with this sample is the greater quantity of items analyzed: 311 instead of 177. A summary of the database features and composition is given for each dialect and grammatical category in Tables 9.4 and 9.5. As one can see, the database contains almost seven thousands tokens, of which about half in each dialect is represented by nouns. It is the subset of nouns (which is the largest grammatical subcategory and by far the most easily analyzable, directly and without preprocessing) that is used in the analysis discussed below. The results are to be seen in Table 9.3, containing the non-weighted average LD between dialects. Some pair of dialects happen to have very similar average LD’s and in order to rank distances and detect the small differences three significant digits have been maintained in the numerical values of the matrix elements. In Fig. 9.12, we have selected three snapshots, according to three thresholds of differentiation T = 0.45, T = 0.59 and T = 0.72—corresponding to low, mid and high distance respectively.

160

J. L. Léonard et al.

Table 9.4 Grammatical Categories and number of tokens Grammatical Category Number of tokens Adjectives Adverbs Conjunctions Conj.sub.

1129

16.54%

326 18

4.78% 0.26%

12

0.18%

Interjections

9

0.13%

Interrogative part

8

0.12%

3770 9 180 27

55.24% 0.13% 2.64% 0.40%

Pref.compl.

21

0.31%

Pref.cont Pref.pers.

11 20

0.16% 0.29%

3

0.04%

Pron. Pron. 1 pl. Inclusive

53 1

0.78% 0.01%

Pron.pers.

15

0.22%

Pron.rel. v. 2s. Imper.

12 8

0.18% 0.12%

Nouns Plural nouns Num. p.p.

Pref.pers.(plural)

v. 3 sg.

18

0.26%

v. future 3 sg.

11

0.16%

v. imper. 1 pl.

9

0.13%

8 375

0.12% 5.49%

17

0.25%

v. inf. vi. 3 sg. vi. 3 sg. Compl. vi. 3 sg. Past

9

0.13%

vi. future 1 sg./3sg.

9

0.13%

vi. impen 2 sg.

8

0.12%

vt 2 sg.

9

0.13%

vt3 pl.

13

0.19%

10

0.15%

vt. 3 sg (= vt.3 sg?) vt. 3 sg vt. 3 sg + obj vt. 3s. vt. subjunctive 3 sg. Total

671

9.83%

8

0.12%

10 9

0.15% 0.13%

6825

100.00%

9 Patterns of Linguistic Diffusion in Space and Time … Table 9.5 Percentage of nouns in the total number of tokens

Location AY CQ DO HU IX JA AS/J l LO MG MZ SO TE Total

161

Total number of Tokens Nouns 651 538

355 297

55% 55%

706 713 661 767 314 583 357 407 510 618 6825

347 380 368 418 184 321 207 239 282 333 3770

49% 53% 56% 54% 59% 55% 58% 59% 55% 54% 55%

Fig. 9.12 LD applied to nouns in Kirk’s data. Three thresholds of normalized mean distance T  0.45, T  0.59, T  0.72. Maps Marco Patriarca and co-authors

162

J. L. Léonard et al.

All networks emerging from this wider and more consistent sample confirm previous results: at T = 0.45, we find again two choremes—one located in the Southern Lowlands, i.e. [JA-IX], and another located in the Central Highlands, i.e. [HU-JIMG]. The latter choreme, though makes up a chain with a very interesting dialect, which was already viewed as ambivalent by Gudschinsky: MZ clusters with [HUJI-MG] in a [MZ[HU-JI-MG]] chain (Table 9.6). The main difference with previous clusters at this stage lays in the boldness of aggregates: MZ would be expected to cluster at a later stage of structural identification with the Highlands choreme, and JA should cluster first with DO instead of telescoping IX. This particular behaviorial pattern of the diasystem is due to the lesser complexity of the data, as suggested above when analyzing phonological variables in Table 9.2: the simpler the morphological patterns, the more straightforward the results. Bolder chains in Fig. 9.12 give therefore more clear-cut hints at the deep structure of the diasystem. When proceeding as explained above at level T = 0.59, an overt extensive rhombus appears, crossing the whole area from west to east, strongly rooted in MZ in the West and SO in the East, with two lateral extensions: TE in the Northwest and AY in the East. One couldn’t dream of a better summary of most of our previous observations: TE and AY are outstanding actors as pivots, or transitional spots; while MZ, HU and SO had already been noted as crucial innovative dialects in the early phases of Gudschinsk’s models of differentiation—stages C and D in (2) and stage IIIa in (3). At T = 0.72, a trapezoid resorting more to a parallelogram than to an isosceles takes form, confirming the far reaching links between TE and IX, going all the way down towards AY and CQ, then climbing up toward MZ and reaching TE in a loop—this geometry actually comprehends the periphery of the diasystem, and may point at a deeper level of structuration. The MST diagram in Fig. 9.13 provides further information. A minimum spanning tree is a spanning tree of a connected, undirected graph such that all the N (here N = 12) nodes are connected together with the minimal total weighting for its (N − 1) edges (the total distance is minimum). The distance matrix, defined by Leveinshtein distances among the dialects, was used as an input to the inbuilt MST function in MATLAB (See Matlab documentation for details). The MST can be done using the Kruskal or the Prim algorithms, summarized here below: Kruskal. This algorithm extends the minimum spanning tree by one edge at every discrete time interval by finding an edge which links two separate trees in a spreading forest of growing minimum spanning trees. Prim. This algorithm extends the minimum spanning tree by one edge at every discrete time interval by adding a minimal edge which links a node in the growing minimum spanning tree with one other remaining node. Here, we have used Prim’s algorithm to generate a minimum spanning tree. The MST on Fig. 9.13 endows the Central Highlands dialect JI with enhanced centrality. The fact that the transitional variety of AY in the Midlands is intertwined with another “buffer zone” dialect, according to Gudschinsky’s model, confirms details of the deep structure of the dialect network.

AY

0.000 0.632

0.629 0.668 0.606 0.607 0.636 0.981 0.562 0.573 0.582 0.708

#

AY CQ

DO HU IX JA JI LO MG MZ SO TE

0.717 0.703 0.666 0.704 0.589 0.978 0.627 0.645 0.636 0.688

0.632 0.000

CQ

0.000 0.689 0.585 0.334 0.643 1.000 0.608 0.639 0.620 0.703

0.629 0.717

DO

0.689 0.000 0.593 0.655 0.346 0.897 0.402 0.481 0.519 0.550

0.668 0.703

HU

0.585 0.593 0.000 0.599 0.616 0.937 0.574 0.639 0.519 0.586

0.606 0.666

IX

0.334 0.655 0.599 0.000 0.617 0.945 0.594 0.604 0.585 0.675

0.607 0.704

JA

Table 9.6 Levenshtein Distance Matrix (LD), data from (Kirk 1966): 311 nouns JI

0.643 0.346 0.616 0.617 0.000 0.841 0.377 0.426 0.462 0.502

0.636 0.589

LO

1.000 0.897 0.937 0.945 0.841 0.000 0.883 0.892 0.884 0.870

0.981 0.978

MG

0.608 0.402 0.574 0.594 0.377 0.883 0.000 0.446 0.490 0.539

0.562 0.627

MZ

0.639 0.481 0.639 0.604 0.426 0.892 0.446 0.000 0.511 0.567

0.573 0.645

SO

0.620 0.519 0.519 0.585 0.462 0.884 0.490 0.511 0.000 0.574

0.582 0.636

TE

0.703 0.550 0.586 0.675 0.502 0.870 0.539 0.567 0.574 0.000

0.708 0.688

9 Patterns of Linguistic Diffusion in Space and Time … 163

164

J. L. Léonard et al.

Fig. 9.13 Minimum spanning tree based on the Levenshtein distance applied to nouns in Kirk’s data

The dendogram in Fig. 9.14 provides an additional piece of information. A dendrogram is basically a tree diagram. This is often used to depict the arrangement of multiple nodes through hierarchical clustering. We have used the inbuilt function in MATLAB to generate the hierarchical binary cluster tree (dendrogram) of the 12 dialects under consideration, connected by many U-shaped lines as shown in Fig. 9.14, such that the height of each U-line represents the average Levenshtein distance between the two connected dialects. Thus, the vertical axis of the tree captures the similarity between different clusters whereas the horizontal axis represents the identity of the objects and clusters. Each joining (fusion) of two clusters is represented on the graph by the splitting of a vertical line into two vertical lines. The vertical position of the split, shown by the short horizontal bar, gives the distance (similarity) between the two clusters. We set the property “Linkage Type” as “Ward’s Minimum Variance”, which requires the Distance Method to be Euclidean resulting in group formation such that the pooled within-group sum of squares would be minimized. In other words, at every iteration, two clusters in the tree are connected such that it results in the least possible increment in the relevant quantity i.e. pooled within-group sum of squares. The dendogram in Fig. 9.14 not only presents an overall picture of the dialect network– it tells us more about the intricacy of communal aggregates and layers of differentiation. It also solves a few problems raised by discrepancies between model (2) and (3) and our results. In this Stammbaum, Highlands dialects actually cluster with Low-

9 Patterns of Linguistic Diffusion in Space and Time …

165

Fig. 9.14 Levenshtein distance applied to nouns in Kirk’s data. Dendogram

lands dialects, while Southern Midlands2 dialects cluster together with a “default” variety—CQ, a near neighbor in the South. In the inner cluster of the Dendogram (including Highlands dialects) we come across the [MZ[HU-JI-MG]] chain we are already familiar with, on the one hand, and on the other a quite heterogeneous subcluster made up of a [IX-SO] chain, associated to the far distant TE Northwestern Highlands dialect, usually classified within the Highlands dialects proper. Last, the LO dialect, though it can be considered a byproduct of a recent Northwestern dialect overdifferentiation (i.e. from TE), does stand on its own, as if it would classify as a totally different language—which it is not, although its differences are indeed phonologically conspicuous, because of a recent vowel shift i > e, e > a, a > o, u > ï.3 In spite of these discrepancies with expected taxon, the main lesson of this dendogram lays in the tripartition [Midlands[Highlands-Lowlands]], and the confirmation of the [MZ[HU-JI-MG]] chain. Figure 9.15 presents a two-dimensional projection of the same data through a multi-dimensional scaling analysis. It mends up the formal oddities we already mentioned, i.e. TE clustering so far from HU, and CQ so close to AY. This representation, obtained with the same date, is far more congruent with standard taxonomy of Mazatec dialects, as in (1) above: it displays a constellation of choremes as [DO-JA] and [JI-HU], and more loosely tightened chains such as [AY[IX]], [MZ[MG[TE]]] and a fairly distant chain [CQ[SO]]. LO, again, stands far apart, as a strongly innovative dialect as far as phonology is concerned—with strong consequences on morphology too. 2 Northern Midland dialects make up a much diversified group, with San José Independencia, Santa

Maria Chilchotla (see Meneses Moreno 2004: 48). According to our ALMaz fieldwork (especially in 2013, see http://axe7.labex-efl.org/node/136), we can assert that both dialects have a strong drive towards the Huautla phonological and inflectional patterns, although the former is somewhat influenced by IX, and the latter is even more strongly similar to HU. 3 stands here for a retracted tongue root high velar vowel, as Japanese/u/.

166

J. L. Léonard et al.

Fig. 9.15 Two-dimensional projection from multi-dimensional scaling analysis (in linguistic space). Nouns in Kirk’s data

Multidimensional scaling is a method to analyze large scale data that displays the structure of similarity in terms of distances, obtained using the Levenshtein distance algorithm, as a geometrical picture or map, where each dialect corresponds to a set of coordinates in a multidimensional space. MDS arranges different dialects in this space according to the strength of the pairwise distances between dialects—two similar dialects are represented by two sets of coordinates that are close to each other and two dialects behaving differently are placed far apart (see Borg 2005) in the space. We used the distance matrix defined using Levenshtein distance to generate N L real vectors x 1 ,…, x NL such that   xi − x j  ≈ di, j , i, j  1, . . . , N L where | … | represents the vector norm. To evaluate the norm one can use the Euclidean distance metric as is with the classical MDS. Effectively, through the MDS one tries to find a mathematical embedding of the NL objects into D-dimensional real space by preserving distances. In general, we choose the embedding dimension D = 2 so that we are able to plot the vectors x i in the form of a map representing all the N L dialects. It may be noted that the x i (and therefore the corresponding representation on a plane) are not unique. Generally, MDS can be obtained through an optimization problem, where (x 1 , …, x i ) is the solution of the problem of minimization of a cost function C, such as

9 Patterns of Linguistic Diffusion in Space and Time …

167

L     iN  xi − x j  − di, j 2 C x1 , . . . , x N L  i1

In order to capture the similarity among the dialects visually, we have generated the MDS plot of 12 dialects. As before, using the International Phonetic Alphabet from the database as an input, we computed the distance matrix using the Levenshtein distance algorithm. The distance matrix was then used as an input to the inbuilt MDS function in MATLAB (see MATLAB documentation). The output of the MDS was the sets of coordinates, which were plotted as the MDS map as shown in Fig. 9.15. The coordinates are plotted in a manner such that the centroid of the map coincides with the origin (0,0).

9.8 Conclusion and Prospects As Nicolaï and Ploog put it (2013: 278), one has to consider two types of categories, when tackling anything which looks like—or is supposed to work as—frontiers: on the one hand, matter or materiality, on the other hand constructs. Matters or materialities rank as follows: geography, geology, biology, ecology, and they partly shape the world we live in, as we are indeed a very adaptive species. Constructs, instead, should be clearly divided in two: compelling patterns on the one hand, elaborations on the other hand. The former ranges from social constraints or norms, laws, beliefs and habits to economic systems; the latter from models to reforms, according to the activities developed in communal aggregates, in reaction to the environment and its contradictions. In this case, matters do matter a lot, as the Mazatec diasystem is vertically structured, from the Lowlands to the Highlands, and some bigger and older centers or town dialects, as JA, HU, MZ, IX indeed weight more than mere villages or hamlets (as JI, MG, AY, CQ, LO). The fact that SO was so peripheral, and ended up as a village nested on top of a resilient hill above the Miguel Aleman dam (along with the village called Viejo Soyaltepec) has consequences on the evolution of certain components of the Mazatec diasystem. The intrusion and the violent reshaping of ecological and socioeconomic settings since the end of the XIXth century, though mercantile in nature, have instead resorted to elaborative constructs, and these have also played a strong role, in smashing previous compelling patterns of intercommunal solidarity or, on the contrary, enmity. Matter and materialities constantly change in nature, indeed, as biology and geology teach us. But cultural constructs change even faster, and they may even loop, recede and proceed, in a nonlinear way—as do diasystems throughout history, and as does the Mazatec diasystem. But the higher plateau or level in the realm of constructivism and elaboration has to be sought in our models and methods to gather and proceed data, as we did here, handling Kirk’s cognate sets, initially collected for a sketch of comparative phonology. We turned it into something quite unexpected, as alchemists used to

168

J. L. Léonard et al.

dream of turning stones or dust into gold. We saw how quantitative tools like the Levenshtein algorithm are designed to measure dialect distance can provide clues from a Complexity Theory standpoint. Various data sets and a variegated array of computational methods (multilayered normalized means, minimum spanning trees, multi-dimensional scaling analysis, etc.) applied on these raw sets of data opened the way to a labyrinth of constructs and representations, which teach us a lot about what mattered in the past, and what matters and will, today and for the future, in the strongly diversified communal aggregates that make up the Mazatec small world (Léonard et al. 2014). A world full of complexity, whose survey with the help of Complexity Theory methods suggest that both tree-models (Stammbaum), chain models, choremes and buffer zones or transitional areas are not sufficient to grasp geolinguistic complexity. We also have to resort to concepts such as pivots, default varieties, and a few more. Neither is the punctuated equilibrium (Dixon 1997) concept enough, as the Mazatec dialect network geometry shows an intricate web of constant interactions.The valley leading from the Lowlands to the Highlands has not only once in a while served as a bottleneck, its seems to be a highway for diffusion and linguistic change which never rests. Corridors from the Northern Midlands, as Santa Maria Chilchotla, and the San José Tenango area, between HU and San José Independencia, may also account for this multisource and multidirectional percolation of change and metatypes between communal aggregates. The intricate geometry of diasystems has still to be disentangled, and this Mazatec case study provides but a glimpse at how to tackle this issue. Complexity Theory undoubtedly should be at the forefront of such a crucial endeavor, for the understanding of how complex adaptive and cooperative systems such as language and society work and mingle together. Acknowledgements Anirban Chakraborti and Kiran Sharma acknowledge the support by the University of Potential Excellence-II grant (Project ID-47) of JNU, New Delhi, and the DST-PURSE grant given to JNU by the Department of Science and Technology, Government of India. Kiran Sharma acknowledges the University Grants Commission (Ministry of Human Research Development, Govt. of India) for her senior research fellowship.

References Anderson, P. W. (1972). More is different. Science, 177, 393–396. Beijering, K., Gooskens, C., & Heeringa, W. (2008). Predicting intelligibility and perceived linguistic distance by means of the Levenshtein algorithm. Linguistics in the Netherlands, 13–24. https://doi.org/10.1075/avt.25.05bei. Bolognesi, R., & Heeringa, W. (2002). De invloed van dominante talen op het lexicon en de fonologie van Sardische dialecten. Gramma/TTT: tijdschrift voor taalwetenschap, 9(1), 45–84. Borg, I., & Groenen, P. (2005). Modern Multidimensional Scaling: Theory and applications. New York: Springer. Castellano, C., Fortunato, S., & Loreto, V. (2009). Statistical physics of social dynamics. Reviews of Modern Physics, 81, 591.

9 Patterns of Linguistic Diffusion in Space and Time …

169

Dixon, R. M. W. (1997). The rise and fall of languages. Cambridge: Cambridge University Press. Goebl, H. (1998). On the nature of tension in dialectal networks. A proposal for interdisciplinary research. In G. Altmann & W. Koch (Eds.), Systems. New paradigms for the human sciences (pp. 549–571). Berlin: Walter de Gruyter. Gudschinsky, S. (1955). Lexico-statistical skewing from dialect borrowing. International Journal of American Linguistics, 21(2), 138–149. Gudschinsky, S. (1958). Mazatec dialect history. Language, 34, 469–481. Gudschinsky, S. (1959). Proto-Popotecan. A Comparative Study of Popolocan and Mixtecan. International Journal of American Linguistics, 25(2). Heeringa, W., & Goosken, C. (2003). Norwegian dialects examined perceptually and acoustically. Computers and the Humanities, 57(3), 293–315. Heinsalu, E., Patriarca, M., & Léonard, J. L. (2014). The role of bilinguals in language competition. Advances in Complex Systems, 17(1), 1450003. https://doi.org/10.1142/S0219525914500039. Jamieson, C. (1996). Diccionario mazateco de Chiquihuitlán. Tucson: SIL. Jamieson, C. (1988). Gramática mazateca. Mazateco de Chuiquihuitlán de Juárez. México: D. F: SIL. Killion, T., & Urcid, J. (2001). The Olmec legacy: Cultural continuity and change in Mexico’s Southern Gulf Coast lowlands. Journal of Field Archaeology, 28(1/2), 3–25. Kirk, P. L. (1966). Proto-Mazatec phonology (Doctoral dissertation). Washington: University of Washington. Kirk, P. L. (1970). Dialect intelligibility testing: The Mazatec study. International Journal of American Linguistics, 36(3), 205–211. Léonard, J. L., dell’Aquila, V., & Gaillard-Corvaglia, A. (2012). The ALMaz (Atlas Lingüístico Mazateco): From geolinguistic data processing to typological traits. STUF—Language Typology and Universals Sprachtypologie und Universalienforschung, 65(1), 78–94. https://doi.org/10. 1524/stuf.2012.0006. Léonard, J. L., & Kihm, A. (2014). Mazatec verb inflection: A revisiting of Pike (1948) and a comparison of six dialects. In J. L. Léonard & A. Kihm (Eds.), Patterns in Mesoamerican morphology (pp. 26–76). France: Hal—CCSD. Léonard, J. L., & dell’Aquila, V. (2014). Mazatec (Popolocan, Eastern Otomanguean) as a Multiplex Sociolinguistic “Small World”. In U. Bereczki (Ed.), The languages of smaller populations: Risks and possibilities. Lectures from the Tallinn conference. Tallin Hungarian Institute’s Series: Miscellanea Hungarica (pp. 27–55). Tallin: Balassi Instituudi & Tallinna Ungari Instituut. Léonard, J. L. (2016). Diversification, diffusion, contact: Modélisation géolinguistique et complexité. Lalies, 36, 9–79. Léonard, J. L., & Fulcrand, J. (2016). Tonal inflection and dialectal variation in Mazatec. In E. Palancar & J. L. Léonard (Eds.), Tone & inflection: New facts and new perspectives. Trends in linguistics. Studies and monographs (Vol. 296, pp. 165–195). Berlin: Mouton de Gruyter. Meneses Moreno, A. B. (2004). Impacto político, social y cultural de la presa Miguel Alemán en la comunidad mazateca de la isla del Viejo Soyaltepec (Master Thesis). Mexico: Universidad Autónoma Metropolitana. Mufwene, S. S. (2001). The ecology of language evolution. Cambridge: Cambridge University Press. Mufwene, S. S. (2012). The emergence of complexity in language: An evolutionary perspective. In Á. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication, and society (pp. 197–218). Berlin: Springer. Mufwene, S. S. (2013). The ecology of language: Some evolutionary perspectives. In E. K. Nakayama Nenoki do Couto, D. B. Albuquerque & G. P. Aráujo (Eds.), Da fonologia à ecolinguística. Ensaios em homenajem a Hildo Honório do Couto (pp. 302–327). Brasilia: Thesaurus. Nicolaï, R., & Ploog, K. (2013). Frontières. Question(s) de frontière(s) et frontière(s) en question: des isoglosses à la “mise en signification du monde. In J. Simonin & S. Wharton (Eds.), Sociolinguistique du contact. Dictionnaire des termes et des concepts (pp. 263–287). Lyon: ENS Editions.

170

J. L. Léonard et al.

Patriarca, M., & Heinsalu, E. (2009). Influence of geography on language competition. Physica A, 388, 174. Pike, K. (1948). Tone languages. A technique for determining the number and types of pitch contrasts in a language, with studies in tonemic substitution and fusion. Ann Arbor: University of Michigan Press. Ross, J., & Arkin, A. P. (2009). Complex systems: From chemistry to systems biology. Proceedings of the National Academy of Sciences, 106, 6433–6434. San Miguel, M., Eguiluz, V. M., Toral, R., & Klemm, K. (2005). Binary and multivariate stochastic models of consensus formation. Journal of Computer Science Engineering, 7, 67–73. Schwartz, D. (2016). Transforming the tropics: Development, displacement, and anthropology in the Papaloapan, Mexico, 1940s–1960s (Doctoral dissertation). University of Chicago. Chicago. Secretaria de Desarrollo Social y Humano (SSDSH). (2011–16). Microrregión 13: Zona Mazateca. Mexico: Author. Solé, R., Corominas-Murtra, B., & Fortuny, J. (2010). Diversity, competition, extinction: The ecophysics of language change. Interface, 7, 1647–1664. Steels, L. (2011). Modeling the cultural evolution of language. Physics of Life Reviews, 8, 339–356. Wichmann, S. (2008). The emerging field of language dynamics. Language and Linguistics Compass, 2(3), 442.

Chapter 10

Common Knowledge in Conversation of Bilinguals and the Ecology of Pressures. The Complex Processes of Using Language and Learning to Coordinate Actions with Other Speakers Roland Terborg and Virna Velázquez

Abstract There seems to be nothing unusual about a conversation between two people who have never been in contact. These people understand each other because they speak the same language. That is commonly believed because the notion of ‘langue’ has always been there when we think about German, Spanish or Catalan languages. The concept seems useful but may be problematic at times. We cannot avoid using this notion to write a language test; nevertheless, the idea limits us, in understanding the process that is taking place in the interaction. The aim of this text is to discuss how individuals develop a physical and social history together with others. By doing it, some pressures involved lead to common actions that modifies the state of the world to which they belong. All the way to seek less effort in their social actions should be coordinated with other actions of the other speakers involved. These pressures are the product of what is called: Upmost Common Routine (UCR) that is in permanent change in relation to the amendments of Direct Common Knowledge (DCK) and Indirect Common Knowledge (ICK). Each common action causes mental changes related, but not identical, to the knowledge of the individuals involved. Their stories are partly shared and partly individual, so a complex network of shared knowledge that gives the illusion that there is a fixed and unchanging knowledge as a real representation of the world is built. These illusions are reinforced by pressures that emerge from the ideologies, values and beliefs. R. Terborg (B) Escuela Nacional de Lenguas, Lingüística y Traducción, Universidad Nacional Autónoma de Mexico, Mexico City, Mexico e-mail: [email protected] V. Velázquez Facultad de Lenguas, Univesidad Autónoma del Estado de Mexico, Toluca, Mexico e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_10

171

172

R. Terborg and V. Velázquez

10.1 Introduction There seems to be nothing unusual about a conversation between two people who have never been in contact. These people understand each other because they speak the same language. That is commonly believed because the idea of “langue” has always been there, even since before Saussure distinguished between “langue”, “langage” and “parole”. But the concept of a default system is also present in the following notions: dialect, code, sociolect, style, code-switching, monolingualism, bilingualism, diglossia and competence. We cannot avoid using this notion to write a language test; nevertheless, the idea limits us, at the same time, in understanding the process that is taking place in interaction. In many ways the notion seems useful but may be also problematic at times. It would not be so problematic to distinguish between two European Languages like Spanish and English with established norms that are widely accepted, but why Norwegian is a different language from Danish or Swiss German a different language from German and Austrian German is not, may provoke a discussion and may be a question of a point of view. But the complex language situation in Mexico, Nigeria, Cameroon, India or Papua New Guinea does not obey to this simplified notion of different languages or dialects. (See Romaine 1994) The same concept is not suitable to describe these different situations. We could probably compare the concept of language with a road that gives us a close look at a landscape; simultaneously, we are also limited to roadsides because the far field is inaccessible to the traveler’s view. One can easily believe that what you see from the road is all that exists throughout the region and can also believe that this road is the only one existing. Roads can be of help to reach the goal and to reconnoiter yet they can be limiting. On the other hand, there are travelers that are not limited to the use of a road but use a combination of several roads, even when there seems to be no connection among them. We can say that many speakers also achieve a combination of languages even with those people who may think these could not be used together. Some use different elements that are not compatible, having a partially common history which allows them to successfully combine what apparently seems to be incompatible. Often code switching as a method of analysis is limited to the analysis of the use of several languages in a conversation. It’s like the road that leads away from any natural phenomenon in the remote countryside. Of course, as just noted, the concepts of language, code, etc. can be very convenient and so the analysis of the alternation of the code can explain many phenomena. What we want here is to present a different point of view which in some situations could be advantageous. We want to avoid notions related to language, code and their alternation, but we do want to emphasize what Varela called bodily and social history and product knowledge, especially when it is shared with others. El verdadero desafío que esta orientación plantea a las CTC (ciencias de tecnologías de la cognición) es que pone en tela de juicio el supuesto más arraigado de nuestra tradición científica: que el mundo tal como lo experimentamos es independiente de quien lo conoce. En cambio, si estamos obligados a concluir que la cognición no se puede entender adecuada-

10 Common Knowledge in Conversation of Bilinguals and the Ecology …

173

mente sin sentido común, el cual no es otra cosa que nuestra historia corporal y social, la inevitable conclusión es que conocedor y conocido, sujeto y objeto, se determinan uno al otro y surgen simultáneamente (1990: 96).

In that sense, we explain how the complex interplay between various stories form a direct common knowledge (DCK) and self-organized indirect common knowledge (ICK) will emerge as an alternative to prescribed possible systems. In other words, most of the ICK emerge without conscious planning. At the same time, ICK reorganized again in multiple events in which dominates the DCK. This is a cyclical process due to pressure from individuals cooperating in a coordinated manner. We try to explain this relation between DCK and ICK using the Ecology of Pressures Model (EPM). The model in its current state has already been presented at different times; a summary is presented here. For further information on the EPM see Terborg (2006), Terborg and García Landa (2006, 2011, 2013) and Trujillo Tamez (2012). We understand what has been named here as DCK and ICK as a way to expand the idea of “common routine” (CR) and “utmost common routine” (UCR) . We rely on these concepts rather than those of competence or communicative competence. First of all, we present the model in a general way. We must warn the reader that we use the concept of ecology as a context in constant change. Second of all, we explain in more detail how CDK and ICK are formed and evolved in ecology as to emerge reorganized into the UCR. Finally, we present an example of data extract taken from a bilingual situation at school where sometimes we may speak of different languages and sometimes we cannot.

10.2 Ecology of Pressures For the idea of an ecology of pressures our starting point is the premise that communication is a part of human action. Our premise was that all human beings feel some kind of pressure to communicate and use a particular code. Some of those pressures are in conflict, and others are not. Communicative actions are the result of a summary of pressures. One of our purposes is to find out which pressures are dominant in each particular situation. In order to answer this question, it is important to have a tool to classify all pressures involved. If we identify the elements that cause pressures, we will be able to find a tool to classify them. We agree with Bastardas-Boada (2003) that an accumulation of empirical studies would not contribute itself to the development of theories within the framework of complex systems. Accordingly, this work will not present empirical results, but rather examples used as an illustration of our proposed classification for pressures involved in communicative efforts. The dynamics of ecology become decreasingly stable and increasingly dynamic as pressures on a sector of speakers increase in contact situations. That is to say, an alteration in pressures occurs first, and therefore, this can be also characterized

174

R. Terborg and V. Velázquez

as ecology of pressures. Thus, it is necessary to find an answer to the following questions: how do pressures appear? How do pressures change? And how can we classify all different pressures? We assume that pressure is the origin of every human action within a context that we call “state of the world”. The state of the world influences pressure and is modified by action. If we aim at creating a tool to analyze linguistic actions within multilingual situations, it will be necessary to classify pressures. We first classified pressures into pressures that depend mainly on the interest of the person who performs a speech act, and pressures that depend more on the state of the world. This means that we have to distinguish between what ‘I want’ and what ‘I am able to do’. We will now probe into pressures arising from the state of the world (I am able to do), limiting our scope to communication tools, i.e. competence. We propose a more appropriate concept for our objectives, which we will call “common routine” (CR) or “utmost common routine” (UCR) , which generally accounts for the use of a particular code in a conversation. Finally, we will present, as an example, a brief analysis of a bilingual community in Mexico. Not all pressures merge a violent coercion, history shared by a community of speakers allows the development of CR and UCR, in which participants in language exchange launch an almost unconscious negotiation, which press a code to prevail over another without a direct imposition. Different mixed pressures simultaneously lead to action. The origin of pressures is the interest in something and the modification in the state of the world. That is, looking for favorable individual or group context, the involved modification considerably reduces pressure on those who have started a given action; however, because the state of the world is dynamic, there can be an increasing pressure in time where such favorable situation may be disturbed by any given change. The state of the world may limit or help us to be successful in a given action. So the state of the world is a condition for pressure, but it is not the only condition because not all persons feel the same pressure to do something. Interest, in turn, is the condition of what we want to do and it is based on our needs (related to our body, i.e. hunger, social relations, sex, etc.), ideologies, values, beliefs and emotions that help categorize interests and the pressures resulting from the interests (see Terborg and García Landa 2013: 213). Emerging UCR from the state of the world influences pressures helping to their classification. Interests are presented as sets and groups; however, it is possible to identify them. When they are the origin of a pressure, they demonstrate their characteristics in the kind of pressures they cause. Characteristics of interest are variable and these variables determine the pressure in combination with the state of the world. We can talk about immediate and permanent interests, as well as individual and common interests. In addition, having its origin in interests, pressures acquire the same characteristics as interests do; so that immediate and permanent, individual and common pressures may interact and lead to a chain of actions. However, not all pressures lead to action, pressure should be evaluated according to the effort required, which depends on the state of the world as well as the force of

10 Common Knowledge in Conversation of Bilinguals and the Ecology …

175

interest underlying. In case of pressures in conflict the effort required also determine which pressures would eventually lead to action. Some interests lead to internal conflicts and pressures can be found both in individuals and in groups; the state of the world prevents all visualized objectives to be achieved, these conflicts require deciding what pressure should be followed. On the other hand, interests also carry external conflicts between two people or groups or between the individual and the group. Human actions are the result of pressures and fluctuate between levels of internal and external conflicts. These modify or maintain the state of the world temporarily or more permanently. In both cases modification not only alters the state of the world but also depends on it because it includes tools for its own amendment. Although many everyday actions are individual acts, communication is a shared and coordinated action, i.e., is an interweaving of speech acts that meet those actions in a context that are part in the state of the world. The state of the world is everything that forms the context of all action. It includes all relevant things at the time of an action, in this way interest, pressure and power are part of the state of the world which are past and present actions as well as future projections. This includes all relevant processes that create pressures: beliefs, concepts, skills, individual and group ideologies, values, tools to modify the state of the world, one’s own interests and pressures that direct action. As you can imagine, many elements are initiated when communicative actions are established. Ideologies, beliefs, stereotypes, expectations and the development of various linguistic levels, as part of the repertoire of the participants in the exchange, come into use immediately. In this interaction the state of the world is mobilized when two or more participants communicate daily, they generate what is here called (utmost common routine) UCR, a complex process in which they arise, trade, share and create parameters; one personal linguistic norm. Interacting daily with different participants at different levels of power creates a cohesive network of communication by UCR. There may be different (common routines) CRs for a given group of speakers who may be multilingual or speakers of different dialects or registers. These speakers can choose between different codes or ways of speaking but they always tend to speak in the way which will be the easiest for most in a specific situation with specific interlocutors and related to a specific topic. This we call UCR which may be a certain language, dialect or a mixture of different codes. Any change in situation, interlocutors and topic may change the UCR in a conversation. As a reflection of many actions, UCR is the result of diverse interests and pressures in interaction. An example of this is the report of Schmidt’s young speaker Dyirbal (Schmidt 1985). Young Dyirbal avoids using this language with adults because from the perspective of the latter their use is inadequate. However, young Dyirbal talk to each other using the language version of “Young people’s Dyirbal” to identify themselves within their group. This occurs through multiple interactions from the new parameters which are established and shared between different subnets that are part of a community. However, those within the community who have less power in a situation may be opposed to the consensus, in UCR. We want to take that concept rather than competence. Competence would be equivalent to the concept of

176

R. Terborg and V. Velázquez

language. From this perspective, an analysis of the mixture of different competences would then be an analysis of code-switching. The explanation we propose is the continuous interaction between the DCK and ICK making UCR to continually evolve in the context of a specific network of individuals. These individuals in permanent coordination increase the efficiency of UCR and at the same time lessen efforts in interaction. That is, with each joint event communication is facilitated. By decreasing the efforts UCR is one of the factors that modify pressures of people on a network that are interacting in a given ecology. Since each individual tries to avoid further efforts to achieve his objective, pressure leads to evading an action that requires more effort. To sum up, if you can communicate in a given language “A” and in a given language “B” but the use of language “A” is more convenient for everyone, interaction in “A” language will take place automatically, this is because “A” equals to UCR in the group. Each communicative event is a learning event, where a personal story (physical and social) is combined with many other (Varela 1990: 96). In a social network, linguistic interaction is automated, the level of information (linguistic experiences, ideology, power position, pressure, interest, etc.) are present all at the same time, expressed in action. UCR is based on common stories, shared knowledge or skills that are overlapping and that allows them to reach a high degree of automatization; therefore, is based on feedback from direct common knowledge (DCK) and indirect common knowledge (ICK) .

10.3 Common Knowledge When a person goes out but is not sure of which way to go s/he can ask anyone how to get to her/his destination even when they have never met before. Whether or not the pedestrian may provide the requested information, it is likely that they both understand and exchange information using a prescribed code. The same happens when the person in our example enters a restaurant and talks to the waiter or in a market with a salesperson. These situations are quite normal for most humans, they can ask a person they have never seen before and make themselves understood at the precise point in the conversation when most people can play with some accuracy all words and statements issued by another. This may vary if there are some restrictions, for example, when traveling to a region where the language is different. The most common viewpoint of why this works is that there is a common language. But we consider that the ability to communicate with strangers is what creates the illusion of a language, a code or a dialect. That illusion is that people are able to communicate in the same previously learned language. Including the use of the ‘correct’ form, a fact that guarantees success in communication. Correctness depends on the use of established rules and norms that also depend on linguistic and communicative competence of speakers. Therefore, the individual speaker must acquire the rules and move within this framework. Being able to move within the framework of

10 Common Knowledge in Conversation of Bilinguals and the Ecology …

177

established rules is what determines a competent speaker. Consequently, competence focuses on the individual. Certainly, it is also possible to consider competence as a social phenomenon. But in this case the concept is slightly different, even when it is used with the same name. A social phenomenon is negotiable whereas an individual phenomenon can be seen as a positivist representation of an external reality. In that sense we want to limit the individual to what we call competence, whether it is linguistic or communicative competence, the latter can also be something the individual has to acquire without control over the preset rules. It is then, a reality outside the human mind and that mind has to take over this reality. Rules are considered correct in the way they are taught in an English course. Although these rules are then used in social acts in which actions are modified by a group of individuals, it is important to separate them but not to distinguish them from the same social phenomenon as Bastardas-Boada says, “we can distinguish, but not separate” (2013: 18). Competence is closely related to the concept of language. An individual can become proficient in a language or may be incompetent. Success in competence depends on her/his verbal acts. Language is a convenient concept for linguists who do not need to explain how communication between strangers who have never seen before works. The concept even enables communication of someone dead for centuries, like a dead author through his writings with people still alive. Now, if we take this chimera called language as an objective and external reality to our mind, we could say that some languages have millions of competent speakers while others have less than a hundred, we can find examples of these languages spoken by few people anywhere in the world. Some of them may have had millions of speakers but now they are in a process of language shift. Examples of this situation are the languages Ayapaneco, Paipai or Kiliwa in Mexico. However, especially in the Amazon and New Guinea, some languages have never exceeded one hundred speakers (see Romaine 1994: 1–16). In such cases there is a high possibility that speakers may know each other. That is, when a speaker of a language A finds another speaker B, B may also be a relative or acquaintance of A. Obviously, these speakers have coexisted for so many years that there are no chances they have to talk to a stranger in her/his own language. If we had the opportunity to engage in everyday activities of these communities, we would probably find a different way to see their linguistic norm. Obviously we do not know to what extent speakers care about prescribed rules in their way of communicating, but it is likely that the idea or concept of ‘a language’ does not match what most speakers of major languages considered as such (see Romaine 1989: 286–7). Language is only an illusion, although very convenient at times as we already emphasized. Now if you come back some 10,000 years in the history of mankind, chances are that all speakers of a certain code knew each other, so when people found a stranger that person usually spoke a different language. In that sense, speakers have a common language and personal story that leads directly to a DCK. So, when speakers communicate they modify their partially shared knowledge. That knowledge led to the dependent pressures of UCR. Signs organize themselves in communicative practice. As we just mentioned, somehow those shared personal stories (personal

178

R. Terborg and V. Velázquez

and social stories) that arise from this communication are more direct than others. That is, these stories rely on a more authentic experience. There are also indirect shared personal stories based on ICK. That is, there is a continuum between DCK and ICK, which in turn form a feedback loop (Morin 2001). To illustrate the difference between DCK and ICK we can take the knowledge created by modern science shared by many people. Nothing could be more perfect than the knowledge we have about travel to Mars. The images and other information we receive from that planet are given indirectly by space probes. So, there is only an indirect knowledge because it is known through stories, books, pictures or videos. This common knowledge is indirect because it is based on reports or documents. Similarly, we can consider ICK, we know about historical events in which there are no survivors. They are events which may not have any DCK. In that sense, not only what we observe in nature but also the acquisition of elements of a language may be the result of conversations with speakers of a language, and those conversations would form part of a common direct personal story and come to form a DCK. But there is also the case that knowledge is the result of an indirect common personal history as the acquisition/learning in the context of formal education. In today’s world, perhaps most cases of acquisition are a combination of direct and indirect common stories. In prehistoric times, human groups departed from the DCK and the common direct personal history. At that time one person could be carrying the relevant knowledge of the whole group. Gradually, with teaching abilities, mystical stories, etc. indirect knowledge was gaining importance and thus appeared increasingly indirect common personal stories, i.e., people who shared knowledge without sharing the same experiences and without knowing each other. Since then began to generalize knowledge. Speech acts began to be modified obeying pressures emerging from ICK. Our modern societies depend largely on generalized rules. The so-called “knowledge society” is a widespread knowledge society, unified and homogenized. In this generalized knowledge (ICK) modern societies are basing their education, the production of goods and though it seems contradictory, the creation of new knowledge. Probably more than 90% of the new knowledge belongs to ICK. We do not acquire this knowledge through our own experiences, i.e., seeing with our own eyes, hearing it with our own ears or feeling it with our own hands. Currently people acquire knowledge through stories, books, newspapers, electronic media, education, etc. The DCK, however, has been and remains central to humanity. The common personal story is the story that is shared between at least two individuals shared personal history and that leads to a common understanding. An isolated person acquires knowledge only as a result of a solitary experience. This knowledge, then, is not common until you relate to someone who will be part of the ICC. DCK and ICK are products of a common personal history.

10 Common Knowledge in Conversation of Bilinguals and the Ecology …

179

If the DCK is paramount, then the ICC emerged along with language. Knowledge is only common in a partial sense. Two people never see the same event in quite the same way. That is, there are different degrees of common understanding between two or more people. By the same token, what is shared is only a part of everyone’s perception. What is common is the basis for future communicative acts. When that common ground satisfies the requirement for the development of common actions that are perceived as successful, the differences in the knowledge of two people can be unnoticed. Both DCK and ICK are part of the state of the world while enacting the state of the world (Varela 1990: 87–120). The state of the world that makes up the context of actions determines the pressures do emerge in the actions (see Terborg 2006; Terborg and García Landa 2011, 2013). If we said that DCK originates in emerging pressures UCR, this is also the product of DCK and partially also ICK. DCK and ICK feed each other. As we said above, DCK is paramount but the ICK is developing more and more with the homogenization of knowledge. DCK emerge in small social networks and micro level is where it is taking shape and fits many immediate needs and is rather the product of immediate pressures. Small personal networks are connected to other networks at the meso level and thus emerges a network of smaller networks whose members are sharing knowledge. Thus, it is possible that some person A has created some new expression that he has only ever shared with B but another time you will hear that expression of C who had no previously shared term. That is, B has transmitted to another network where this term has been accepted and is now in use. Thus, emerges ICK trigger back with the DCK and the DCK is transmitted to different social networks becoming ITC which in turn feed a pop DCK. In communities where communication consists mainly of talking, the net has its limits. However, there are degrees of development of the ICC and in other communities it is more evolved, due to the form of communication that can be spoken, written, etc. and due to the organization of their networks. The organization is increasingly complex and there is not only a network of networks but a network of network of networks at the macro level, which could be considered as a super network. Individuals do not only belong only to one social network (Milroy 1982), but several and serve as links between these different networks. Knowledge diversifies and people specialize in circles of knowledge they are sharing with certain networks. In this dynamic it is increasingly important the development of the ICK to communicate in a successful way. The DCK has ceded many domains to ICK, however, the DCK is still important because the ICK could not evolve without DCK. There are always emerging DCK creations that are becoming ICK in a complex interplay between both of them. The DCK still exists but is often unnoticed. The DCK continues feeding the ICK but, simultaneously, the ICK also feeds the DCK. Both are forming a loop (Morin 2001: 213–269) in which it becomes difficult to distinguish between. In the same way Bastardas-Boada writes that the language is in society that is in the language. Esta aproximación permite así, finalmente, la superación adecuada de antinomias de larga tradición que bloquean nuestra comprensión de la realidad y nos distraen con discusiones estériles. Lo real es, pues, co-existente, co-dependiente: el individuo está en la sociedad

180

R. Terborg and V. Velázquez

que está en el individuo; la mente está en la cultura que está en la mente; la lengua está en la sociedad que está en la lengua. Lo que intuimos que ciertamente ocurre es así ‘decible’, formulable: las interdependencias, las imbricaciones, la realidad de los elementos que evolucionan influyéndose y determinándose mutuamente (Bastardas-Boada 2003: 6).

So we can also say that the ICK is in the DCK that is in the ICK. Both are interdependent. Using standardized language with normative grammatical rules people always get creative under these rules but also by ignoring these rules. Through the web of networks a part of the DCK will be part of the ICK. Power is related to ICK. It is the knowledge considered ‘proper knowledge’. Of course, not all knowledge of the ICK is considered proper knowledge, but proper knowledge is part of ICK. So there is a value added to ICK it did not have before. This value creates interests from which new pressures emerge (Terborg 2006) to act in a specific manner. The act of speaking must be within the framework of the norms. Thus, logically the illusion of a code or a language arises. Speakers often believe that the ICK is the only valuable knowledge although these speakers in their networks are negotiating meaning in a complex combination of ICK and DCK. Becoming more widespread the ICK is also becoming less modifiable and the DCK will be needed for successful communicative actions. Thus, the illusion of language becomes a mental reality that for many people seems to be an external reality which is independent of the speakers. And it is this belief of the external reality that creates power. People who are not part of the super-network of the highly valued ICK do lack power and suffer disadvantages because they do not dominate the norms of high prestige. The illusion of the ‘correct code’ hinders the comprehension of the interplay between DCK and ICK, especially in bilingual situations. In this sense, even bilingualism and multilingualism may be seen as an illusion. As Massip-Bonet points out “Two or more systems can change in response to one another, in a process of co-adaptation. … Language emerges from continuous human social interaction and the structure is fundamentally shaped by cognitive abilities, by processing idiosyncrasies and limitation/abilities and by the network of circuity in the human brain” (Massip-Bonet 2013: 47). Instead we may think in the ICKs of two or more separated super-networks getting in contact. From a structural view seen we may say that elements and established rules are mixed. In this mixture we may identify elements from ICK “A” and elements from ICK “B”. But this is not always the case. For example, when a group of immigrants speaking language “A” are living for some time in the territory of speakers from language “B” we may suppose that all newcomers from A who know both languages will be able to communicate with all immigrants and their descendants in territory B because they are able to use both ICKs. Surprisingly these newcomers may have serious problems in communication, at least at the beginning of their stay, as Romaine states. For the isolated bilingual individual, the norms for the use of the two codes may be those of the separate communities in which the languages are native, but for the bilingual in a community of other bilinguals, speakers may create their own norms which are quite different. This may

10 Common Knowledge in Conversation of Bilinguals and the Ecology …

181

lead to the creation of a new language and have quite a dramatic cumulative effect, once there is a younger community of native speakers of the new system who have had no exposure to the pre-contact system. New members who enter such communities often find they are not able to understand the community’s variety of language. One Yugoslavian-born woman who was fluent in Serbo-Croat and had a passive knowledge of English when she moved to Milwaukee, said she could not understand the language used by the Serbian community. (Romaine 1989: 163–164)

The ICK by itself is limited and has to adapt to the real situations for the emergence of a successful conversation. Thus, the DCK emerges in a new network. Only an appropriate combination of ICK and DCK leads to success. As said before, unlike competence, it is UCR that permits successful communication with minimal effort. Competence guides itself by ICK but UCR emerges in an appropriate combination of ICK and DCK. Bastardas-Boada exemplifies with a dance of two persons (2013: 20). Speakers in an interaction coordinate their ICK with their DCK and thus among all participants emerges a new DCK. Parents treat their child who starts learning to speak different from other people. They have to coordinate very carefully the ICK and DCK because the child wouldn’t understand communication among adults. Parents and their child do not share the same ICK. They often repeat the pronunciation of the child’s words to make it sure to be understood. So ‘inappropriate’ words in ICK are becoming successful in DCK. The same words may be incomprehensible for a person that does not stay in contact with the child. In the first part of their life humans hardly share much knowledge and ICK nearly doesn’t exist. Something similar happens when adults are using a lingua franca because they have to look for a poor ICK while DCK nearly doesn’t exist. Only with enough pressure among all DCK will emerge. In these cases, linguistic norms may lose importance. Lacking a lot of ICK participants have to base on individual knowledge and try to develop a functional DCK. When there is no appropriate coordination between ICK and DCK there will be no success in communication. We would like to present an example when ICK do not satisfy the needs in conversation that consists in a record of the Spanish lesson in a Primary school in Xocen. Xocen is a village of about 1000 people in the Mayan area of Yucatan, Mexico. The native language of most people in the village is Mayan although it is not far from the city of Valladolid where mainly Spanish is spoken. Most children of Xocen had their first contact with Spanish at school. The lesson was from a third-grade group and there was no difference in the method and the book that were in used in the rest of Mexico during the time of data collection. No difference was made, using the same didactic as anywhere in the nation. But the teachers themselves were speakers of Mayan and used their knowledge shifting from Spanish to Mayan and vice versa to make the children understand. Grammar was taught to these children in the same way as it is taught to native speakers, i.e. by making the grammatical rules the speaker is using conscious and not as second language teaching. The teacher faced up with an inadequate method prescribed by the book. She tries to solve the problem with instant translations and instructions in Mayan as we will see in our examples.

182

R. Terborg and V. Velázquez

(1) teacher: a ver este e por ejemplo tech Andrés ‘a’alten hum peel enunciado cin dz’iib te pizarrono Spanish translation: tú Andrés dinos un enunciado para que lo escriba en ese pizarrón (2) student: los caballos corre en el campo (Let me see, for example, you, Andres, give us a sentence I will write on the blackboard) (Horses run in the field) The ICK is the knowledge imposed by the book. The teacher begins with Spanish and shifts to Mayan while asking for a sentence and DCK is not yet visible because what was said in Mayan might be ICK as well. Changing elements between both languages is DCK because that’s the way she tries to solve the problem with her school class. For example, “pizarrono” bears the Spanish element “pizarrón” (blackboard) and the Mayan suffix “o” (those). There are many similar examples from the lesson. The way the teacher brings the discourse close to the success because it keeps the structure of Mayan grammar. We have to bear in mind, that the problem for the teacher here is enormous compared to lessons in schools in town where children are monolingual Spanish speakers. She uses the structures of the ICK and combines them with new structures of the DCK. The way one of the children answered “los caballos corre en el campo” (the horses are/is running in the field) does not inflect the plural of the Spanish word that would be “corren” (they are running). The child says “corre” (it runs). This rises suspicion that the sentence might have been remembered from a former lesson and wasn’t invented by the child. If this is true, the remembered sentence was from the book, that means it is part of the ICK. But what the child remembers is the DCK because it is the memory of the teacher’s discourse from a former lesson. In this sense DCK is a special code between the teacher and the children. Perhaps it will never leave the network of persons of the classroom. We may find and understand it only among the members of the network. So, part of this DCK will never be ICK. Perhaps the children will never know how to find the subject or the verb of a sentence, which was the target of the lesson. The conversation between children and the teacher was partly successful. Because all teachers’ questions have been answered to the satisfaction of the teacher. Between the teacher and the children emerged a DCK that is starting to be part of the UCR of this group of people.

10 Common Knowledge in Conversation of Bilinguals and the Ecology …

183

10.4 Conclusions If we would only analyze the mix of different ICKs coming from different supernetworks, we would only look for rules of code-switching and we wouldn’t become aware of the creativity of the bilingual speakers. In contrast, the inclusion of the DCK not only focuses on code-switching, but building fugacious codes or permanent codes. These are the codes that only do appear among persons who know each other. The phenomenon emerges in bilingual and monolingual networks as well. Each individual develops a corporal and social history. Pressures lead people to joint actions modifying the state of the world while they are part of the state of the world. Diminish efforts in actions is a common unconscious goal of everybody. The actions have to be coordinated and adapted to other persons. These pressures are mainly the product of UCR that is under permanent modification in relation to the permanent changes of ICK and DCK. Each joint action causes mental changes that are related, but these changes are not identical in peoples’ knowledge. Their histories are partially shared and they are partially individual. The conversation between teacher and children partly can be called successful but probably there is a failure too. In this situation she is building a DCK on an ICK which is absent in the children. It is not the inability of the teacher but the teaching plan of the book. The elements introduced in Spanish could not be translated into Mayan without problems. To maintain success in conversation a new DCK that is changing quickly is developed. The ICK wasn’t appropriate for the situation and it had to be carefully balanced with an appropriate DCK. On the other hand, it is possible that the failure may be unobserved in the future because conversation was successful. The children answered satisfactorily to her questions. It is almost possible that in the future they will not have any need to remember what is a verb or a subject. There are elements of parts of conversation that are only used between people who know each other. It is only DCK that makes success possible in combination with elements of different ICKs which are connected because of a lot of overlaps. ICK is the part of the road that is paved. But DCK is built with different ICKs which are combined with difficult pathways. Two persons who have never met before are able to communicate because of their ICK. But in the moment of contact they start developing a DCK that depends on environment. DCK facilitates the use of ICK. Persons who are members of a longstanding network count on a well-developed DCK which enables communication with less effort. When two persons have a background of different ICKs their DCK will feed from these ICKs and new forms will emerge. In this way we have presented a different point of view to show how bilingual discourse emerges as a product of pressures depending on DCK. Aknowledgments The article was supported by a research project grant on the vitality of indigenous languages in México (PAPIIT IN404313) Situaciones y prácticas bilingües que contribuyen al mantenimiento-desplazamiento de lenguas. Análisis del ‘conocimiento’ y de la ‘máxima facilidad

184

R. Terborg and V. Velázquez

compartida’ como sistemas complejos funded by the National Autonomous University of Mexico, Department of Applied Linguistics at the Teaching Centre of Foreign Languages (CELE-UNAM).

References Bastardas-Boada, A. (2003). Lingüística general: elementos para un paradigma integrador desde la perspectiva de complejidad. LinRed: Lingüística en la red, 12, 1–23. Retrieved from http://hdl. handle.net/10017/24698. Bastardas-Boada, A. (2013). Sociolinguistics: Towards a complex ecological view. In Á. MassipBonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 15–34). Heidelberg: Springer. Massip-Bonet, Á. (2013). Language as a complex adaptive system: Towards an integrative linguistics. In Á. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 35–60). Heidelberg: Springer. Milroy, L. (1982). Language and social networks. Oxford: Basil Blackwell. Morin, E. (2001). El método I. La naturaleza de la naturaleza. España: Ediciones Cátedra. Romaine, S. (1989). Bilingualism. Oxford: Basil Blackwell Inc. Romaine, S. (1994). Language in society. An introduction to sociolinguistics. Oxford: Oxford University Press. Schmidt, A. (1985). Young People’s Dyirbal. An example of language death from Australia. Cambridge: Cambridge University Press. Terborg, R. (2006). La ‘ecología de presiones’ en el desplazamiento de las lenguas indígenas por el español. Presentación de un modelo. Forum Qualitative Sozialforschung/Forum: Qualitative Social Research, 7(4), 39. URN: urn-nbn-resolving.de/urn:nbn:de:0114-fqs0604396, http://dx. doi.org/10.17169/fqs-7.4.167. Terborg, R., & García Landa, L. (2006). Cómo los conceptos pueden influir en la planificación del lenguaje: la competencia y su impacto en las relaciones de poder y la desigualdad. In R. Terborg & L. García Landa (Eds.), Los retos de la planificación del lenguaje en el siglo XXI (Vol. I, pp. 163–182). México: CELE, UNAM. Terborg, R., & García Landa, L. (Eds.). (2011). Muerte y vitalidad de las lenguas indígenas y las presiones sobre sus hablantes. México: CELE/UNAM. Terborg, R., & García Landa, L. (2013). The ecology of pressures: Towards a tool to analyze the complex process of language shift and maintenance. In Á. Massip-Bonet & A. BastardasBoada (Eds.), Complexity perspectives on language, communication and society (pp. 219–246). Heidelberg: Springer. Trujillo Tamez, I. (2012). La vitalidad lingüística de la lengua ayuk o mixe en tres comunidades: Tamazulapam del Espíritu Santo, San Lucas Camotlán y San Juan Guichicovi. (Doctoral dissertation). UNAM. México. Varela, F. (1990). Conocer. Las ciencias cognitivas: tendencias y perspectivas. Cartografía de las ideas actuales. Barcelona: Editorial Gedisa.

Part IV

Discourse Analysis

Chapter 11

Discourse Analysis: The Constructivist Perspective and Transdisciplinarity Esperanza Morales-López

Abstract This paper explores how discourse analysis can benefit from the main tenets of complexity theory: including its holistic (or systemic) perspective in the research of any object, always in relation to its emergency conditions; and transdisciplinarity as methodology. If applied to the study of discourse, it revitalizes ethnography as an empirical methodology, constructivism as a theoretical starting position, and the integration of discourse analysis with rhetoric, argumentation theory and semiotics, among other disciplines.

11.1 Introduction My aim herein is to consider how a great deal of the ideas in the theoretical approach of complexity may be very valuable for enhancing our understanding of discourse from a functionalist and constructivist perspective. To that end, I begin with a brief reference to complexity, with the thoughts of one of its key authors, Edgar Morin. In one of his books, he defines complexity based on an opposition to what he calls the paradigm of simplicity: The paradigm of simplicity puts order in the universe, and chases out disorder. Order is reduced to one law, one principle. Simplicity can see either the one or the many, but cannot see that the One is perhaps at the same time Many. The principle of simplicity either separates that which is linked (disjunction) or unifies that which is diverse (reduction). (1990: 89)

As a consequence of the effect of this paradigm’s predominance during the nineteenth and twentieth centuries, the purpose of studying any scientific object was to determine each of its components, and to describe and analyze them separately. For example, let us consider the human being, a case proposed by Morin himself; it can be considered approached from a biological perspective (as a being with anatomical and physiological characteristics, etc.) or from a cultural perspective (as a subject which speaks, and which has ideas and consciousness). Each of these aspects has E. Morales-López (B) Universidad de A Coruña, A Coruña, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_11

187

188

E. Morales-López

been the subject of a scientific discipline, in which the primary objective has traditionally been to determine its simplest units. The same is true of other objects of study: physical matter, the universe, etc. However, the obsession with the simplest units ended in the early twentieth century, when a group of physicists realized that disorder is always present in the universe; the ideal of science as the search for the simple was therefore revealed to be a mere pipe dream. One of those physicians was Boltzmann, who proved that heat is simply the random movements of molecules and atoms. When water is heated, its molecules begin to swirl about, and some fly off into the atmosphere and dissipate. Disorder is also present in the life, linked to work and to transformation; disorder is therefore as inherent in nature as order is. Indeed, a new order may emerge from disorder—a new quality, as occurred at the beginning of the universe (the Big Bang theory). Complexity, continues Morin, is therefore the paradigm that is based on the assumption that in order to understand what happens in life, it is necessary to consider the relationship between order, disorder and organization. This relationship also means that complexity accepts contradiction as an inherent fact in the study of life. While reflecting further on the various theories related to complexity that have been proposed, for the purposes of this study we can consider systems theory, formulated by the Austrian biologist Ludwig von Bertalanffy, in the mid-twentieth century. For an explanation of this type of metatheory, we begin with the excellent book by the physicist Fritjof Capra (1996) The web of life. According to Capra, the key to formulating a complete theory of living systems (the systemic perspective) lies in a synthesis of the following three characteristics: the study of the pattern, the structure and the process of life. The organizational pattern of any system, living or otherwise, is the configuration of the relationships between its component parts, which determines the essential characteristics of the system. For example, some abstract relations must be present for something to be recognizable as a bicycle: in this case, the functional relationships between the handlebars, pedals, wheels, etc. The structure of a system is the corporeality of its pattern of organization: in the case of the bicycle, its various shapes, size, the physical composition of its components depending on the type of bicycle (city, mountain, etc.). In a bicycle, the parts are designed to form a structure with fixed components. However, in a living system, the components change continuously because within it there is growth, development and evolution. This is the third component necessary for a complete description of the nature of life: the process of evolution. These three criteria (pattern, structure and process) are completely interdependent, so that the pattern of organization can only be recognized if it is embodied in a physical structure; in living beings it is in turn a continuous process. The pattern of a living being is also considered to be an autopoietic (or selforganized) unity, i.e. the system remains stable despite evolving, and does so autonomously (Maturana and Varela 1992; Massip-Bonet 2013). At the same time, the pattern of a living being is a dissipative structure (a term proposed by the physicist Prigogine), which means it is open to the conditions of its environment and flows of

11 Discourse Analysis: The Constructivist Perspective and …

189

matter and energy. Because of this openness, the system’s state of equilibrium can break down at crucial times or bifurcation points. These are periods of instability, in which new forms of disorder can arise spontaneously, as a result of various other different evolutionary structures. Furthermore, in a living being the processes of life are identified by cognition, which is the process of knowing. This statement implies a completely new idea about what the mind is; it is no longer a thing but instead a process. In other words, the organizing activity of living systems is a mental activity at all levels of life; the interactions of a living organism—a plant, animal or human being—with its environment are cognitive, mental interactions. Life and cognition are thus inseparably linked. This new concept of cognition is much broader than that of thought; it includes perception, emotion and action: the entire process of life. The brain is the specific structure through which this process operates. In human life, human cognition also includes language, conceptual thinking and all the other attributes of human consciousness (Capra 1996: 185 and 188). Ideas such as those proposed by systems theory show that we are moving from the idea of a simple world explained by laws and principles (the predominant position in science according to the rationalist and positivist ideal), towards a new idea of the world as a completely complex entity (Delgado Díaz 2007). This also calls into question the very purpose of scientific disciplines and the need to interlink their focus of study in what we call multidisciplinarity, interdisciplinarity and transdisciplinarity (Nicolescu 2007). Multidisciplinarity is the study of a single object through various disciplines at the same time. Interdisciplinarity has a different perspective; it is concerned with the transfer of methods from one discipline to another. Transdisciplinarity is the dynamic that arises as a result of the simultaneous investigation of several levels of reality with a dialogic purpose, i.e. creatively relating concepts from various disciplines in order to understand the processes of life in an increasingly complex way (Massip-Bonet 2013: 36; Pujante and Morales-López 2013). This different approach to science enables the object in question to be studied as a creative process. It also paves the way to a conception of the scientific process as a discontinuity, in contrast to the idea of a continuum advocated by positivism (Couceiro Bueno 2012: 166). This leads to the re-emergence of disciplines traditionally banished by positivism, such as hermeneutical, historical, rhetorical approaches, etc., all within a holistic view of knowledge. This implies a revaluation of the historical contexts that compete on an equal footing with the prevailing logical models in the description of the received conception. Postulates such as the idea of a universal language of science, truth as correspondence, and the separation between fact and theory, among others, thereby enter a state of crisis. According to Couceiro-Bueno, the splintering of the scientific community gives free rein to creativity, which sees its own work as a process rather than a product.

190

E. Morales-López

11.2 Discourse—A Complex Notion Discourse is an object of study about which it is difficult to explain its nature using the dominant linguistic theory; in other words, from the perspective of structuralist formalism, in turn inspired by the positivist philosophical tradition (an example of the paradigm of simplicity). In the 1970s, the semiotician Umberto Eco argued (1976: 161) that the results of neopositivism have been positive for the exact sciences (a statement that is debatable even today), but limited (and even dangerous) for the social sciences. There are several reasons for the prevailing inadequacy of linguistic theory in the twentieth century in the discursive field. First, there is no balance in the communicative field because a great deal of meaning is constructed in the process of interaction, and is closely linked to the context; this meaning is therefore completely dependent on those conditions. The laws of grammar impose the minimum constraints necessary to ensure mutual intelligibility, but also preserved some degree of uncertainty from the point of view of meaning due to their dependence on context (Verschueren 1999: 111; Juarrero 1999: 168); speakers choose and select lexical-syntactic forms and constructs according to their specific intentions and the socio-cultural constraints imposed by the communicative situation in which they are participating. As Halliday pointed out (1982: 45, 51 and 66), creativity in language consists of the ability of a social actor to build other meanings in new contexts of situation, based on the semantic options that have been codified as grammatical choices in the language he/she uses. As a result, according to this author, what the speaker “can mean” (the semantic choices) equates to what the speaker “can say” (the functions) in the discursive process. Second, it is also impossible to deduce the overall meaning that is constructed in a given interaction from the knowledge and use of those rules; it is necessary to interpret what the users of the language have wanted to communicate. Chomsky’s hypothetical-deductive method is therefore a distant illusion; even Austin’s proposal of developing a series of conditions for compliance for speech acts (Juarrero 1999: 54–55) is insufficient. In discursive transformation, what is communicated is both explicit and implicit, with a continuous transgression of the rules of communication (Pujante and Morales-López 2013). Third, in the study of discourse it is necessary to link the functions (the pattern) and the forms (the structure) that convey these functions. This is because what is important is not only the communicative purpose, but also showing how that communicative purpose is constructed within the discursive process. The form is not merely an envelope for the function, but is also a constituent part of the overall meaning (White 1987; Morales-López 2017a, b). Hence the importance of the following statement by Halliday (1970/2002: 173–4), in which he explains how to understand the nature of language: We cannot explain language by simply listing its uses… Malinowski’s ethnographic account of the functions of language, based on the distinction between ‘pragmatic’ and ‘magic’, or Bühler’s well-known tripartite division into the ‘representational’, ‘expressive’ and ‘cona-

11 Discourse Analysis: The Constructivist Perspective and …

191

tive’ functions, show that it is possible to generalize; but these generalizations are directed towards sociological or psychological inquiries, and are not intended primary to throw light on the nature of linguistic structure. At the same time, an account of linguistic structure that pays no attention to the demands that we make of language is lacking in perspicacity, since it offers no principles for explaining why the structure of language is organized one way rather than in another… It is necessary to look at both the system of language and its functions at the same time; otherwise we will lack any theoretical basis for generalizations about how language is used.

In a similar vein, Halliday clearly shows that his starting point is the function: what is important is what users do with language, and with what concrete and specific purposes of a culture. However, his idea is new in that it does not consider the study of language beyond the relationship between form and function. As suggested by the anthropologist Malinowski or Bühler (mentioned above), the study of the communicative functions in themselves may be an objective in sociological or psychological research, but Halliday argues that it is insufficient for those interested in the nature of language. However, a purely structural analysis that pays no attention to the communicative functions is simple descriptive research, and unable to provide explanatory principles for how language functions. In this respect, when explaining the dialectical relationship between function and form, his argument is based on the idea that the basic unit of communication is the speech act. According to this concept, the speaker makes a selection from a range of related options in a given communication situation. These options represent the potential for meaning—of language—for Halliday, a similar concept to Hymes’ notion of communicative competence (Halliday 1982: 50); a potential that is presented systematically, constituting what is known as the grammar of a language. The speaker selects a specific option from the possible forms and constructions, which are not in a vacuum, but are instead, as mentioned above, always related to the situation context in which the speaker finds himself (Halliday 1970/2002: 174). The specific contributions on the relationship between speech and communicative functions are made not by this author, but rather by his followers. An example is Martin (2001), who continues to consider cohesion, as Halliday and Hasan did (1976), as part of the process of texture: the property of “being a text”, of functioning as a semantic unit while taking into account its contextual conditions (texture, in turn, as a macro-function forming part of consistency). Martin also reformulates the classification of Halliday and Hasan’s bonds of cohesion (1976: reference, ellipsis, substitution, conjunction and lexical cohesion), within the more general idea of a number of semantic meta-functions that organize the cohesion in a text. What is interesting about this idea is that Martin continues to move forward with the programme that Halliday had begun—the dialectical relationship between form and function; because this perspective means that any structural analysis is inseparably linked to a more general study of functions. However, we must bear in mind that meaning is also completed in relation to the subject (its subjectivity or emotional state) and with the human action that the subject performs in the communication process. Cognition is therefore also always linked to discourse and to communicative activity (as already foreseen at a very early stage

192

E. Morales-López

by Vygotsky 1934/1986). In reality, we can conclude that it is in an individual’s communicative activity with others where the interrelation between the cognitive and the social takes place. On this point we follow Maturana and Varela (1992), and Maturana (1996), as advocates of constructivism in the explanation of knowledge; we agree that interaction is the way to consider the cognitive, because cognition is the process of living itself (a similar idea is expressed by Vilarroya 2014). Finally, discourse is a subject of study that has been approached from many social disciplines and with various perspectives. In this aspect, de Beaugrande (1996: 22–26) makes an interesting argument that discourse is itself a transdisciplinary field, strategically located to analyze different issues from different angles. Since the 1970s, discourse research has benefited greatly from the reflections of disciplines such as the philosophy of language, constructivist sociology, ethnography of communication, analysis of conversation and ethnomethodology, cognitive psychology, and the reflections of various discourse theorists (Foucault, Bourdieu, Habermas, Bakhtin, Voloshinov, Pêcheux, etc.), in traditions that are otherwise essential for those beginning work in this field of research (Morales-López 2011). Long before that, rhetoric and argumentation theory studied discourse in terms of persuasion in the public agora (Perelman and Olbrechts-Tyteca 1989; Pujante 2003, 2017a; Meyer 2008); semiotics emerged later, and related discourse (or the study of texts) to all the other signs (an update in Kress 2010; and Eco 2011). The perspective of complexity merely reinforces our idea that discourse can only be studied from the crossroads referred to by Beaugrande. At this point, in addition to considering this characteristic of discourse as a bridge between disciplines, the description by Scollon and Wong Scollon (2001), of the importance of the study of discourse to obtain a better understanding of some of the most topical social issues in the contemporary world, is interesting. One of these issues is interculturality; a very good way of finding out about people’s cultural similarities and differences is by analysing communicative interaction within large corporations, for example, where people from very diverse cultural backgrounds and with partially different world views work. Many other social issues are currently being explored through discourses: these include the construction of the health professions (Salvador et al. 2013), the communicative management of health risks (Nespereira García 2014), ideological conflicts such as those over the rights of women and immigrants, the social versus the capitalist economy, renewable energies, terrorism, etc. (Morales-López 2012b, c, 2017a; Martín Jiménez 2013; Salvador 2014; Pujante 2017b; see other examples in Morales-López & Floyd 2017). A social issue that has interested us a great deal in recent years is the analysis of the discourses of social change; those making proposals for socio-political change. The current crisis of the neoliberal capitalist model is leading to the creation of new narratives and alternative discourses which not only attempt to formulate a sociopolitical criticism, but also to explain the various aspects of the crisis based on a cognitive framework that is different to the previous status quo and to use more

11 Discourse Analysis: The Constructivist Perspective and …

193

creative discursive processes to articulate (and/or construct) new conceptions of the world, able to achieve more active participation by citizens (de Sousa Santos 2011; Castells 2012; Morales-López 2012a; Pujante and Morales-López 2013; MoralesLópez 2017a; Montesano-Montessori and Morales-López 2015).

11.3 Discourse Analysis and Complexity Theory The holistic view that advocates complexity in the study of nature and society entails an investigation of any scientific object in its relationship to the whole. In the field of linguistics, it always means taking the centrality of speech and communication as a starting point; because it is in interaction with others, in specific contexts and actions, where we complete our life’s cognitive process (Larsen-Freeman and Cameron 2008: 107–108). The study of any linguistic phenomenon must therefore always consider the complete discourse, the interaction between different levels of language (the linguistic and the pragmatic-rhetorical) , the discursive genre, the actors conveying it in a particular context, and those discourses’ relationship with simultaneous action. We therefore reiterate that the constructivist ideas about knowledge of the biologists Maturana and Varela (to which we have already referred to extensively in previous works: Morales-López 2011 and Pujante and Morales-López 2013) appear to be highly topical: what makes us human is not language itself, but the act of communicating. According to Maturana (2006: 96): The human lineage arose as a living together in networks of conversations conserved from one generation to the next in the learning of the children. From its very origin humanness arose and occurs in networks of conversations, and all that we human beings do as human beings occurs in networks of conversations. Indeed, the different manners of living that we live, the different worlds that we generate in the course of our living, all occur as different networks of conversations, and in particular, the different cultures that we live are different manners of living in closed networks of conversations… we generate the worlds that we live as networks of conversations in the dynamics of the interplay of our languaging and emotioning as different domains of objects, entities and relations.

This centrality of communication, of language as a process in which the rational and the emotional converge, entails adopting two methodological requirements in research on discourse. First, the recovery and/or updating of ethnography as a research methodology which enables the investigation of meaning as a holistic construction in the communication process. On this subject, Blommaert and Jie (2010) consider that ethnography is not only a method for collecting data, as many believe (a common reductionism), but instead a method of researching language and communication, with a history dating back to the work of Boas and Malinowski, which includes both an ontological and epistemological perspective. In the ethnographic tradition, categories of analysis are not considered on an a priori basis, i.e. as “reifications” with an independent existence, but rather as mere abstractions (Scollon and Wong Scollon 2000: 540; Gumperz 2001). Any category

194

E. Morales-López

of the object studied therefore needs to be subjected to empirical analysis. And this complex empirical analysis should include the collection of all or some of the following types of data: generalizations made by the social actors in communication processes, objective observations by the researcher, the individual experience of some specific actors (which may deviate from the common experience) and the interactions of the observer with individuals he/she is researching (Scollon and Wong Scollon 2001: 19–20). Ethnography is thus an inductive approach, which focuses on the analysis of a particular case; a specific case belonging to a larger series of similar cases to which it is related; and which also allows generalizations and theoretical interpretations to be made (Blommaert and Jie 2010: 12). The level of interpretation of the discursive data advocated by ethnography involves including some degree of uncertainty in the analysis. In this respect, ethnography is close to the hermeneutic tradition (Duranti 1997). In ethnographic research, the ideal of objectivity in rationalist research is replaced by rigorous data collection, in the analysis itself and in the interpretation of the discursive data carried out based on the theoretical tradition. However, as well as the study of any discursive phenomenon in its communicative development, studies of complexity remember that all communication processes are always linked to human action: either because they are discourses arising from certain actions, or because they are discourses undertaken to cause actions. The discourse researchers who have done most to clarify the relationship between discourse and action include Scollon and Wong Scollon (2005). For these authors, an urgent task in critical discourse analysis is indeed this theorizing about how discourse becomes action, and action becomes discourse. In a more recent study, Scollon (2008: 15) makes this relationship explicit as follows: The question is: what action is being taken by what social actor in a concrete material place in the world at a specific time and how is the document or text (or any other mediational means) used by the social actor as a tool for taking that action?

The question is one of analyzing specific actions and discourses by actors at certain points in time in order to reconstruct a particular story or narrative of the events that goes inseparably from the social to the individual and from the individual to the social. Scollon and Wong Scollon (2005: 107) therefore believe that Bourdieu’s concept of habitus (1990) blurs the distinction between the collective and the individuality of a particular social actor. A second aspect caused by complexity is the need to develop research methodologies that move towards transdisciplinarity. In the case of discourse analysis, this involves the joint consideration of disciplines that have historically provided analytical solutions to the study of communication: rhetoric, argumentation, linguistics (pragmatics and/or discourse analysis) and semiotics. Rhetoric can be found in all types of discourses; furthermore, the study of orality undoubtedly originates in rhetoric, as Albaladejo (2013) reminds us. Eco also alluded to it (1976: 110) when he described the interest of the sophist in the persuasive effect of the word as pragmatic. However, we must also remember that rhetoric has also been a fragmented discipline throughout history: due to the opposition between

11 Discourse Analysis: The Constructivist Perspective and …

195

rhetoric and philosophy, between rhetoric and religious dogmatism, and between rhetoric and rationality (Perelman 1997; Meyer 2008; Pujante 2011, 2017a). Today, it needs to regain its unity, after its re-emergence as a force in the second half of the twentieth century. As Meyer points out (2008: 23ff), it needs to be a discipline that comprehensively addresses ethos (the skill of the speaker), logos (the study of discourse itself, reasoning) and pathos (emphasizing the relationship with the audience, by means of attention to the emotions) (Martín Jiménez 2014; Molpeceres Arnáiz 2014). The rhetorical relationship rests on these three components, and as such there is no reason for its fragmentation, as occurred in the past. Pujante (2017a) considers that at present, rhetoric can only be unified and revitalized within a constructivist paradigm that he calls “constructivist rhetoric”. This same need for unification also requires reconciliation between rhetoric and non-formal argumentation, as attempted by Perelman and Olbrechts-Tyteca (1958), although these authors based their work on the centrality of the logos. Likewise, pragmatics (in its interactional perspective) is essential in an area like the study of argumentation, as argued by van Eemeren and Grootendorst (2004) in their discussion of Pragmadialectics (Pujante and Morales-López 2009). In the communicative process, argumentation is related to acceptability and the need to have an effect on those being addressed; as a result, van Eemeren and Grootendorst (2004: 12) consider the argumentative process to be a complex speech act: “the performance of the complex speech act of argumentation [which] aims to convince a reasonable critic of a certain standpoint.” The pragmatic tradition is also important for recognizing that the general aim in the communication process is to reveal the meaning that is constructed and which is also relevant to the participants, while being aware that this meaning is mostly activated inferentially. The socio-constructivist tradition of Goffman (1974) and the proposal by Gumperz (1982) therefore remain very topical—especially Gumperz’s notion of contextualization cue—a kind of linguistic or generally semiotic trail that activates the various inferences in a discourse.1 The common characteristic of these various disciplines is that they emphasize the study of discourse based on both a functional and a constructivist approach. The first perspective, functionalism, was also the original focus of the anthropological tradition, according to Blommaert and Jie (2010: 7): “language from an anthropological perspective is almost necessarily captured in a functionalist epistemology.” The second, constructivism, is clarified by authors such as Gumperz (2001: 218) when he says that his goal is not only to determine the meaning that is conveyed, but also to discover the interpretive processes in order to relate them to the linguistic processes through which the former are negotiated. The semiotic perspective in discourse analysis needs to be revitalized, especially when the use of new technologies increasingly favours the construction of multimodal discourses (Kress 2010). Human communication is already in itself multi-

1 Indeed,

with the death of Professor Gumperz, I would like to take advantage of this reference to pay tribute to him; and also to remember his kindness during the seminars that I attended in 1990.

196

E. Morales-López

modal (voice, gesture, gaze, etc.), but the Internet has been added to the landscape, as well as the voice and the written text, to construct new meanings simultaneously: When I speak of ‘meanings’, I refer to semiotic resources encountered in the many communities in which I have participated and in which I now lead my life… There are meanings of different kinds at different levels… The image strongly suggests that it is impossible to separate the conceptual – not to mention the cognitive – and the affective. Affect is inevitably part of such mediation; it must have a central place in this theory of meaning. Indeed, my inclination is to erase the boundary of affect and cognition in this frame…. (Kress 2010: 109)

Peirce’s traditional division of signs into icons, indices and symbols (as iconic, indexical and symbolic) takes Kress’s idea (2010: 64–65), which he calls Multimodal Social Semiotics, one step further. In his proposal, arbitrariness is replaced by motivation, in all instances of sign-making and for any kind of sign, due to the fact that in the relationship between form and meaning, the strength of convention (socially imposed) is as important as transparency (or motivation) in the recognition of the relationship between form and meaning in communication.2 Eco (1976: 59) reminds us that the classification of a sign as an index, icon or symbol depends on the circumstances in which it appears and the meaningful use to which has been allocated: “Hence I can use the historic photograph of those executed during the Paris Commune as both an arbitrary and conventional symbol of “revolutionary martyrs”, or as an icon or as an index, in the sense of an ‘imprint’ which testifies to the veracity of a historical fact.” The study of the sign in relation to the message, discourse, and society is therefore as important as the study of the sign in relation to codes; in other words, “the cycle of semiosis, the life of communication, and the use and interpretation made of signs; it is society which uses signs to communicate, to inform, to lie, to cheat, to dominate and to release” (Eco 1976: 19–20). However, we must remember that Eco made this argument in favour of the relationship between semiotics and communication in the 1970s, at the same time as arguing that semiotics had to be a discipline of the code, in the same way as linguistics was for the a system of a language (1976: 136). It was still the era of structuralism. In a more recent work, Eco (2011: 248–249) emphasizes the process of construction in every sign, since reality is a segmentable continuum. In addition to articulating pragmatic-discursive, rhetorical-argumentative and semiotic perspectives, in our opinion it is necessary to add another dimension to complete the cycle of meaning: the cognitive perspective (Pujante and MoralesLópez 2013). We advocate a socio-cognitive vision of discourse, based on which the various pragmatic-discursive, rhetorical-argumentative and semiotic resources selected by social actors are placed at the service of the construction of the world

2 As

an example of the importance of this transparency between form and meaning, Kress uses the choice made by two chess players when they discover they have lost one of their pieces. Due to the strength of convention, they know that any object could replace the missing piece, but because of the strength of transparency, they would choose an object with a similar shape, size and colour to the original piece.

11 Discourse Analysis: The Constructivist Perspective and …

197

activated within a given cognitive framework.3 If this was not the case, the study of the communicative cycle would be incomplete, because we would be excluding a key process and fundamental organ in the relationship between knowledge and human action—the process between the mind and the brain. At this point, it is necessary to address the relationship with one of the key concepts in the philosophical and cognitive traditions—that of representation (Moscovici 1981). When we refer to this notion today, we turn to the interpretation of the concept by Wilson (2001: 401), “representation refers to the issue of how language is employed in different ways to represent what we can know, believe, and perhaps think.” Two opposite positions have been considered as a response: First, the realist position that postulates the existence of a reality outside language; a reality we become aware of by means of a series of universal primitive concepts; from this position, language is the (independent) vehicle for the transmission of thought. According to Edwards (1997), this approach uses the mapping metaphor to explain the relationships between the world, the mind and language; this metaphor also reduces language-thought relations to individuals and their mental processes. Second, the relativist position that advocates the relationship between language and thought; from this perspective, the experience of the world is not a given in advance, but mediated through language. Molpeceres Arnáiz (2014: 9) refers to these two traditions respectively as the tradition of logical-philosophical thought and the tradition of rhetorical-symbolic thought. Wilson (2001: 401) also describes how many discourse analysts are aware of the mediating role of language, but still advocate a realistic epistemological position, according to which the function of discourse analysis mainly consists of revealing those deviations from previously agreed realities that are external to discourse; hence their interest in the institutional discourses that violate democratic principles that are assumed to be universal.4 However, it is also possible to look at another different tradition in the constructivist relationships between language, thought and reality. One of the first references is in Vico (1744/2006). And in the twentieth century, one branch of this tradition comes from Bartlett, Mead, Bateson, Goffman and Gumperz, subsequently reformulated in cognitive terms by Lakoff (2004). Another branch has precursors in Bakhtin, Voloshinov and Vygotsky (Scollon and Wong Scollon 2005); and is also present in the theoretical works of Berger and Luckmann, Castoriadis, Bourdieu and White, among others (references in Morales-López 2011, 2014, 2016b; Morales-López and Floyd 2017). When we use the concept of the frame in our research, rather than the traditional concept of representation, we are not talking about a simple change in terminology; we are referring to a different theoretical model, which some authors have called post3 I refer to an earlier work (Morales-López

2011) to explain the origin of the term framework in the American tradition. In Montesano-Montessori and Morales-López (2015), we have linked the term framework with the concept of narrative or narrativity (Somers 1994). 4 In Montesano-Montessori and Morales-López (2015), we addressed the relationship between these two traditions, and their application to an analysis of the discourses of social change.

198

E. Morales-López

cognitivism (Gomila and Calvo 2008), in which cognition is considered as a unified process that is the result of the relationship between the many factors mentioned above. The process of signification in discourse is part of this more holistic cognitive process. This position advocates the idea that the experience of the world that we express through language is not given in advance, but instead mediated through it. In other words, the discursive process involves a construction of a dialectical relationship with the subjectivity of the actors (their emotions), their actions and their surroundings. Cognitive processes, including the process of signification that emerges in our communicative interactions, cannot therefore be separated from our biological characteristics, or from the socio-cultural relations in which we are immersed. For Maturana and Varela (1992: 241), this centrality of communicative interactions in the action of knowing implies having found the middle path between two extremes: objectivism and idealism. Capra (1996: 300) recalls that the Latin origin of the word consciousness is conscire: ‘knowing together’; consciousness is never a completely individual process, but is instead social and cultural at the same time. These ideas about cognition have therefore been called embodied cognition or distributed or networked cognition (see also Varela et al. 1991/1996). In its conclusions, this post-cognitive period is consistent with neurological research on the relationship between knowledge and emotions. One of these investigations is that of Antonio Damasio, who in one of his books (2010) used a very interesting idea—the image—which could be considered a correlate in neuroscience of the socio-cognitive concept of the frame. According to Damasio (2010: 431–434), images are like momentary maps that the brain creates of something; some of them are of real things that happen outside the brain, but others are reconstructed from memory, according to a selection process based on their value, in which the emotions play a key role. This image creation process is not only used in the biological regulation of life, but in any other type of higher knowledge. We therefore see how different disciplines converge on the same idea: the unitary view of cognition as a process that includes various deeply interconnected processes, providing a systemic vision of human life.

11.4 Analysis of the Data In order to briefly illustrate the theoretical framework and methodology discussed above, I have selected for analysis the statement by member of the Catalan cooperative CIC, a post-capitalist socioeconomic alternative, that emerged in Catalonia in the wake of the 15M social movement (although the project had begun several years earlier). In Spain, the 15M movement was a response to major cuts in funding for social services in the aftermath of the 2008 economic crisis. This social group takes its name from a spontaneous peaceful protest in the main squares of Madrid and Barcelona

11 Discourse Analysis: The Constructivist Perspective and …

199

that took place on 15 May 2011, before spreading to other cities across the country (The group is also referred to as the ‘Outrage Movement’ (los indignados) after Stéphane Hessel’s 2010 essay, Indignez-vous). While Spain is no stranger to the struggle for democracy, the 15M group presented a number of new characteristics, including its emphasis on peaceful resistance and the movement’s imaginary of a new democracy or world view, conveyed through inventive placards and slogans designed by the citizens themselves and posted in public squares and spaces and/or on various websites (Pujante and Morales-López 2013). Our main hypothesis (see also Montesano-Montessori and Morales-López 2015) is that their various messages (in combination with the numerous demonstrations and the overall dynamic of the movement) functioned in the first instance, as a sign of protest, but also as a way to reframe the population’s understanding of the economic and social crisis, and to rearticulate the identity of the country’s citizens, transforming them from victims into agents. The following is a verbatim excerpt from the oral contribution by a member (a middle-aged man) of the CIC to a video documentary about its establishment and development, in which the activist describes his participation in this process in a personal way: (1a) Yo veía que había cosas que no funcionaban bien en la sociedad para las personas, pero no me llamaba la atención a mí hacer la función de protesta… Entonces, cuando realmente me involucré en movimientos sociales, fue cuando se dio la oportunidad de crear cosas desde la gente y para la gente… Estamos en un proceso de desaprender hábitos y conceptos que traemos; entonces, la idea de trabajo, pues estamos desarrollando otras formas de entenderlo quizás… English translation: ‘I saw that there were things that did not work properly in society for people, but the action of protest didn’t attract my attention… So when I really got involved in social movements was when the opportunity came up to create things from the people and for the people… We are in a process of unlearning habits and concepts that we bring; hence the idea of work, as we are perhaps developing other ways to understand it…’ (1b) Yo siento una diferencia enorme de cómo vivía antes de estar involucrado en la red y en este proyecto y aquí tenemos la oportunidad de crear (nuestras) propias condiciones de desarrollo, también de prosperidad, ¿por qué no? Y también de realizarlo a la manera que cada uno pues le motive más… En general, la idea, lo que intentamos es buscar un equilibrio en el cual todo lo que hagamos lo podamos disfrutar y que no nos sintamos forzados a hacer algo”. English translation: ‘I feel an enormous difference between how I used to live before I was involved in the network and in this project, and here we have the opportunity to create (our) own conditions for development, and for prosperity, why not? And what’s more, to do it in the way that motivates everyone most… In general, the idea, what we are trying to do, is to find a balance in which everything we do is something we enjoy and we do not feel forced to do something.’

Ethnographic work (by means of ongoing participatory observation within a social group) shows the significance of the discursive data selected from a range of many other discourses to which the researcher is exposed for a specified period of time (Morales-López 2012a, 2016a). In my case, I carried out the observation in the CIC, attending its various activities in the spring of 2014.

200

E. Morales-López

The opinion of the activist selected seemed significant to me because his contribution enabled me to consider the various components of discursive process that we are analyzing. This person gives his opinion as a member of the cooperative in a promotional video for the CIC. It is therefore a discursive genre that has a specific communicative function: conveying the aims of the CIC in order to convince and persuade the public of the need to change a specific socio-economic model and if they wish, to join their project. The relationship between what the activist says and the communicative purpose of the discursive genre of which his contribution forms a part thus builds his participation status; that is, his social role of principal (in Goffman’s terms, 1981: 145), in this particular organization. The communicative function of the contribution also needs to connect with the local and global context and the socio-political action of the CIC. The local context provides the specific relevance for this activist’s opinion, within the opinions of other colleagues. This person is part of one of the cooperative’s most creative and innovative initiatives, as the CIC became consolidated after a massive urban social protest movement, the 15-M, after the sit-in protest in Plaça Catalunya, in the centre of Barcelona. The activist, together with others, is part of a community initiative to return to rural life (the majority have no prior knowledge of the agricultural world), albeit involving some of them continuing their previous professional work. This local context gives meaning to the statement in (1a): “We are in the process of unlearning habits and concepts that we bring; hence the idea of work, as we are perhaps developing other ways to understand it…” Inferentially, we can see that the process of unlearning which he is referring to is a new (post-capitalist) means of resolving the divorce between the rural and urban world engendered by capitalism. The global context is also present in this contribution. The CIC is not an isolated initiative for socio-political socio-economic change, but is instead connected with many movements for autonomy and self-management that have been emerging in many parts of the world: the best known is the Zapatista movement in Chiapas, which they refer to explicitly. Hence the group’s interest in the communication and dissemination of its messages through its website (www.cooperativa.cat), among many other initiatives. The activist mentions that he became involved in “the network” (simultaneously local and global) (1b); this is a significant lexical term in this context, when referring to the cooperative, because it alludes to how this socio-economic initiative works: in collaborative networks, as opposed to competitive principles of capitalism. From the local context (a rural innovation initiative), we have therefore moved into the global context (one involving independent but interconnected social groups working through social networks in different parts of the world); and from there to a certain degree of socio-political action that is simultaneously local and global (glocal) and that is currently being formulated as an alternative to neoliberal capitalism. Finally, he also discusses how his participation in this initiative for specific social change came about; to do so he reveals something of his subjectivity to us. He does not “introduce himself” in the public space (Goffman 1974) as the prototypical activist. In this context, this is what would be expected in his role as a spokesperson for the CIC (mentioned above); in other words, a person acting in situations of protest,

11 Discourse Analysis: The Constructivist Perspective and …

201

condemnation and even disobedience before the status quo. However, this person says in (1a) that “the action of protest did not attract my attention”; he thus shows us a new side of being an activist: someone who is able to create new types of coexistence, new livelihoods by means of manual labour (in the face of the current high unemployment levels due to the crisis) and above all, a balance between effort and the ability to enjoy this work. The terms and expressions used, such as create things from the people and for the people, development, prosperity, motivation, enjoyment produced a cognitive framework in which this manual work is presented in a positive way, because it is an action performed freely and in solidarity and cooperation with other people with whom he shares ideals. According to Maturana and Varela (1992: 241), this cognitive frame has revealed his own world, as the one which he brings forth in coexistence with others. From the argumentative point of view, this contribution begins with an initial assumption in which he describes his personal situation as an observer of social reality: “I saw…. didn’t attract my attention…”. He acted as a passive but not completely indifferent observer, as indicated by the verbal lexeme “see”. This is followed by an implicit cause-effect argument, which he uses to describe, by means of a construction of time (“When I got involved…”), the cause (the opportunity to “create things from the people and for the people”) that led him to react (the consequence). And in (1b) he presents the conclusion of his decision and experience using emotional premises, in which he mainly describes his psychological state: “I feel an enormous difference…” It is a pragmatic argumentative conclusion (Perelman and Olbrechts-Tyteca 1958) in which the value of an act or event is appreciated according to its favourable or unfavourable consequences. The argumentative force is therefore on this subjective level, and also coincides with the emotional tone of the lexical selection that we have previously discussed. The relational analysis of the various components activated in this brief discursive fragment finally appears to show an example of a fractal structure (Massip-Bonet, 2013: 4): the ideological principles on which the cooperative is based, such as freedom, self-management, post-capitalism, cooperation and solidarity emerge in this fragment of discourse through the connection of the discursive form with the constructed meaning and the context. The whole (the CIC) is therefore replicated in the part (the activist) although he also uses his individual creativity in order to explain the particular role that he wants to perform in the social construction of the CIC. The reader can consult the full analysis of the data in the following studies: Morales-López (2016b, 2017a).

11.5 Conclusions In this study, we have examined how discourse analysis, considered from the perspective of complexity theory, entails the inclusion of the following principles: (a) the dialectical relationship between form and function, as well as the other components of communication: the subjectivity of social actors, the (more or less comprehensive)

202

E. Morales-López

context and human action; (b) the constructionist perspective, according to which reality is constructed (i.e. emerges) in communicative action; and (c) transdisciplinarity as overcoming the traditional division in the humanities in which discourse (or communication) is the object of study. This implies that the discourse analyst requires a solid background in the pillars of functionalist linguistics, while acknowledging that this dimension is only a part of what his/her theoretical-methodological framework should be. The remainder must be completed (and integrated) with research on the rhetorical tradition, the theory of argumentation and semiotics. Furthermore, this person will need to consider research in other disciplines that sheds new light on the construction of reality in specific communicative activities and actions. Aknowledgments This research is part of the projects RECDID and CODISCO, financed by the Spanish Ministry of Economy and Competition, and European Feder Funds (FFI2013-40934-R and FFI2017-85227-R; periods: 2014–2017 and 2018-2020; website: http://cei.udc.es).

References Albaladejo, T. (2013). Rhetoric and discourse analysis. In I. Olza, Ó. Loureda, & M. Casado (Eds.), Language use in the public sphere: Methodological perspectives and empirical applications (pp. 19–51). Frankfurt am Main: Peter Lang. Blommaert, J., & Jie, D. (2010). Ethnographic fieldwork. A beginner’s guide. Bristol: Multilingual Matters. Bourdieu, P. (1990). Language as symbolic power. Cambridge: Polity Press. Capra, F. (1996). La trama de la vida. Una nueva perspectiva de los sistemas vivos. Barcelona: Anagrama. Castells, M. (2012). Redes de indignación y de esperanza. Madrid: Alianza. Couceiro Bueno, J. C. (2012). La carne hecha metáfora. La metaforicidad constituyente del mundo. Barcelona: Edicions Bellaterra. Damasio, A. (2010). Y el cerebro creó al hombre. ¿Cómo pudo el cerebro generar emociones, sentimientos y el yo? Barcelona: Destino [Original title: Self comes to mind]. de Beaugrande, R. (1996). The story of discourse analysis. In T. A. van Dijk (Ed.), Introduction to discourse analysis (pp. 35–62). London: Sage. Delgado Díaz, C. J. (2007). Hacia un nuevo saber. La bioética en la revolución contemporánea del saber. La Habana: Acuario. de Sousa Santos, B. (2011). Las epistemologías del sur. Retrieved from http://www. boaventuradesousasantos.pt/media/INTRODUCCION_BSS.pdf. Duranti, A. (1997). Antropología Lingüística. Madrid: Cambridge University Press. (2000). Eco, U. (1976). Signo. Barcelona: Labor. Eco, U. (2011). Kant, Peirce, and the platypus. In F. Stjernfelt & P. F. Bundgaard (Eds.), Semiotics. Critical concepts in language studies (Vol. 1, pp. 229–283). London: Routlege. (Philosophy). Edwards, D. (1997). Discourse and cognition. London: Sage. Goffman, E. (1974). Fame analysis: An essay on the organization of experience. Nueva York: Harper & Row. Goffman, E. (1981). Forms of talk. Philadelphia: University of Pennsylvania Press. Gomila, T., & Calvo, P. (2008). Directions for an embodied cognitive science. Toward an integrated approach. In P. Calvo & T. Gomila (Eds.), Handbook of cognitive science: An embodied approach (pp. 1–25). Amsterdam: Elsevier.

11 Discourse Analysis: The Constructivist Perspective and …

203

Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press. Gumperz, J. J. (2001). Interactional sociolinguistics: A personal perspective. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 215–228). London: Blackwell. Halliday, M. A. K. (1970/2002). Language structure and language function. In M. A. K. Halliday (Eds.), On grammar (pp. 173–195). London: Continuum. Halliday, M. A. K. (1982). Exploraciones sobre las funciones del lenguaje. Barcelona: Editorial Médica y Técnica, S. A. Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London: Longman. Juarrero, A. (1999). Dynamics in action. Intentional behavior as a complex system. Cambridge Mass.: The MIT Press. Kress, G. (2010). Multimodality. A social semiotic approach to contemporary communication. London: Routledge. Lakoff, G. (2004). Don’t think of an elephant. Know your values and frame the debate. Vermont: Chelsea Green Publishing. Larsen-Freeman, D., & Cameron, L. (2008). Complex systems and applied linguistics. Oxford: Oxford University Press. Martin, J. R. (2001). Cohesion and texture. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 34–53). Malden, Mass.: Blackwell. Martín Jiménez, A. (2013). El componente retórico y el componente simbólico en la publicidad. Análisis de los anuncios de energía eólica de Iberdrola. Cuadernos de Investigación Filológica, 39, 159–186. Retrieved from http://publicaciones.unirioja.es/ojs-2.4.2/index.php/cif/article/ view/2560/2386. Martín Jiménez, A. (2014). La retórica clásica y la neurociencia actual: las emociones y la persuasión. Rétor, 4(1), 56–83. Retrieved form http://www.revistaretor.org/pdf/retor0401_jimenez. pdf. Massip-Bonet, À. (2013). Language as a complex adaptative system: Towards an integrative linguistics. In À. Massip-Bonet & A. Bastardas-Boada (Eds.), Complexity perspectives on language, communication and society (pp. 35–60). Berlin & Heidelberg: Springer. Maturana, H. (1996). La realidad: ¿objetiva o construida? (Vol. 1). Barcelona/México DF: Anthropos, Universidad Iberoamericana. Maturana, H. (2006). Self-consciousness: How? when? where? Constructivist Foundations, 1(3), 91–102. Maturana, H., & Varela, F. (1992). The tree of knowledge. The biological roots of human understanding. Boston: Shambhala. Meyer, M. (2008). Principia rhetorica. Una teoría general de la argumentación. Madrid/Buenos Aires: Amorrortu. (2013). Molpeceres Arnáiz, S. (2014). Mito persuasivo y mito literario. Bases para un análisis retóricomítico del discurso. Valladolid: Universidad de Valladolid. Montesano-Montessori, N., & Morales-López, E. (2015). Multimodal narrative as an instrument for social change: Reinventing democracy in Spain -the case of 15M. Critical Approaches to Discourse Analysis Across Disciplines (CADAAD), 7(2), 200–219. Retrieved from http://www. cadaad.net/journal. Morales-López, E. (2011). Hacia dónde va el Análisis del Discurso. Tonos Digital, 21. Retrieved from http://www.um.es/tonosdigital/znum21/secciones/estudios-21-discurso.htm. Morales-López, E. (2012a). Discourses of social change in contemporary democracies: The ideological construction of an Ecuadorian women’s group based on “solidarity economy and finance”. Text and Talk. An interdisciplinary Journal of Language, Discourse and Communication Studies, 32(3), 329–348. Morales-López, E. (2012b). Atos de fala e Argumentação: um debate entre uma companhia transnacional (Repsol) e ativistas em um site. EID&A (Revista Electrônica de Estudos Integrados em Discurso e Argumentação), 3. Retrieved from http://www.uesc.br/revistas/eidea/espanol/index. php?item=conteudo_revistas_eletronicas.php.

204

E. Morales-López

Morales-López, E. (2012c). Análisis de discursos ideológicos en la empresa: La deslegitimación y la defensa de las energías renovables. FORUM. Qualitative Social Research, 13(3), Art. 20. Retrieved from http://www.qualitative-research.net/index.php/fqs/article/view/1864. Morales-López, E. (2014). La reflexión metadiscursiva como función comunicativa en el proceso de construcción de discursos de cambio social en un grupo de mujeres. Sociolinguistic Studies, 8(2), 249–269. Morales-López, E. (2016a). De la perspectiva etnográfica al análisis crítico del discurso: investigación en un grupo de mujeres ecuatorianas. In B. Crespo, I. Moskowich, & C. Núñez-Puente (Eds.), Queering women’s and gender studies (pp. 45–66). Newcastle: Cambridge Scholar Publishing. Morales-López, E. (2016b). Frame construction in post-15M speeches. Res Rhetorica, 1, 50–67. Retrieved form http://resrhetorica.com/index.php/RR/article/view/2016-1-4. Morales-López, E. (2017a). Cognitive frames, imaginaries and discursive constructions: Post 15M’s discourses with reference to eco-social alternatives. In E. Morales-López & A. Floyd (Eds.), Developing new identities in social conflicts: Constructivist perspectives on discourse studies (pp. 249–272). Amsterdam: John Benjamins. Morales-López, E. (2017b). Epilogue. In E. Morales-López & A. Floyd (Eds.), Developing new identities in social conflicts: Constructivist perspectives on discourse studies (pp. 273–284). Amsterdam: John Benjamins. Morales-López, E., & Floyd, A. (2017). Developing new identities in social conflicts: Constructivist perspectives on discourse studies. Amsterdam: John Benjamins. Morin, E. (1990). Introducción al pensamiento complejo. Barcelona: Gedisa. Moscovici, S. (1981). On social representations. In J. P. Forgas (Ed.), Social cognition. Perspectives on everyday understanding (pp. 181–209). New York: Academic Press. Nespereira García, J. (2014). Estrategias discursivas en la comunicación de crisis sanitarias (Retórica y Teoría de la Argumentación): el caso de la gripe A en 2009. (Doctoral dissertation) Universidad de Valladolid. España. Nicolescu, B. (2007). La transdisciplinariedad, una nueva visión del mundo. Retrieved from http:// nicol.club.fr/ciret. Perelman, C., & Olbrechts-Tyteca, L. (1958/1989). Tratado de la argumentación: La Nueva Retórica. Madrid: Gredos. Perelman, C. (1997). L’empire rhétorique. Rhétorique et argumentation. París: Librairie Philosophique J. Vrin. Pujante, D. (2003). Manual de Retórica. Madrid: Castalia. Pujante, D. (2011). Teoría del discurso retórico aplicada a los nuevos lenguajes. El complejo predominio de la elocutio. Rétor, 1(2), 186–214. Pujante, D. (2017a). The discursive construction of reality in the context of rhetoric: Constructivist rhetoric. In E. Morales-López & A. Floyd (Eds.), Developing new identities in social conflicts: Constructivist perspectives on discourse studies (pp. 42–65). Amsterdam: John Benjamins. Pujante, D. (2017b). I am, I am not Charlie. The discursive conflict surrounding the attack of Charlie Hebdo. In E. Morales-López & A. Floyd (Eds.), Developing new identities in social conflicts: Constructivist perspectives on discourse studies (pp. 83–106). Amsterdam: John Benjamins. Pujante, D., & Morales-López, E. (2009). Los aspectos argumentativos de las respuestas de Rajoy a un grupo de ciudadanos en el programa de televisión española Tengo una pregunta para usted. Oralia, 12, 359–390. Pujante, D., & Morales-López, E. (2013). Discurso (discurso político), constructivismo y retórica: los eslóganes del 15-M. Language, Discourse & Society, 2(2), 32–59. Retrieved from http://www. language-and-society.org/journal/issues.html. Salvador, V. (2014). El debate social sobre las fuentes de energía: representaciones semánticas y gestión social de los conocimientos. Culture, Language and Representation, 13, 221–243. Salvador, V., Macián, C., & Marín, M. J. (2013). La construcción de las profesiones sanitarias a través de las revistas especializadas. Discurso & Sociedad, 7(1), 73–96.

11 Discourse Analysis: The Constructivist Perspective and …

205

Scollon, R. (2008). Analysing public discourse. Discourse analysis in the mapping of public policy. London: Routletge. Scollon, R., & Wong Scollon, S. (2000). Discourse and intercultural communication. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 538–547). Malden, Mass.: Blackwell. Scollon, R., & Wong Scollon, S. (2001). Intercultural communication. Malden, Massachusetts: Blackwell. Scollon, R., & Wong Scollon, S. (2005). Lighting the stove. Why isn’t enough for critical discourse analysis. In R. Wodak & P. Chilton (Eds.), A new agenda in (critical) discourse analysis (pp. 101–117). Amsterdam: John Benjamins. Somers, M. R. (1994). The narrative constitution of identity: A relational and network approach. Theory and Society, 23, 605–649. van Eemeren, F., & Grootendorst, R. (2004). A Systematic theory of argumentation. The pragmadialectical approach. Cambridge: University Press. Varela, F., Thompson, E., & Rosch, E. (1991/1996). De cuerpo presente. Las ciencias cognitivas y la experiencia humana. Barcelona: Gedisa. Verschueren, J. (1999). Para entender la pragmática. Madrid: Gredos. (2002). Vico, G. (1744/2006). Ciencia nueva. Madrid: Tecnos. Vilarroya, Ó. (2014). Una comunicació sense informació basada en vivències. In À. Massip & A. Bastardas-Boada (Eds.), Complèxica. Cervell, societat i llengua des de la transdisciplinarietat (pp. 39–54). Barcelona: Universitat de Barcelona. Vygotsky, L. (1934/1986). Pensamiento y lenguaje. Buenos Aires: La Pleyade. White, H. (1987). El contenido de la forma. Narrativa, discurso y representación histórica. Barcelona: Paidós. Wilson, J. (2001). Political discourse. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 398–415). Malden, Mass.: Blackwell.

Chapter 12

A Complex Approach to Prosodic Discourse Variation Raquel García Riverón and Alejandro F. Marrero Montero

Abstract The confluence of several fields of knowledge, of the work of several linguistic and non-linguistic disciplines allow us to approach a holistic perspective of our research on speech within the complexity theory. It allows us to design the interconnections that emerge from language within the interactions of its means of expression. This perspective leads us to find a new phonological interpretation of prosody, lexical, grammar and speech systems in virtue of intention through speech multidimensional analysis. This also makes possible to explain and describe the bases of a group of semantically and pragmatically founded attractors defined for the study of intonation and a new concept of underlying structure formed by emergent features from the morphogenetic processes of the language systems. This is shown in a case study of a dialogue from the Cuban movie Fresa y chocolate.

12.1 Linguistic Thought and Complexity At a given time the development of the science of language, the methods and techniques established from the more traditional view, the plain vision of language system, are not enough in the research of such a complex object as language. Then, since the last century some approaches have emerged, approaches that take into account the phenomena that arise from the boundaries of different disciplines. The object of study extends to different types of discourses and communicative situations under the influence of the ethnography of communication and discourse 1 analysis ; psycholinguistics based on the activity theory (Leontiev 1962; Vigotskii 1 An

overview in Calsamiglia and Tuson (2004) and Morales Lopez (2011).

R. García Riverón (B) Grupo-Red de Complejidad y Lenguajes de La Habana, Havana, Cuba e-mail: [email protected] A. F. Marrero Montero (B) Universidad Central “Marta Abreu” de Las Villas, Santa Clara, Cuba e-mail: [email protected] URL: http://www.uclv.edu.cu © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_12

207

208

R. García Riverón and A. F. Marrero Montero

1966); Van Dijk’s work (1983, 1997, 2003), especially those texts that advocate a multidisciplinary perspective of discursive phenomena; the second Wittgenstein’s treatises (1973); Austin (1962) and Searle (1969), who formulated the theory of speech acts. Likewise, we take these speech acts as a sign of an intention according to their semantic and pragmatic determination.

12.1.1 Objectives The main objective of the ideas that are outlined below is to show some ways to approach the relationship of the expression’s means of intention in language (grammatical, prosodic, and non-verbal means), intention defined as determination of the will towards an end. As it has been stated, the intention is important and this fact is reflected in discourses, as shown in the bibliography, or in the results that have been obtained from the viewpoint of prosody, in our research.

12.1.2 What Is the Thought of Complexity?2 Postmodernism was reinforced since the second half of the last century as an epistemological reconstruction period. First, the notion of totality was recovered, the holistic vision. With the disciplinary barriers being challenged, the whole is currently not seen as sum of the parts. Quite the contrary, the components of the whole are analyzed in interaction processes that take place within the network. This view leads the researcher to focus on the key points where the intersections between different disciplines occur, the joints (engarces, in Spanish), as we call them. The scientific ideal is placed in transdisciplinarity.3 This transdisciplinary approach is for us a general form of knowledge that is based on inquiry tools in constant development, according to what the different ways of examining the object indicate.4 The notions of reflexivity, object and subject vary, and we begin to talk about omniobjetivity. The researchers are now considered the product of the social interactions in which they are immersed, of the powers and knowledge that surround them and in 2 On

the new epistemology, the different variants and specifically on the thought of complexity, we will only mention some key works: Prigogine (1983); Capra (2002); Goodwing (1998); Juarrero (1999); Morin (1994); Kauffman (2003); Kennealy (2010); among others. 3 Many researchers define multidiscipline as the “convergent research effort in various disciplines”, such as biophysics or sociolinguistics. Interdisciplinarity presupposes multidiscipline insofar as it has the same goals, but it involves the determination of a new object of study. It is the case of genetic engineering. Transdisciplinarity includes the two previous ways of doing science, but it tries to find points of contact between the different branches, finding these points of contact in a step by step analysis of the different possible interactions. 4 A proposal of these tools for this kind of intonation studies can be found in García Riverón (2005a).

12 A Complex Approach to Prosodic Discourse Variation

209

which they are formed. As a result, their reality is reflected in their relationship with the object of inquiry. It also varies the criterion of determinism in science to include chance, randomness and uncertainty that explain many of the facts that could not be apprehended from rigid and determinist positions.

12.1.2.1

Differences Between Simple Systems and Complex Systems

We see that the semantic value we have referred as complexity is wide-ranging, although it has become popular in many disciplines and even in everyday life. As expected, we can find many definitions of this view of science and thought; however, we will only present the main turning points of relevance for our work. We will also list consecutively some of the characteristics of simple systems and complex systems so that the reader could grasp the subject we deal with more easily. The experienced reader will immediately extrapolate these paradigmatic features to many of the models that have been applied in language studies: Let us first see the features of simple systems: • Simple systems are closed. That is, they do not interact with the environments or the contexts in which they occur. • Simple, closed systems are composed of very few variables. • They are static systems in equilibrium. • Usually, they are studied in isolation because researchers do not denote the connections with other types of systems that might be in their environment and therefore could influence the studied dynamics. Then, the obtained research results may be influenced by many other conditions that are not taken into account. • Inasmuch as this type of system operates with simple phenomena, they are deterministic, meaning, a cause always involves the same effect. At the same time, they are considered as linear systems which stipulate that the whole is the sum of the parts. • This type of system, being understood as a closed system, could not dissipate energy because this concept eliminates systems’ interaction with the environment conducive to these phenomena. • The concept of invariant in simple systems is immovable, unchangeable and nondynamic. • The concept of time is seen as exogenous, external to the system itself. • The strict and linear order is the main feature of this type of system. In contrast, let us see some of the properties of complex systems: • Complex systems are developed with possible bifurcations, that is, unexpected changes in the course of the interactional processes. In a complex non-linear dynamic system, constant flows take place. There can be flows of: mass, understood as matter or substance; energy, understood as psychic or other type of energy, or as a higher and dynamic form of integration of matter; information, which occurs in recurring encoding and decoding processes; sense, which unfolds in different ways, depending on the nature of the given system.

210

R. García Riverón and A. F. Marrero Montero

• These systems are far from equilibrium, on the edge of chaos. This implies that the system can change at any time. They constantly move between order and disorder, which is its fundamental characteristic. • They constantly dissipate energy in their interactions with the environments. • The complexity of their structures and interactions supports the fact that changes in these systems are irreversible. • Every complex system consists of many other systems that, according to authors, are called microsystems or subsystems. At the same time, the subsystems are constituted by elements which also form systems. • All these systems interact as a sort of network in which each system generates fairly similar features in each of the nodes of the network. To some extent, these systems are self-similar. • In complex systems, the criterion of invariant is dynamic and it is defined as a set of relationships that scientists mark as trajectories around some specific information. It is important to note that many of the models in use in our discipline are applied as standard, simple models with very few variables and considering only one of the levels of language. However, the thought of complexity has shown that the behavior of interactions is not determined in a single level. Consequently, if only one level or just the levels that lie beneath it are studied, it is not possible to define the essence of the system in question. This is the fundamental challenge of the study of language as a non-linear, dynamic and complex system. To avoid these limitations, we use the method of the sliding lens, in which we take into account the dynamics established as being fundamental, and we try to point out, although as a hypothesis, the joints with the underlying levels and environments.

12.1.2.2

Key Concepts of the Theory of Complexity

To conclude this introductory part, we will define some essential terms when we face the study of these new realities: Emergence In dynamic and complex systems, new properties appear, arise, and emerge as a consequence of interactions between the different variables and also as a result of interactions with the environment. The emergence of new properties in this type of system is called emergence process. The resulting dynamics is not reducible to any of the interacting parts, but a new reality, a new totality. Self-organization As a result of the interaction of the parts and the emergence processes, new structures, patterns, interactions and senses are created, and deserve to be examined. The capacity of these systems to move from order to disorder and vice versa, and to keep their goals is called system self-organization.

12 A Complex Approach to Prosodic Discourse Variation

211

Attractor The term attractor arises in mathematics given the need for specialists to solve nonlinear equations. In this type of equations, a figure or shape appears whose pattern shows the most common trends of a variable in chaotic motion. In other disciplines, such as semantics, the term shows trends or propensities of sense within a semantic space or field of fuzzy boundaries.

12.2 The “Grupo-Red de Complejidad y Lenguajes de La Habana” (Group-Network of Complexity and Languages of Havana)5 : From Prosody to Semantics, from Semantics to Discourses and Environments The thought of complexity has allowed us to realize that several features of the actions of sense/meaning acting in discourses are emergent properties of self-organizing phenomena in the discursive activity. This self-organization is given by the specific relations of the noesis-semiosis processes, which are nucleated around semantic and pragmatic attractors that can be defined in discourses and encompass intentionality, consciousness, psychosocial and sociocultural variables, among others. We started from intonation. Typically, theories and methods in use in linguistic studies analyze the system of intonation and language in a general way, according to different perspectives and degrees of complexity. Intonation patterns, which as a rule can be classified as being simple, can be summarized in two major groups: those based only on the acoustic aspect (usually referred only to pitch) and the ones that, to some extent, include meaning, although in an unsystematic way, these are the least. Since 2006 we proposed the idea of trying to develop tools to address the research of some aspects of language as a subsystem of the human, the biological, the social and the physical.

5 The Grupo-Red de Complejidad y Lenguajes de La Habana was created by Raquel García Riverón

in 2006 as part of the groups that separated from the Cátedra para el Estudio del Pensamiento Complejo (Department for the Study of Complex Thought) of our country. It brought together specialists from all branches of linguistics and other sciences. This perspective, although it finds in Complexity the appropriate framework for its optimal development, it goes back to the joints developed by Leandro Caballero and Raquel Garcia Riverón from semantics and pragmatics, and from Intonology. We must not ignore the key role played by the Russian linguistics of dialectical epistemological bases as the background of this linguistic thought in Cuba. Raquel Garcia Riverón presented the initial vision of this work at the International Biennial Seminar on philosophical epistemological and methodological implications of complexity theory, at the Palacio de las Convenciones in Havana. This Seminar was organized by the Institute of Philosophy of the Academy of Sciences of Cuba (January 2006).

212

R. García Riverón and A. F. Marrero Montero

12.2.1 Previous Results of the Grupo-Red In opposition to the reductionist positions, we achieved the previous definition of the system of intonation of Cuban Spanish (García Riverón 1996, 1998). It was clear that in the intonological analysis, language and speech are a result of cultural and psychosocial interactional patterns of the human being. The system becomes discourse in the communicative activity of people, and cannot be abstracted from this. The model obtained advocates the study of intonational suprasegments from the approach that the system of language is dynamic, complex and non-linear, and for this reason, it must be described by using methods of qualitative analysis (delimitation of its functionality or communicative values) that enable the segmentation of intonation units that emerge in the different discourses of a type of highly complex system: the systems of natural languages. This segmentation takes into account its distinctive and significant capacity (García Riverón 2005a) in the system and in the interaction with other linguistic and non-linguistic expression means involved in orality. The results were obtained using qualitative research techniques that focus on the semanticpragmatic analysis of meaning (or, in this case, sense) of the oral text. Once the prosodic continuum is segmented using the qualitative methods, the quantitative methods are applied: acoustic analysis by measuring the variables of pitch, time and intensity. By trying to find answers to some questions, a survey of qualitative classification of logatomes was projected. It was concluded that the meaning of intonation has a relative autonomy from the rest of the (lexical-grammatical) means of expression: the meaning of intonation is socio-culturally encoded.6 However, even if this is so, in some cases, the sense emerges from semiosis processes in different systemic levels. For these reasons, it was presented the hypothesis of the need to examine a unit seen as a sign that we have called interactional speech act (acto de habla interaccional, AHI, in Spanish),7 which was defined as8 “a complex sign (with all the connotations that the concept of sign implies) comprising the interaction of the expression means (systemic, lexical, grammatical, prosodic and kinesic levels in oral discourse). Further on we opened the concept (García Riverón 2005b) by including “different semiotic codes”.

6 García

Riverón (2002, 2003). is very encouraging that the proposal of a cognitive and comprehensive approach we propose of the communication unit at a first level, a unit that certainly has its roots in the Vygotskian psychology, also arises in other latitudes in which other trends in psychology have proliferated. See McNeill (1999) for a similar proposal to develop the “growth points” for the study of gestures. 8 García Riverón (1998: 77). 7 It

12 A Complex Approach to Prosodic Discourse Variation

12.2.1.1

213

The First Semantic-Pragmatic Attractors

This approach has permitted to see more clearly the joint of the type of interactional speech act (AHI) with other semantic-pragmatic units that allowed to discretize the senses that emerged in the interactions of the different intonation patterns in discourses. This first level of analysis showed phenomena that are organized around five attractors that determine a large proportion of intonation system data and converge with the features that are obtained from an extensive fact-discourse material9 : 1. 2. 3. 4. 5.

The enunciative The interrogative The evaluative The exhortative The appellative

Let us see in detail the evaluative AHI, in which the sender expresses the receiver an evaluation (in an evaluative discourse) of an object (referent) that is intentionally considered (communicative and illocutionary strategy) and located on a scale of grammatical and semantic values linguistically structured (semantic modality evaluation,10 strongly related to the subjectivity of the individual and the categories of knowledge, motivation and desire). It has been called the semantic dominant semantics of evaluation. We have a clear example in the speech act (6)11 : 2 (3) -¿Dónde pongo los caramelos? (4). Dame un 1 7 cartucho. (Le dan el cartucho) (5) ¡Mami, (6)¡Qué 6a cartucho¡ (with a movement of the hand up and down) = negative evaluation 2 [ (3) –Where do I put the candies?(4). Give me a 1 7 paper bag. (He is given the roll of paper) (5) Mom, (6) What kind of paper bag 6a is this! (with a movement of the hand up and down) = negative evaluation]

9 For a full understanding

of the joints see the cited literature, especially: Pardo and Losada (2004); García Riverón et al. (2008). 10 Caballero (1995, 1996). 11 The number above the statement, signals in each example the intonation pattern according to the system described for the Spanish variant of Cuba (Appendix 12.1). The number in parentheses refers to the order of the statements in the sample. Indication of the gestures is placed in brackets in the same line of the statement.

214

R. García Riverón and A. F. Marrero Montero

It is important to confirm how these semantic-pragmatic dominants of the system of intonation in oral discourse in samples of Cuban Spanish,12 articulate with the macro-categories that are grounded on the dimensional discourse analysis. This articulation is presented today as a hypothesis and we aim, with its study, to explain, describe and define the bases for a new concept of underlying structure of intonation with a semantic-pragmatic basis, unlike other schools in the current intonology,13 which do not take into account from the beginnings of research the vital ontological trait of language: the meaning.

12.2.1.2

Discourses: Emergence of Expression Means and Senses

In the semantic-pragmatic analysis of prosody and discourse in general, the macrocategories or attractors found by our researchers in discourses, marking the discursive strategy or intentionality, are defined as: evaluation, loyalty (manipulation), determination, interest, appreciation and expressiveness. Also, the more specific features related to the noetic-semiotic processes of agents or speakers are tracked with a higher degree of precision, by establishing the relation thought-complex systemlanguage-discourse, this is, the relation sign-sense. There are attempts to describe these correlations in the illocutionary dimension (intentionality), in the modal dimension as modal features (elaboration of the targeted object made by the agent) and in the referential dimension (elaboration of the referents made by the agent) of discourses.

Results: Approach to the Analysis of Discursive Praxis Let us see as an example, one of the types of discourse studied: a conversational text of an audiovisual discourse, the Cuban film Fresa y Chocolate (Strawberry and Chocolate), by Tomas Gutiérrez Alea,14 in which only some of the expression’s means of interactions are present on the joints. This will allow us to observe how the various means of expression (lexical-grammatical structure, prosody, gestures and actions) interact, based on the communicative intention (analysis of the behavior of the semantic, pragmatic, and illocutionary features) or the discursive strategy and the emerging senses of the interactions. SCENE The scene takes place in the second half of the eighties in Coppelia, a busy ice cream parlor in Havana, it is daylight. Diego, a homosexual intellectual, arrives and sits at 12 These

dominants were defined in 1998 (García Riverón 1998). Later, researchers of the GrupoRed have continued working to expand or improve the semantic-pragmatic discursive categories found analyzed in the different discourses. 13 See García Riverón (2005a). 14 We reproduce the full text of a scene of the film, which was fully analyzed, although for reasons of length and to avoid unnecessary tautologies, only the discursive analysis of a few sentences in which these processes are clearly manifested is submitted for the consideration of the reader.

12 A Complex Approach to Prosodic Discourse Variation

215

the table of David, a heterosexual university student member of the Union of Young Communists (Unión de Jóvenes Comunistas, UJC, in Spanish).15 Diego: Con permiso.29 No pude resistir la tentación…30 Me encanta la fresa.31 Ummm (saborea el helado) Es lo único bueno que hacen en este país. 32 Ahorita lo exportan. 33 Y para nosotros agua con azúcar.34 ¡Uy! Hoy es mi día se suerte; 35 (encuentra una fresa en el helado) Encuentro maravillas. 36 (mira directamente a David) Bueno, volemos en alas de la imaginación; 37 porque en otra cosa no se puede.38 ¿Où est le cahier du cinéma? 39 (Mientras busca en una bolsa coloca algunos libros sobre la mesa entre los que se encuentra Conversación en la catedral de Vargas Llosa. David se queda mirando el libro) ¿Te interesa Vargas LLosa? 40 Este está dedicado.41 Pero en casa tengo otro ejemplar.42 Además tengo a Severo Sarduy y a Goitizolos completos.43 ¿Vamos a buscarlos? 44 [Diego: May I? 29 I couldn’t resist the temptation… 30 I love strawberry! 31 Mmm. (He tastes the ice cream) it’s the one good thing made in Cuba. 32 Soon they’ll export it 33 and for us, water and sugar 34 Today is my lucky day! 35 (He finds a strawberry in the ice cream) I find wonderful things. 36 (He stares at David) let’s fly on the wings of imagination. 37 It’s the only way we can. 38 ¿Où est le cahier du cinéma?16 39 (While he is looking in a purse, he puts some books on the table, among which Vargas Llosa’s Conversación en la Catedral 17 can be found. David stares at the book) Interested in Vargas Llosa? 40 This one is autographed 41 but I have another copy at home. 42 I’ve also got the works of Severo Sarduy and Goitizolos. 43 Shall we go get them? 44] David: Yo no voy a casa de … (lo mira de arriba hacia abajo) gente que no conozco.45 [David: I don’t visit… (he looks him up and down) strangers 45] Diego: Aprovecha, niño.46 ¿Dónde vas a encontrar esos libros? 47 (David se cambia el carné de la Unión de Jóvenes Comunistas, UJC, de bolsillo) Ummm Capté.48 Solo puedes leer los libros que te autorizan en la juventud.49 Los forras, viejo. 50 Ten imaginación.51 [Diego: Don’t let this chance go, honey. 46 Where else would you find these books? 47 (David takes the UJC card out of one of his shirt’s pockets and puts it in another) Mmm, got it! 48 You only read books authorized by the Union of Young Communists 49 Cover them, buddy … 50 Be imaginative! 51] David: No tengo que forrar nada.52 Yo leo lo que me da la gana.53 Y no tengo ganas de hablar.54 ¿Está bien?55 [David: I don’t have to cover anything. 52 I’ read what I please … 53 and I don’t feel like talking. 54 Is it clear? 55] Diego: Yo a ti te conozco.56 Te he visto muchísimas veces saliendo de la universidad.57 [Diego: I know you. 56 I’ve seen you many times leaving the University. 57] David: No soy yo.58 15 The

numbering in the text corresponds to the general numbering of the statements in the film. French in the original: Where is the movies magazine? 17 Conversation in the Cathedral. 16 In

216

R. García Riverón and A. F. Marrero Montero

[David: It wasn’t me. 58] Diego: Sí, niño, como no vas a ser tú.59 [Diego: Of course it was you! 59] David: No soy yo.60 [David: It wasn’t! 60] Diego: Ay, perdona.61Compañero Torvaldo.62 Sólo quería prestarte unos libros y regalarte unas fotos de cuando actuaste en Casa de muñecas. 63 [Diego: Excuse me, then! 61 Comrade Torvald. 62 I just wanted to lend you some books … and give you some photos of you from “A Doll’s House”. 63] David: ¿Fotos mías? 64 [David: Pictures of me? 64] Diego: Nora y Torvaldo…65 estuviste maravilloso.66 [Diego: Nora and Torvald … 65 you were wonderful. 66] David: ¿Y esas fotos? 67 [David: And those pictures? 67] Diego: Fotos de la obra.68 Soy fotógrafo. 69 Tengo muchísimas.70 Todo el que las ve se queda loco, 71 ¡pero loco! 72 [Diego: From the play. 68 I’m a photographer. 69 I’ve got many.70 everyone who sees those goes crazy. 71 Crazy! 72] David: Me las tienes que dar.73 [David: So give them to me! 73] Diego: Por supuesto son tuyas.74 [Diego: If you want them, they’re yours. 74] David: ¿Y dónde están? 75 [David: Where are they? 75] Diego: En mi casa.76 Yo vivo aquí cerquita.77 Con mis padres, mis sobrinos y unas tías además que no salen nunca.78 [Diego: In my apartment. 76 I live nearby. 77 With my parents, nephews and to boot aunts, who never go out. 78] David: Está bien, 79 vamos a buscarlas.80 Y me las das todas, ¿okey? 81 [David: Okay 79 let’s go get them. 80 but I want them all! 81] Diego: Déjame terminar el helado, no.82 Ya.83 [Diego: Let me finish my ice cream. 82 Ready! 83]. It is clear that speakers belong to different social groups. Diego is a man of a wide artistic and literary culture, professing religious beliefs and an ideological vision18 that is contradictory to the ideology of the person he speaks to. David has a peasant origin, a university student in political science, but with a very restricted cultural spectrum and militant of the UJC, which advocated atheism at the time portrayed in the scene. This forms a pattern of social interaction that determines their motivations, desires, knowledge and intentions, and power relations, which vary throughout the entire play (in the macrotext) and even, in more modest margins, within the fragment analyzed. 18 We

understand ideology as Van Dijk (2003), as a set of beliefs shared by a social group.

12 A Complex Approach to Prosodic Discourse Variation

217

The macrostructure of the discourse shows that the underlying intention of the speaker Diego is to take the other person, David, to his home to, in the medium term, win the bet made with another character of taking David to bed, which is not revealed in the film but at the end. In this way, a discursive strategy is built throughout the fragment and is marked by constant units of sense: evaluation, loyalty, determination, interest, appreciation and expressiveness. The strategies composed by different illocutionary, modal and referential features behave as shown in the Table 12.1. In contrast, David does not have a defined strategy a priori, its discursive strategy is built as a reaction to Diego’s strategy in a defensive position. Units of evaluation, interest, determination and expressiveness predominate in him. The analysis of the illocutionary, modal and referential dimensions are shown in the Table 12.2. Now, we proceed to the analysis of the realization of the strategy in examples of speech acts within the fragment, in the analysis of each example we will always take into account the general discursive strategy described: EXAMPLE 1 During the development of the defined strategy, different interactions of expression means are used, from which the different senses emerge. Diego’s AHI 30—No pude resistir la tentación interacts with E-1. It is part of the macro-category, as an indirect explanatory act, for the “tentación” is not supported by strawberry ice cream, but by the interlocutor. The speaker continues his strategy with AHI 31 Me encanta la fresa, which interacts with E-1 with prosodic emphasis on (-can-) (Fig. 12.1.) and a whisper of the following syllables characterized by a low sound that reinforces the subtext (Fig. 12.2). It is an evaluative, categorical AHI that can move to an explanation of cause: he entered the ice cream parlor because he likes David. We found the modal

Table 12.1 Table of character David’s different illocutionary, modal and referential features

218

R. García Riverón and A. F. Marrero Montero

Table 12.2 Character Diego’s different illocutionary, modal and referential features

semantic-pragmatic features good-approving manners and the referential features of knowledge: reality and quality.

Fig. 12.1 Pitch in semitones ahi me encanta la fresa

Fig. 12.2 Time in seconds ahi me encanta la fresa

12 A Complex Approach to Prosodic Discourse Variation

219

It is also an indirect act. Insinuation, the subtext, is reinforced with a stare at the partner. We cannot forget that the ramifications of this dialogue contain traces of various subtopics and subsequent sub-plots that then will run through the whole film. From this explanation of initial state, the adventures of the characters begin, through a process shaken by constant bifurcations leading to the transformation of the (deeply stereotyped) initial state in another state of complex and contradictory nuances that goes beyond the narrow mold of the dichotomy homosexual/communist. EXAMPLE 2 Diego: 32 Ummm (Mientras saborea el helado) Es lo único bueno que hacen en este país. 33 Ahorita lo exportan. 34 Y para nosotros agua con azúcar. Here begins the ideological characterization of the speaker. In AHI 32 we have an evaluation marked by interjection Ummm and the tasting of ice cream, followed by a disapproving evaluation of society that, along with AHI 33 and 34, make up a critic of the prevailing social system. The modal semantic-pragmatic features of AHI 32 are: bad-disapproving, categorical, and in the referential dimension, it moves within the knowledge of reality. The following two acts are informative presenting similar features with a causal coherence relationship. The tonal scheme by levels of the speech act Ahorita lo exportan, is pronounced above the average tonal level, with a resulting higher pitch that reinforces the ideological content that we have mentioned (Fig. 12.3). EXAMPLE 3 In speech act 37 Bueno/volemos en alas de la imaginación, we have a connector on a high rising tone that allows Diego to change the subject of conversation and continue with the selected discursive strategy (Fig. 12.4). This change of subject introduces explicitly the arguments of the fundamental (ideological) contradiction between the two characters as a mechanism for local coherence of explicitness of this contradiction that rules the entire story, while being constituent of the macrostructure of the film. In AHI 38 Porque en otra cosa no se puede, that interacts with VE-1a, we have

Fig. 12.3 Tonal pattern by levels of ahi ahorita lo exportan (normal tonal levels for this unit are in between the gray lines)

220

R. García Riverón and A. F. Marrero Montero

Fig. 12.4 Spectrogram and behavior of the fundamental frequency (blue line) of ahi bueno/volemos en alas de la imaginación, porque en otra cosa no se puede (all spectrograms were generated with Praat)

an intensification of an expressive and evaluative act that reinforces the ideological position of the speaker of criticism of the prevailing system with the argument: “You cannot leave the country”. To do this, let us mainly focus on the modalization of the speech act: evaluation with bad-disapproving features, categoricity and certainty, interacting with the referential feature of knowledge of reality. EXAMPLE 4 Diego: 44 ¿Vamos a buscarlos? David: 45 Yo no voy a casa de … gente que no conozco.

Fig. 12.5 Spectrogram and pitch behavior of ahi yo no voy a casa de… gente que no conozco in interaction with pause and gesture

12 A Complex Approach to Prosodic Discourse Variation

221

In this example that interacts with exhortatory E-3, we have a polite invitation with which Diego seeks to achieve its objective and to which David responds with an explanatory act whose modal semantic-pragmatic features are uncertainty-certainty, disjunction. In this case, there is also an evaluation with a bad-disapproving feature that is expressed by interaction with David’s evaluating look up and down to Diego as an update process of the object focused by the speaker. In the prosodic dimension we have that the look coincides with a pause after which the euphemism gente que no conozco, meaning homosexual is introduced (Fig. 12.5) Likewise, we can observe power relations: although at first, they show balance, at a certain moment of discourse, it is lost given the relationship with Diego’s higher knowledge of facts that he uses in his discourse with the intention of taking David home. David’s desire to recover the photos makes Diego complete his intention for which he has built a manipulative discourse strategy with declarative, appellative, evaluative and interrogative speech acts, in complex interactions of different means of expression.

12.3 Conclusions and Prospects Starting from prosody, we have analyzed its emerging roles in “orchestrating the intentional subject of the author”. The work with a film sample has allowed us to deepen in the study of the subsystem of characters and establish hypothesis about its evolution in the narrative and the ideological characterization, by analyzing intentionality and strategies expressed by illocutionary, modal and referential features. The study of the characters’ subsystem, as seen in the examples, allows to understand their characterization individually, and their relationship as elements of this narrative subsystem. Pavis (1988) has referred to this as the semantic and semiotic characterization of the characters. From there, we can build maps of intentionality and strategies of the characters’ subsystem whose outline would reflect interactions with other narrative categories such as conflict and the development of action in the narrative continuum. All this would be, at the same time, an image of narrative as emerging from the system of interactions among its categories. Currently, only the film Fresa y Chocolate has been analyzed, but as part of an ongoing project, another sample is being studied. A recording of the reading of the short story El lobo, el bosque y el hombre nuevo has been made, in the voice and the intentionality of its author: Senel Paz. This will test the effectiveness or not of this analysis in the study of intentionality and strategies of the narrator’s subsystem. Then, we should be able to describe the narrator and the characters as identities within the story, by their relation and contrast. Consistently with our complex approach, current research is conducted by a team of specialists in literature and linguistics. At the same time that the synchronic states of characters in the story could be described, it is also possible, by analyzing the narrative synchrony, to reach a diachronic description of characters and the evolution of the intentionality and strategies throughout the narrative. In this area, superstructural categories come into play.

222

R. García Riverón and A. F. Marrero Montero

That is, characters in the development of the story in relation to the characters in the end. The application of the matrix of illocutionary, modal and referential features to narrative studies from prosody provides an operationalized model of a more scientific objectivity. Moreover, the study of prosodic indicators in relation to the character’s ideology opens this vast field of ideological mutations of characters and their negotiation processes according to the development of narrative conflict. All this, connected with referentiality to the context. Fresa y Chocolate is a perfect example because it was the film that dealt with, for the first time, the taboo subject of homosexuality in relation to the scale of values of the Cuban Revolution in a story about the friendship between a homosexual and a communist militant. This component also determined its success. The story of these characters and the projected conciliation also involve an ideological proposal. Finally, the analysis allows to describe multiple rhetorical and stylistic devices associated with the management of focus, which are suitable to the film and literary art. On the whole, we have exposed some theoretical tools and a multidimensional model, those enable and refer to: • Explain and demonstrate the intentionality of speakers in the different types of discourses that are being studied with semantic and pragmatic features established in elaborated matrices from data belonging to the corresponding discourses. • Describe the relationship of discourse with actants in different situations of social interaction. • Describe the features of the subjectivity of the participants in discourses. • Describe the referential features that may influence the analyzed discourses. • Describe the different expression means of communication strategies of speakers in the different discourses. All this, with a holistic view that tries to cover the diverse discursive levels studied or, in this case, leave the track established for future research.

12 A Complex Approach to Prosodic Discourse Variation

223

12.1 Appendix Intonation System of the Spanish Variant of Cuba (García Riverón 1996).19

19 The

(RE).

system is made of Entonemes (E), Entonemes Variants (VE) and Entonemes Realizations

224

R. García Riverón and A. F. Marrero Montero

References Austin, J. (1962). Cómo hacer cosas con palabras. Barcelona: Paidós. Caballero, L. (1995). Semiótica y diccionario [Unpublushed book]. Caballero, L. (1996). Ilocuciones valorativas. In G. Wotjak (Ed.), El verbo español. Aspectos morfosintácticos, sociolingüísticos y léxicogenéticos (pp. 93–109). Madrid: Vervuert-Iberoamericana. Calsamiglia, H., & Tuson, A. (2004). Las cosas del decir. Manual de análisis del discurso. Barcelona: Ariel Lingüística. Capra, F. (2002). The hidden connections. New York: Doubleday.

12 A Complex Approach to Prosodic Discourse Variation

225

García Riverón, R. (1996). Aspectos de la entonación hispánica. I La metodología y II Análisis acústico de muestras del español de Cuba. Cáceres: Universidad de Extremadura. García Riverón, R. (1998). Aspectos de la entonación hispánica. III Funciones de la entonación en el español de Cuba. Cáceres: Universidad de Extremadura. García Riverón, R. (2002). El significado de la entonación: primer acercamiento a los datos. Oralia, 5, 53–74. García Riverón, R. (2003). El significado de la entonación. In Z. E. Herrera & M. Butragueño (Eds.) La tonía: dimensiones fonéticas y fonológicas (pp. 245–265). México: El Colegio de México. García Riverón, R. (2005a). El estudio de la entonación. Moenia, 11, 141–176. García Riverón, R. (2005b). Nota a la reseña de Francisco Torreira sobre R. García Riverón “El significado de la entonación”. Estudio de Fonética Experimental, XIV, 364–368. García Riverón, R., Losada, M., & Pardo, A. (2008). El acto de habla interaccional como unidad para el estudio de la oralidad: una visión desde la complejidad, Oralia, 11, 333–351. Goodwing, B. (1998). Las manchas del leopardo (A. García, Trans.). Metatemas 51. Barcelona: Tusquets Editores. Juarrero, A. (1999). Dynamics in action: Intentional behavior as a complex system. Cambridge: The MIT Press. Kauffman, S. (2003). Investigaciones. Complejidad, autoorganización y nuevas leyes para una biología general (L.E. de Juan, Trans.). Metatemas 76. Barcelona: Tusquets Editores. Kennealy, C. (2010). Talking heads. New Scientist, 20, 33–35. Leontiev, L. A. (1962). Jazyk, rech´i rechevaja deatel´nost. Moscú: Ed. Nauka. McNeill, A. (1999). Triangulating the growth point. Arriving at consciousness. In L. S. Messing & R. Campbell (Eds.), Gesture and speech and sign. Oxford: Oxford University Press. Morales Lopez, E. (2011). Hacia dónde va el análisis del discurso. Tonos Digital. Retrieved from www.um.es/tonosdigital. Morin, E. (1994). El Método III. Madrid: Cátedra. Pardo, A., & Losada, M. (2004). Un acercamiento semántico-cognitivo a la tropología: la metáfora y la metonimia. Moenia, 10, 143–158. Pavis, P. (1988). Diccionario del teatro. Dramaturgia, estética, semiología. Vols I & II. La Habana: Edición Revolucionaria. Prigogine, I. (1983). La nueva alianza. Metamorfosis de la ciencia. Madrid: Alianza. Searle, J. (1969). Actos de habla. Madrid: Cátedra. Van Dijk, T. A. (1983). Texto y contexto. Un enfoque interdisciplinario. Barcelona: Paidós. Van Dijk, T. A. (1997). Discourse as social interaction. London: Sage. Van Dijk, T. A. (2003). Ideología y discurso. Una introducción Multidisciplinaria. Barcelona: Ariel. Vigotskii, L. S. (1934/1966). Pensamiento y lenguaje. La Habana: Edición Revolucionaria. Wittgenstein, L. (1973). Philosophical Investigations (G.E.M. Anscombe, Trans.). (3rd ed.) New York: Prentice Hall.

Chapter 13

Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic Performances H. Samy Alim and Geneva Smitherman

Abstract A large number of American researchers have recently been dedicated to theorize about language and races as a unified social process rather than as two independent disciplines. Within this “new” field of study called raciolinguistics, in which it is intended to apply the different linguistic methods to analyze and solve the most relevant problems on the relations between language, race and power, one can observe how language shapes ethnoracial identities as well as the role it plays in racialization and its importance in the concepts of race and racism, in general. The scope of this paper is to also theorize about language and race but with special attention to how both processes nurture and constitute each other. Through President Barack Obama’s discourse, it is intended to demonstrate what speaking as a racialized subject and as one of the most important social figures in today’s America implies. Likewise, through what is said or what media mentions about his way of speaking, one is able to distinguish why this interdiscipline proposes that race is sociolinguistically constructed and not only socially.

13.1 Introduction You go to the cafeteria… and the black kids are sitting here, white kids are sitting there, and you’ve got to make some choices. For me, basically I could run with anybody. Luckily for me, largely because of growing up in Hawai’i, there wasn’t that sense of sharp divisions. Now, by the time I was negotiating environments where there were those kinds of sharp divisions, I was already confident enough to make my own decisions. It became a matter of

H. S. Alim (B) Department of Anthropology, University of California, Los Angeles, CA 90095, USA e-mail: [email protected] G. Smitherman Michigan State University, East Lansing, MI 48824, USA e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_13

227

228

H. S. Alim and G. Smitherman

being able to speak different dialects. That’s not unique to me. Any black person in America who’s successful has to be able to speak several different forms of the same language… It’s not unlike a person shifting between Spanish and English.1 —Barack Obama

In June of 2015, Americans were stunned by a violent White terrorist attack on an African American church in Charleston, South Carolina, in which a young White man walked into a Black church and murdered the reverend and eight other members of the church. At the eulogy for slain Reverend Clementa Pinckney, Barack Obama spoke lovingly and powerfully to all those in attendance, explaining that “more than any particular policy or analysis,” what was needed in that moment was “an open heart, a reservoir of goodness.” Then, in what was a wholly unexpected move, and perhaps a singular moment in American political oratory, President Obama began singing the emblematic African American spiritual, “Amazing Grace.” The predominantly African American audience responded immediately with uproarious applause, joining him, singing line-by-line, with hands raised in the air, bearing witness to the amazing grace of God in moments of horrific and devastating tragedy. Even the organist began playing musical accompaniment, until the entire church was singing with the President. After having roused the spiritual sentiment and soul of the audience, Barack Obama reached a rhetorical climax of the eulogy: Obama: Clementa Pinckney, found that grace! Audience: Yes! Obama: Cynthia Hurd, found that grace! Audience: Yes! [organ punctuates response] Obama: Susie Jackson, found that grace! Audience: Yes! [organ punctuates response] Obama: Ethel Lance, found that grace! Audience: Yes! [organ punctuates response] Obama: DePayne Middleton-Doctor, found that grace! Audience: Yes! [organ punctuates response] Obama: Tywanza Sanders, found that grace! Audience: Yes! [organ punctuates response] Obama: Daniel L. Simmons, Sr., found that grace! Audience: Yes! [organ punctuates response] Obama: Sharonda Coleman Singleton, found that grace! Audience: Yes! [organ punctuates each response] Obama: Myra Thompson, found that grace! Audience: Yes! [organ punctuates each response] Obama: …through the example of their lives! Audience: Yes! My my! 1 One

of Barack Obama’s most insightful interviewers was none other than “Sir Charles”—NBA legend Charles Barkley, that is. His book, Who’s Afraid of a Large Black Man? (New York: Penguin Press, 2006) features interviews on race in America with Barack Obama, Bill Clinton, Jesse Jackson, Tiger Woods, Morgan Freeman, George Lopez and Ice Cube, among others. Obama’s quote is from p. 25.

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

229

Obama: They’ve now passed it on to us. Audience: Uhuh! Obama: May we find ourselves worthy… Audience: Uhuh! Yes! Obama: …of that precious and extraordinary gift… Audience: My my! Obama: …as long as our lives endure. Audience: Mm-hmm! Yes, sir! Obama: May grace now lead them home. Audience: Well! Yes! Obama: May God continue to shed his grace… Audience: Uhuh! Yes! Obama: …on the United States of America. Audience: Yes! [thunderous applause].2 By tapping into the African American oral tradition in what was a masterful calland-response performance, with perfectly timed use of repetition—two rhetorical devices that the audience clearly understood—President Obama was able to honor the lives lost by transforming himself from America’s “President-In-Chief” into America’s “Pastor-In-Chief.” As we show later in this chapter, President Obama’s use of what many refer to as a “Black preacher style” was critical to him becoming president of the United States. It is raciolinguistic performances such as these that enabled his rise in American politics, ultimately to the highest office in the land. These performances will be examined within the context of the rise of an emerging field of raciolinguistics in the United States.

13.2 Raciolinguistics In the U.S., a critical mass of scholars is now committed to theorizing language and race together rather than as separate social processes (Alim et al. 2016). Our goal is to theorize language and race together, paying particular attention to how both social processes mediate and mutually constitute each other. Collectively, we have forged a new field of raciolinguistics that is dedicated to bringing to bear the diverse methods of linguistic analysis to raise and answer critical questions about the relations between language, race, and power across diverse ethnoracial contexts and societies. To be sure, linguistic anthropologists have produced substantive research examining race and language (Urciuoli 1996; Zentella 1997; Spears 1999; Makoni et al. 2003) and continue to do so (Reyes 2007; Mendoza-Denton 2008; Reyes and Lo 2009; Alim and Reyes 2011; Bucholtz 2011; Dick and Wirtz 2011; Ibrahim 2014; Roth-Gordon 2016; Rosa 2017), but it is only recently that there has been a focused, 2 The

eulogy can be accessed here: https://www.youtube.com/watch?v=IN05jVNBs64.

230

H. S. Alim and G. Smitherman

collective effort to theorize race and ethnicity within and across language studies. This emerging focus addresses both the fields of linguistics and race and ethnic studies by foregrounding the role of language in shaping ethnoracial identities. Raciolinguistics is interested in complicating the increasingly vexed relations between race, ethnicity, and language. As raciolinguistics—the interdisciplinary field of “language and race”—continues to have an impact across anthropology, linguistics, education and other fields (Rosa, Flores, Lo, AERA panel), these scholars aim to focus these fields on both the central role that language plays in racialization and on the enduring relevance of race and racism in our lives. A growing number of language scholars hold that rather than fixed and predetermined, racial and ethnic identities are (re)created through continuous and repeated language use. Further, while race and ethnic studies research has focused critically on race and ethnicity as social constructs, this emerging field demonstrates that race is not only socially constructed, but sociolinguistically constructed. In addition to understanding how the social processes of language and race mediate and mutually constitute each other, raciolinguistics specifically seeks to: (1) Employ the diverse methods of linguistics to raise critical questions about the relations between language, race, and power; (2) Highlight research that explores how ethnoracial identities are styled, performed and constructed through minute features of language (variations in phonological and morphosyntactic features, for example) as well as diverse modes of interaction, from embodied, face-to-face conversations to widely-circulating media discourses; (3) Integrate theoretical areas of style, stance, and performance (among others) in order to more closely examine how these approaches might inform each other and processes of racialization; (4) Take intersectional approaches that understand race as always produced in conjunction with gender, sexuality, class, citizenship, religion and other axes of social differentiation; (5) Look comparatively across diverse linguistic and ethnoracial contexts to better understand the role of language in maintaining and challenging racism as a global system of capitalist oppression; (6) Emphasize the linguistic and discursive construction of race and ethnicity while at the same time noting their endurance as social realities for subjugated racial and ethnic minorities, (im)migrants, and other oppressed groups; (7) Consider the complexities of racialization within the rapid demographic shifts and technological advances of the 21st century; (8) Consider the implications of research for developing various anti-racist strategies that can lead to social transformation in the areas of language, race, and education, among others.

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

231

13.3 Returning to Barack Obama and the Project of Racial Becoming In 2008, the U.S. elected its first “Black Language”-speaking president. We, along with many others, were fascinated not only by how Barack Obama used language, but also by how America responded to his language use. As scholars of language and race, it became clear that Barack Obama’s linguistic production (how he talks) and the metalinguistic commentary that surrounded it (how Americans talked about how Barack Obama talked) revealed the cultural tensions around language and race in the U.S. As Barack Obama took the national stage by storm, folks from across the racial and political spectrum began commenting on his language, from linguist John McWhorter’s playful use of “Blaccent,” to Hip Hop icon Snoop Dogg’s observation that Obama had “the right conversation” to independent candidate Ralph Nader’s quip, “he’s gonna do that White-talking thing again. “Later, Harry Reid’s racialized comments about Barack Obama—that he “speaks no Negro dialect, unless he wants to have one”—gave us all pause. What exactly did all of this commentary mean? Why were so many Americans so closely monitoring the first Black president’s language? As we wrote in Articulate While Black: Barack Obama, Language, and Race in the U.S. (2012), Barack Obama had to constantly navigate multiple and intersecting discriminatory discourses of language, race, citizenship, and religion. While President Obama was the object of our analysis (what were the underlying reasons why so many referred to him articulate, the adjective?), the more important questions for us centered on the social, cultural, linguistic, and educational implications of what it means to articulate while black. For us, that was the deeper, underlying philosophical question of our work: What does it mean to speak as a racialized subject in contemporary America? This is one of the central concerns of the emerging field of raciolinguistics (see discussion below). In using President Obama as a starting point, we sought to race language and language race (Alim 2009)—that is, to view the politics of language through the lens of race, and the politics of race through the lens of language—in order to gain a better understanding of language and/in the process of racialization. The following discussion explores how Barack Obama translates himself as “Black” through his raciolinguistic performances. We take the view that language varieties are not just lists of features that belong to a given “race”; similarly, linguistic features do not just straightforwardly belong to a given “language.” Rather, we prefer to use the term “linguistic resources” in order to highlight the fluid and evolving nature of language. We view linguistic resources as being employed by speakers as they shape and engage in processes and projects of identification. President Barack Obama’s use of what comes to be racialized as “Black Language,” for example, is very much a conscious raciolinguistic project. In the same way that the President selected “Black” on the U.S. Census to mark his racial identity, he also regularly employs particular linguistic resources in the multifaceted project of becoming Black (Ibrahim 2014), or more generally, of racial becoming.

232

H. S. Alim and G. Smitherman

Barack Obama’s chronicles his search for “a Black identity” in his best-selling book, Dreams from My Father: “Away from my mother, away from my grandmother, I was engaged in a fitful interior struggle. I was trying to raise myself to be a black man in America, and beyond the given of my appearance, no one around me seemed to know exactly what that meant.3 ” (76). Many Americans who are racialized as “Black,” especially those on the margins of what many view as a normative Black identity are very familiar with this process. Sudanese scholar Ibrahim (2003) described the process in these terms: “To become black is to become an ethnographer who translates and searches around in an effort to understand what it means to be black in North America.” It is a process of “entering already pronounced regimes of Blackness.” Black feminist Joan Morgan, who identifies as Jamaican, described the process of becoming Black in America in these terms: “As a matter of both acclimation and survival, we learn [African American] history. We absorb the culture. Some of us even acquire the accent.” (2009: 63). Racialization, then, is a process of socialization in and through language. It is a continuous project of “becoming” as opposed to “being.”

13.4 “Nah, We Straight”: Race, Language, and Sociolinguistic Style Turning our attention directly to Barack Obama, we now present some of our findings regarding the intersection of race, language, and style. Throughout our conversations with and surveys of Americans of the Obama generation (at the time, mostly 18–24, with a few in their early 30 s), it became clear that he was extremely highly regarded as a communicator and speaker. Barack Obama was sometimes even referred to as “the most powerful speaker of our age.” Our surveys and conversations further revealed that listeners believed Barack Obama to be a speaker who was “strategic” and “hyper-aware” of his audience. While this hyper awareness of your audience may be central to many politicians, what distinguished Obama was his successful deployment of various sociolinguistic styles. For politicians with diverse electorates and constituences, sociolinguistic style takes on a heightened level of importance. Obama was seen as someone who could speak directly and comfortably with folks across regions, generations, socioeconomic divisions, racial and ethnic groups, and political and religious views. Barack Obama’s global family history, diverse life experiences, and multicultural socialization, along with his biracial background, surely helped him sharpen his styleshifting skills.8 At the beginning of this chapter, we quoted Obama’s description

3 From

Barack Obama’s Dreams from My Father: A story of race and inheritance (New York: Crown Publishers, 1995, pp. 73–74).

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

233

of his experience as a young man of Color growing up in American schools as one where you had “to make some choices.” In the American context, where sharp racial divisions in friendship groups are still the order of the day, Barack had to learn to speak “several different forms of the same language.” In much the same way that many bilingual/bicultural Americans “codeswitch” between two languages (English and Spanish, for example), many bilingual/bicultural Americans “styleshift” between varieties of the same language (Puerto Rican English and White Mainstream English, for example).9 While Barack Obama’s ability to styleshift is one of his most compelling and remarkable linguistic abilities, it is also quite common for many Black Americans who travel in and out of Black and White social worlds. In fact, Black Americans in our conversations and surveys were more likely than Whites and others to note Obama’s styleshifting abilities. While Black respondents differed in terms of their view about the degree to which Barack Obama sounded “Black,” many made critical distinctions between his language and his style. One young African American woman from California, however, made observations about instances of Barack Obama using Black Language style as well as syntax. Noting both the range of Obama’s styleshifting and distinguishing between his language and style, she noted that he reserves “the nonstandard grammatical structure” of Black Language for “settings that are primarily Black.” She then offered an example of Obama’s language from his visit to Ben’s Chili Bowl, which she described as “a racially mixed, very informal location in the heart of D.C.” This now famous example was captured on YouTube.11 In the clip, Barack Obama is seen interacting with a Black cashier. When offered his change, he declined with the statement, “Nah, we straight.” While this may seem like a simple phrase, in these three words we have three different linguistic features that are aspects of Black Language: (1) “No” is rendered as “nah” through a phonological process known as the monophthongization of diphthongs. Varying between these different pronunciations is a linguistic hallmark of Barack Obama’s styleshifting. In a South Carolina speech with a racially-mixed audience, which we analyze below, Barack says “Wit mah Bahble” for “With my Bible.” In this case, the diphthong in “my” and “Bible” was rendered as “ah.” (2) Despite some hilarious misinterpretations of the word “straight” (no, he was not referring to anyone’s sexual orientation!), Barack Obama used the word in the Black, now-crossover youthful sense of the word, meaning he was “okay,” “fine,” “alright” without getting his change back. Many observers have noted Barack Obama’s use of Black slang in relation to Hip Hop Culture, using such words as “flow” or “tight.”12 Other than its ever-evolving slang, the lexicon of Black Language is not as widely known outside the Black community. Barack Obama also uses words and phrases from this less widely known dimension of

234

H. S. Alim and G. Smitherman

the Black Lexicon, which have survived for generations in the Black community, like, “trifling,” “high-yella,” “Tom/Uncle Tom,” and “house nigger.4 ” (3) In addition to words, phrases and pronunciation, the third Black linguistic feature in Obama’s “Nah, we straight” is the use of zero copula. The “copula” refers to “is” and “are” and other forms of the verb “to be.” This feature is one of the most important and oft-studied features of Black Language. In fact, sociolinguists sometimes describe the copula as Black Language’s “showcase variable” because it gives Black Language its distinctiveness, setting it apart from other varieties of American English.5 Non-Black Americans sometimes deride this use of zero copula as a sign of “Black people’s ignorant ways” or “their lazy, ungrammatical speech.” Contrary to popular opinion, Black Language actually has a more complex verbal system than any other White American variety of English. This is due mostly to its origins as a Creolized form of African and

4 Barack Obama uses all of these lexical items in Dreams from My Father. In Geneva Smitherman’s

Black Talk: Words and phrases from the hood to the amen corner (Boston & New York: Houghton Mifflin Company, 1994, 2000), trifling “describes a person who fails to do something that he/she is capable of doing; irresponsible; inadequate” (p. 285). Barack uses trifling in exactly this sense on page 226 in Dreams: “We’re trifling. That’s what we are. Trifling. Here we are, with a chance to show the mayor that we’re real players in the city, a group he needs to take seriously. So what do we do? We act like a bunch of starstruck children, that’s what…” Smitherman defines yella/high yella as a term used to describe “a very light-complexioned African American; praised in some quarters, damned in others. Community ambivalence stems from high yellas’ close physical approximation to European Americans…” (p. 303). In Dreams, Barack writes about becoming “familiar with the lexicon on color consciousness” (p. 193) in the Black community and uses the term high-yella on page 273 in Dreams: “…the high-yella congregations that sat stiff as cadets as they sang from their stern hymnals…” Tom/Uncle Tom is described by Smitherman as “a negative label for a Black person, suggesting that he/she is a sell-out, not down with the Black cause. Tom comes from the character Uncle Tom in Harriet Beecher Stowe’s nineteenth century novel, Uncle Tom’s Cabin, who put his masters wishes and life before his own…” (p. 284). Barack uses these terms to describe his puerile attempt to belittle another Black classmate in college: “Tim was not a conscious brother. Tim wore argyle sweaters and pressed jeans and talked like Beaver Cleaver… His white girlfriend was probably waiting for him up in his room, listening to country music… ‘Tim’s a trip, ain’t he,’ I said, shaking my head. ‘Should change his name from Tim to Tom.’” (p. 101–2). House nigger, Smitherman explains, historically referred to “an enslaved African who worked in Ole Massa’s house,” rather than in the field (field nigga), and “was viewed as loyal to Massa.” (p. 130). Malcolm X updated this term in the 1960s to refer to the working-class Blacks as field niggas and middleclass Blacks as house niggas. House niggas were “more likely to deny the existence of racism or make excuses for it, to identify with whites and the system, and thus unlikely to engage in protest of rebellion.” This is precisely how Barack Obama used the term when he realized that his Muslim grandfather, whom he always imagined to be “an independent man, a man of his people, opposed to white rule” in Kenya, turned out not to be anything but that. “What Granny had told us scrambled that image completely, causing ugly words to flash across my mind. Unlce Tom. Collaborator. House nigger.” (p. 406). 5 Rickford, J., Ball, A., & Blake, R. (1991). Rappin on the copula coffin: Theoretical and methodological issues in the analysis of copula variation in African American vernacular English. Language Variation and Change, 3, 103–132.

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

235

European language varieties.6 Now, before explaining further, we want to share five important points about the copula. The first point is that while speakers of White American varieties of English only have two ways of representing the copula, speakers of Black Language have three. In Black Language, you can say all three of these, depending on the situation: (1) We are straight, (2) We’re straight, or (3) We Ø straight. In White varieties of English, you are restricted to the first two forms. Black Americans can shift between these three variants, all of which have the same literal meaning but differ in social meaning. The second point is that copula absence follows a very well-documented set of linguistic constraints. That means that speakers cannot just decide to always use the zero copula form (as in “We Ø straight”). Take this example from a Black minister in San Francisco: “The Black Man Ø on the rise, and the White man, he Ø runnin scared now, because we Ø wide awake today and he know we Ø not just gon lay down and accept things as they are.” While the copula can be absent before prepositional phrases and locatives (The Black Man Ø on the rise), progressive verbs (he Ø runnin scared), adjectives (we Ø wide awake), negatives, and the future marker gon (we Ø not just gon lay down), it cannot be absent when it is in sentence-final position (as they are). The copula cannot be absent in the first-person singular form either. For example, if Obama had said something like, “Nah, I straight,” that would be ungrammatical in the Black Language system. The third point is that these linguistic constraints on copula use are also ordered such that the copula is more likely to be absent in decreasing order before gon (She Ø gon do it), Verb + ing (She Ø doin it), locatives (She Ø at the bus stop), adjectives (She Ø happy), and noun phrases (She Ø the boss). Fourth, copula absence also depends on phonological [pronunciation] constraints, e.g., if there’s a vowel or a consonant before or after its use. Lastly, Black Americans also shift their use of the copula in regularly patterned ways depending on the race, gender, and cultural knowledge of the person to whom they are speaking (Alim 2004). For example, on the one hand, Black youth display high levels of zero copula in their peer groups and when talking with Black male Hip Hop fans. On the other, they are much more likely to use the 6 The

copula is just one example of BL’s complex verbal system and the Africanization of American English. In fact, according to John Rickford, copula absence “provides one of the strongest arguments for possible Creole and African influences on the grammar” of Black Language. Many Caribbrean Creoles and West African languages do not have the copula in some grammatical environments and patterns of its absence in Black Language mirror that of its absence in creoles (See Alim’s You Know My Steez, 141–160, for strong evidence of this from Black youth in the San Francisco Bay Area in Cali). Rickford also notes that “the very presence of certain aspect categories in [Black Language]—particularly the completive (marked by done) and the present durative, or habitual (marked by be)—may be attributed to their prevalence in West African languages, which is well documented in the work of William Welmer and others. Even the existence of a category of remote past (marked by BEEN) may go back to distinctions in languages like LuGanda and KiKongo. Moreover, the tendency of [Black Language] to encode its most important tense-aspect distinctions through a series of preverbal markers (be, bin, done, BIN, fitna, had, and so on) rather than through verbal affixes strikingly parallels the pattern in Caribbean Creoles.” (from John Rickford and Russell Rickford’s Spoken Soul: The story of Black English, New York: John Wiley & Sons, Inc., 2000, p. 154).

236

H. S. Alim and G. Smitherman

copula (is or are) when speaking to White women who know nothing about Hip Hop. Of course, sociolinguists take for granted that Black Language is “rule-governed” and “systematic” like any other language variety. Now, if we compare Obama’s language use in Ben’s Chili Bowl with his language use in Ray’s Hell Burger, we can get a really good sense of how Obama shifts styles. One of Obama’s favorite restaurants, Ray’s Hell Burger, is located in Arlington, Virginia and has a predominantly White clientele. In this nearly eight-minute clip, Obama’s language is informal (“How’re you doin, man?”; “We’ll check that out.”), but he does not use features of Black Language.7 Obama is seen, however, interacting briefly with a Latina employee, whom he winks at and says, “Hola,” in a form better than el español típico de los gringos. His Spanish greeting was taken up by the employee who responded right away, “Hola!” Revisiting Obama’s Ben’s Chili Bowl experience, we have generally the same speech situation (lunchtime at an informal restaurant), speech event (a service encounter between customer and employee) and speech act (ordering food).8 We see Barack Obama in a crowded restaurant with a multiracial crowd in Washington D.C.: Obama: [Handing over his money to the cashier] You just keep that. Where’s my ticket. You got my ticket? Cashier: [Offers Barack his change] Obama: Nah, we straight. [Reaching over to take his soda] Customer: You got cheese fries, too? Obama: Nah, nah, that’s you, man… [Video cuts away and returns after Obama receives his chili dog] Obama: Now, do y’all have some Pepto Bismol in this place? All present: [Laughter] Obama: [Walking back up to the counter, addressing cashier again] Hey, how come he’s got some cheddar cheese on his and I don’t have any on mine? All present: [Laughter] Woahhh! Cashier: Whatever you like, sir. Obama: We got some cheese, you can sprinkle on it? [Gesturing the sprinkling of cheese, then signifyin] Not, not, not, not the Velveeta but the…

7 Check it: http://www.youtube.com/watch?v=TDy9I9C1xUM&feature=related. Last accessed: 0901-2011. 8 Linguistic anthropologists will recognize these terms. They refer to an approach to the scientific study of a culture and their communication patterns known as “the ethnography of communication.” A speech situation, the largest level of the three levels of analysis, describes the social occasion in which speech may occur (in our example, lunchtime at an informal restaurant). You will hear many speech events inside of a speech situation (in our example, a service encounter between customer and employee). A speech acts refers to each action of speech inside of a speech event (in our example, ordering food). Check John Gumperz and Dell Hymes’s edited volume for an early classic, Directions in Sociolinguistics: The ethnography of communication (New York: Holt, Rinehart & Winston, 1972).

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

237

Customers: [Laughter] Customer: The cheddar cheese! Obama: The cheddar cheese In addition to the three main features we discussed in “Nah, we straight,” we can see here that Obama’s language is generally informal with phrases like, “You got my ticket?” and “Nah, nah, that’s you, man.” We also see his use of other features of Black Language (and Southern varieties of English), which speakers often use in more casual environments. Obama’s use of y’all (“Do y’all have some Pepto Bismol in this place?”), for example, is the preferred way to mark the second-person plural on such occasions. In addition to Obama’s language (grammatical structure), we can also look at his style (language use). The Pepto Bismol joke shows Obama’s use of humor, which flows into a type of “banter” that many African Americans know well. In this case, Obama expresses his discontent about not getting cheese on his chili dog in a lighthearted and humorous example of signifyin.9 “Not, not, not, not the Velveeta…” is characteristic of a sometimes subtle mode of discourse in Black communication that includes acts like snappin, bustin, crackin, playin the dozens or dissin someone through wit and humor. Here, the President of the United States wanted some real cheese, not that fake Velveeta stuff!

13.5 The Art of Signifyin and Talkin Trash Other examples of Obama’s signifyin abilities include his “roast” of Donald Trump at the White House Correspondent’s dinner. The roast included some classic signifyin that, although light-hearted and entertaining, was incredibly witty and cutting. In the weeks before the dinner, Donald Trump appeared on numerous media outlets championing the “Birther Movement,” insinuating that he’d run for President, and directing some aggressive questions at the President. Well, the President had some answers: …Donald Trump is here tonight. Now I know that he’s taken some flack lately, but no one is happier, no one is prouder to put this birth certificate matter to rest than Donald. And that’s because he can [letting out a laugh under his breath] finally get back to focusing on the issues that matter, like, did we fake the moon-landing? [Crowd laughter] What really happened in Roswell? [Crowd Laughter] And where are Biggie and Tupac? [Big laughter and 9 Signifyin

has been described as a means to encode messages or meanings in conversation, usually involving an element of indirection. According to Claudia Mitchell-Kernan: “The black concept of signifying incorporates essentially a folk notion that dictionary entries for words are not always sufficient for interpreting meanings or messages, or that meaning goes beyond such interpretations. Complimentary remarks may be delivered in a left-handed fashion. A particular utterance may be an insult in one context and not in another. What pretends to be informative may intend to be persuasive. Superficially, self-abasing remarks are frequently self-praise.” Check out her classic article, “Signifying and Marking: Two Afro-American Speech Acts” in John J. Gumperz and Dell Hymes, eds., Directions in Sociolinguistics (New York: Holt, Rinehart & Winston), 1972, p. 82.

238

H. S. Alim and G. Smitherman

applause] All kidding aside, obviously we all know about [gesturing out towards Trump] your credentials and breadth of experience [Crowd laughter]… um, for example, um… [Donald Trump is shown uncomfortably scratching the side of his neck with his index finger]… No, seriously, just recently in an episode of Celebrity Apprentice [Crowd laughter], at the Steakhouse, the men’s cooking team did not impress the judges from Omaha Steaks, and there was a lotta blame to go around, but you, Mr. Trump, recognized that the real problem was a lack of leadership. And so ultimately you didn’t blame Lil Jon or Meatloaf [Crowd laughter], you fired Gary Busey! [Crowd laughter] [Then matter-of-factly, Barack adds] And these are the kinds of decisions that would keep me up at night. [Uproarious crowd laughter and applause]

First, Barack Obama made out the conspiracy theory that somehow the President of the United States is not a citizen of his own country to be completely foolish. Second, he framed Trump as inane and inept for being preoccupied with making calls for the President to “prove” his citizenship instead of focusing on more serious issues. Third, the President cut deeper into Trump by highlighting his lack of legitimate political experience. He framed him as nothing more than a reality show star, busy makin “serious” decisions like who was to blame for “the failure of the men’s cooking team.” Then in a one-two punch, he “praised” Trump for firing the right chef on his show, and quickly followed with, “And these are the kinds of decisions that would keep me up at night.” With this one-liner, he underscored the enormous difference between himself and Trump in terms of their political experience and capacity to govern. (In 2016, one wishes that Democratic candidates for president, Hillary Clinton and Bernie Sanders, would take a page out of Obama’s book.) Another example of Obama’s signifyin was recorded in an interview with Hip Hop mogul P. Diddy.10 Four years before his election as President, Diddy described an exchange he had with Barack Obama in these terms: “I had the privilege to meet Barack Obama, interview him… and also joke around with him, have some, you know, we had some funny banter back and forth. We was really like snappin on each other…” In the clip, Obama, after wiping the sweat off of his forehead repeatedly throughout the interview, is urging young people to vote (for Jim Kerry, at the time, instead of George W. Bush): Well, some people just gotta remember what happened in Florida, you know, when George Bush won the Presidency, he thinks, based on just a tiny [pronounced “tahny,” the monophthongization of diphthongs again] margin of votes… And like I said, don’t let people over promise what you can do through politics. It’s not gonna solve the problems of the entire world, but it makes a little bit of difference… You get registered, you vote – that takes about 15 min. And if you can’t spend 15 min on deciding what your community’s gonna look like and what your country’s gonna look like, then you don’t have any cause to complain.

Diddy: He makin sense [No copula needed]. That’s what we need. We need people to make sense. We applaud you. [Then the signifyin begins]… And I wanna apologize for not sweatin, but I do this so much… Obama: [Begins to protest and takes off his suit jacket] [Unseen staff start laughing] 10 This

interview was taped for www.diddy.com. You can catch it at: http://www.youtube.com/ watch?v=Ne_87Kw35pE. Last accessed: 09-01-2011.

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

239

Diddy: … I’m so cool. I just want y’all to see. Everybody I’m interviewing is sweatin. I’m not even touchin my brow. [Laughter from staff continues]. I’m so cool. [Barack still wiping sweat off his face] And I wanna apologize. [Then he really begins clownin] I ain’t tryna make you look bad or nu’in like that but I’m just so cool. Um, we, we… Obama: [Talking to the camera, pointing at Diddy] He, he wearin a T-shirt… [No copula needed] Diddy: [Bent over laughing] Obama: … I tell ya, if he was wearin one of those fancy designer clothes he’s designin, he’d be sweatin just like me. Diddy: The guy’s good. The guy’s good y’all. Let’s give it up for him. Obama: [Slapping Diddy on the back, smiling] I appreciate you guys, thank you. Diddy: Peace, peace, thank you. Obama has proven himself to be a pretty skilled signifier. Now, there are two points of interest here in terms of styleshifting. One, of course, is how Obama was able to engage the Black cultural mode of discourse known as signifyin. The other is how he accommodated his linguistic style closer to the style of P. Diddy throughout the interview. What began as extremely formal and reserved, with mostly “standard” English responses, ended with several examples of Black phonological and grammatical features. Diddy, generously using zero copula throughout the interview, finally gets one back from Obama (He Ø wearin a T -shirt) as Obama shifts his linguistic style structurally to match the Black discourse genre. That example of shifting between discourse modes and linguistic forms in the same interaction is a rich instance of styleshifting. These few examples (and there are many others) show that Barack Obama is skilled in the art of signifyin and talkin trash. In fact, he talks about learning these skills in high school while playing on the university basketball courts: “where a handful of black men… would teach [him] an attitude that didn’t just have to do with the sport. That respect came from what you did and not who your daddy was. That you could talk stuff to rattle an opponent, but that you should shut the hell up if you couldn’t back it up.11 ” Further, in his high school years, Obama writes about an illustrative exchange between him and his friend Ray, who introduced him to “the black parties that were happening on the army bases.12 ” Ray and other Black friends “eased [Obama’s] passage through unfamiliar terrain” in the process of his becoming Black, or becoming well-versed with Black American cultural signs, symbols, tropes, practices, and worldviews. As we discussed in the beginning of the chapter, language use was central to the cultural process of socialization.

11 From

Barack Obama’s Dreams from My Father: A story of race and inheritance (New York: Crown Publishers, 1995, p. 79). 12 Ibid, p. 72.

240

H. S. Alim and G. Smitherman

13.6 “Pastor-in-Chief”: Barack Obama’s Use of “the Black Preacher” Style Barack Obama, as President-In-Chief of the U.S., might also be referred to as “PastorIn-Chief” for the way he uses a style of speech associated with Black preachers. Despite his ability to shift into Black Language styles across many contexts, and signify with the best speakers, Americans most often described Barack Obama’s speech as “mirroring” that of a Baptist preacher. Black Americans, in particular, were not only more likely to frame Obama as a Black preacher, but they also usually provided more nuanced readings of his “preacher style.” Collectively, Black Americans touched on Obama’s cadence, timing, effective use of pauses, metaphors, rhythm and repetition, as well as Black discourse modes of signifyin and storytelling. They described Obama’s “preacher-like” speech as: – having a slow and pointed cadence… words intermittently separated with pauses pregnant with meaning…. – he uses the passion and rousing speech tools of preachers in Black churches… such as signifyin’, using words that have double meaning that blacks pick up…. – he uses repetition… altering pitch and stress…. – he adopts a more Pastorial African-American vernacular and references more Biblical verses…. – bringing… citizens along with storytelling and narration… his storytelling ability usually wraps around to connect to a larger theme…. – often he uses metaphors and stories… His ability to tell stories is one of his greatest strengths… through persuasive storytelling, he taps into the unconscious mind where we make decisions, making it that much easier for him to influence the audience through his language…. – he consciously uses sophisticated code-switching and rhythmic patterns…. Generally, White Americans were not as descriptive, but almost all of those who noted his preacher style linked him to iconic Black preachers and ministers, such as Martin Luther King or Malcolm X. Having presumably less experience worshipping in Black churches, some White Americans even described Obama as “singing” in his speeches. While Black Americans do often refer to a “sing–song” quality in Black speech, the Black survey respondents did not describe Barack this way, probably due to the relative flatness of his speech when compared to the best Black preachers. One White male respondent compared Barack Obama’s style to that of a “preacher” and then immediately made the direct link to “MLK’s” “singing” or “chanting”: “At his best, he has the deliberate and enthusiastic pace of a talented preacher. It’s almost as if he’s singing or chanting as opposed to talking. This, of course, is not unlike how other talented orators, like MLK, sound.” The next example is the most detailed description of Barack Obama’s preacher style provided by the survey’s White respondents: Obama’s composure always remains cool and collected with a strong sense of inner peace – he never lets emotional intensity take over his speeches. At the same time, he is also the 21st century echo of African-American preacher style characterized by such strong

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

241

orators as MLK… Additionally, his speech is effective because his delivery is not boring or monotonous, but rather like a song. The way Obama alters his pace, tone, and rhythm is similar to the way a preacher speaks, which is essentially close to singing. The intonation, emphasis, and pauses and silences that characterize his speaking style are churchy and religious.

As we saw in the opening of this chapter, Obama is indeed particularly well-versed in a mode that draws on Black preacher style. No doubt his time in Trinity was a first-hand language immersion experience in the Black Church’s ways with words. In his “A More Perfect Union” speech (aka “The Race Speech”), for example, which we analyze in great depth in Articulate While Black, Obama painted a vivid picture of Reverend Wright’s Trinity services and connected with many Black church-goers and others who recognize the church as an important Black cultural institution: “Like other black churches, Trinity’s services are full of raucous laughter and sometimes bawdy humor. They are full of dancing and clapping and screaming and shouting.” Obama learned a mode of Black sermonizing. While not attempting to duplicate it to the letter in the political sphere, he readily engaged in a “stylistic sampling” of the Black Church’s Oral Tradition, as one respondent put it. In his “More Perfect Union,” for example, Obama began his speech-sermon by framing slavery as America’s “original sin.” Opening with this religious frame primed the audience for the most important moment of the speech, which to us, sounded like the climax of a sermon. From approximately the last 8 min of his speech, Obama uses a number of Black Preacher style rhetorical devices. He cites Scripture, offers the flock (“Americans”) a choice between good and evil/right and wrong, and then through effective use of timing, repetition and narrativizing offers us a way to perfect our character (ourselves and the Union). The only way to truly witness the man’s skillz is to examine the lengthy excerpt below, which is notated to highlight his multilayered use of repetition13 and which includes a truncated sample of his storytelling: In the end, then, what is called for is nothing more, and nothing less, than what all the world’s great religions demand – that we do unto others as we would have them do unto us. Let us be our brother’s keeper, Scripture tells us. Let us be our sister’s keeper… For we have a choice in this country. We can accept a politics that breeds division, and conflict, and cynicism. We can tackle race only as a spectacle… We can play Reverend Wright’s sermons on every channel, every day and talk about them from now until the election… We can pounce on some gaffe by a Hillary supporter as evidence that she’s playing the race card, or we can speculate on whether white men will all flock to John McCain in the general election regardless of his policies. We can do that. [He repeats “We can” and articulates the entire phrase in a lower, breathy voice to give it the sound of genuine feeling. Pausing to add rhetorical effect] But if we do, I can tell you that in the next election, we’ll be talking about some other distraction. And then another one. And then another one. And nothing will change. 13 This excerpt is notated to demonstrate the multilayered use of repetition. For example, the phrase “we can” is marked in bold. Each instance of “this time” is underlined. Each use of “we want to talk about” or “we want” is in italics. Overlapping repeated phrases like, “This time we want to talk about,” are marked with “This time” underlined and in italics. Whole phrases like: I am here because of Ashley are marked in bold and underlined.

242

H. S. Alim and G. Smitherman

That is one option. Or, at this moment, in this election, we can come together and say, “Not this time.” This time we want to talk about the crumbling schools that are stealing the future of black children and white children and Asian children and Hispanic children and Native American children. This time we want to reject the cynicism that tells us that these kids can’t learn… They are our kids, and we will not let them fall behind in a 21st century economy. Not this time. This time we want to talk about how the lines in the Emergency Room are filled with whites and blacks and Hispanics who do not have health care… This time we want to talk about the shuttered mills that once provided a decent life for men and women of every race, and the homes for sale that once belonged to Americans from every religion, every region, every walk of life. This time we want to talk about the fact that the real problem is not that someone who doesn’t look like you might take your job; it’s that the corporation you work for will ship it overseas for nothing more than a profit. This time we want to talk about the men and women of every color and creed who serve together, and fight together, and bleed together under the same proud flag. We want to talk about how to bring them home from a war that never should’ve been authorized and never should’ve been waged, and we want to talk about how we’ll show our patriotism by caring for them, and their families, and giving them the benefits they have earned… There is one story in particular that I’d like to leave you with today - a story I told when I had the great honor of speaking on Dr. King’s birthday at his home church, Ebenezer Baptist, in Atlanta. There is a young, twenty-three year old woman, a white woman, named Ashley Baia who organized for our campaign in Florence, South Carolina. She had been working to organize a mostly African American community since the beginning of the campaign… Now Ashley might have made a different choice. Perhaps somebody told her along the way that the source of her mother’s problems were blacks who were on welfare and too lazy to work, or Hispanics who were coming into the country illegally. But she didn’t. She sought out allies in her fight against injustice. Anyway, Ashley finishes her story and then goes around the room and asks everyone else why they’re supporting the campaign… And finally they come to this elderly black man who’s been sitting there quietly the entire time. And Ashley asks him why he’s there. And he does not bring up a specific issue… He simply says to everyone in the room, “I am here because of Ashley.” “I’m here because of Ashley.” By itself, that single moment of recognition between that young white girl and that old black man is not enough… But it is where we start…

In the 8 min excerpt quoted above, Obama repeats “we can” seven times in succession (and emphasizes it once for rhetorical effect) before moving listeners to “we want,” shifting the focus from our failures to our collective goals for action. He then effectively uses a combination of the phrases “Not this time,” “this time,” and “this time we want to talk about” to begin ten different successive ideas.14 All the while, he presents us with the choice between “division, and conflict, and cynicism” and 14 Georgetown University professor Michael Eric Dyson also notes Obama’s use of “anaphora,” which is the repeating of the “same word or phrase at the beginning of successive sentences.” What’s interesting here is that Obama layers his repetition of multiple words and phrases, creating an advanced use of this strategy, one that is common in the Black preacher tradition. See Dyson’s full comments and other examples of Obama’s use of this rhetorical device at: http://www.smh.com.au/news/opinion/a-presidentpreacher-from-anaphora-to-epistrophe/2009/ 01/18/1232213445525.html. Last accessed: 09-02-11.

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

243

coming together to say, “Not this time.” Finally, he humanizes that choice with the story of young White Ashley and an older Black man and suggests that together we can work to perfect the union. In this “A More Perfect Union” speech, Obama was addressing a national audience with folks across the racial and linguistic spectrum. In majority Black contexts, however, where Black linguistic norms prevail, Barack Obama’s been known to employ his Black speech style even more. Specifically, as we saw in the opening narrative of the chapter, he can shift into a deep Black style of call-and-response, a communicative strategy that breaks down conventional divisions between “audience” and “speaker”.15 Shot through with action and interaction, call-and-response is concentric in quality, with the “audience” becoming both observers and participants in the speech event. The “audience’s” verbal and non-verbal responses co-sign the power of the “speaker’s” call. Barack Obama’s speech in front of a predominantly Black crowd in South Carolina provides a quintessential example. Walking across the stage, he looked out into the audience: Obama’s call: They’re tryna bamboozle you… [Pause] Crowd’s response: [Black woman seen waving her sign like a fan, Black men shaking their heads in recognition] [Crowd Laughter] Yes! Obama’s call: It’s the same old okey-doke… [Pause] Crowd’s response: [Laughter, agreement] That’s right! Obama’s call: [Looking out to audience with a half-smile] Y’all know about okeydoke, right? [Pause] Crowd’s response: Yeahhh! Yes! [Laughter] Obama’s call: It’s the same old stuff! Crowd’s response: Yeahhh! Obama’s call: Just like if anybody starts gettin one of these emails sayin, “Obama is a Muzlim.” [Pause] Crowd’s response: Yes! They do it! Obama’s call: I’ve, I’ve been a member of the same church for almost twenty years… [Pause] Crowd’s response: C’mon now! Alright! Obama’s call: Prayin to Jesus!! Crowd’s response: [Hits a climax with uproarious shouts and applause!] Obama’s call: Wit mah—wit mah Bible [pronounced Bahble]… [Pause] Crowd’s response: Amen! [Continued applause] Obama’s call: Don’t LET people turn you around [continued applause] because they’re just makin stuff up! Crowd’s response: Yes they are! Obama’s call: That’s what they do! Crowd’s response: Yes they do! Obama’s call: They try to bamboozle you! 15 For

great examples of Obama’s rendering of call-and-response in text, check Dreams from My Father, pp. 293–295. You can also hear this portion of Obama’s South Carolina speech here: http:// www.politico.com/news/stories/0309/19538_Page2.html. Last accessed: 09-01-11.

244

H. S. Alim and G. Smitherman

Crowds response: [Laughter] Hoodwink you! Obama’s call [now a response to “Hoodwink you!”]: Hoodwink you! [Laughter]. Barack Obama’s masterful use of the call-and-response mode of Black communication literally transformed this venue in South Carolina into a Baptist church filled with spirit. However, while Black Americans were shouting, hollering “Amen!”, and going back and forth with Obama in a culturally familiar verbal dance—until the lines between “caller” and “responder” were literally blurred—most White Americans on the scene were either looking on blankly or smiling quietly. It’s quite possible that White Americans knew that something else was going on but couldn’t quite figure out what it was. Of course, to Blacks in the audience, and to most reading this now, Obama was also borrowing some of Malcolm X’s (ironically, a Muslim himself) most famous lines about White people trying to “bamboozle” and “hoodwink” Black Americans. Like a coded verbal game of catch, Obama threw it out, the Black audience caught it, and then threw it back for him to catch. It wasn’t that White audience members didn’t approve of what was being said; it was that they were simply unable to play the game.16

13.7 Obama’s Successful Raciolinguistic Project: Familiarly White, Familiarly Black, Familiarly Male, Familiarly American, Familiarly Christian In this final section, we present an examination of the syntax and style of Obama’s language use in order to return to one of the central questions for the project of raciolinguistics that we raised in the opening of this chapter: What does it mean to speak as a racialized subject in contemporary America? All of Obama’s flexible linguistic abilities that we have described thus far were critically important to his being elected. This was perhaps the single most consistent finding in our survey: Barack Obama’s mastery of White mainstream English ways of speaking, or “standard” English, particularly in terms of syntax, combined with his mastery of Black cultural modes of discourse, in terms of style, was an absolutely necessary combination for him to be elected America’s first Black President. One respondent in particular articulated this 16 Writing about Black music, Imani Perry writes about another level of call-and-response. “To make

something good… means in part to effectively employ the call-response trope on several levels, and, just as important, to know what is good requires a sophisticated… understanding of the symbolic references and cultural history from which the music derives.” (Prophets of the Hood: Politics and poetics in hip hop, Durham, NC: Duke University Press, p. 36). Relating this to Obama’s speech in South Carolina, Barack put out the encoded Malcolm X call and his Black audience responded. He was also employing another level of signifyin, one that is central to the Black literary tradition. According to Henry Louis Gates, Jr.’s The Signifyin(g) Monkey; A theory of African-American literary criticism (New York & London: Oxford University Press, 1989), signification relies on one’s knowledge of previous texts and the author’s (speaker’s) ability to reinterpret them in new ways. Certainly, signifyin on a Muslim minister’s words to ensure that he himself was seen as anything but a Muslim qualifies.

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

245

sentiment perfectly. When asked about Barack Obama’s language and language use, she explained: When Obama was on the campaign trail, his speeches mirrored that of a Baptist preacher. The way certain words were stressed and the rise and fall in his speech were very reminiscent of the church. Sprinkled with imagery, metaphors and historical references, coupled with an underlying theme and you had speeches that captivated not only Americans, but the world… I feel like Obama has been able to balance his multi-racial identity and his Black experiences. His speeches are a great example of that balance. Obama has the ability to use Standard English in a “Black” context by using the “preacher” format to develop his speeches and then delivering them in Standard English. By combining these two experiences, Obama was able to appeal to a larger audience of people. Whites did not feel alienated by his language, and Blacks felt a sense of familiarity with his speech pattern.

Of course, mastery of so-called “standard” English is made necessary in American politics (we’ll return to this later), but it was Obama’s ability to combine this variety with Black male ways of speaking that was ultimately crucial. His linguistic style mattered in at least three ways. First, Barack Obama’s mastery of White mainstream ways of speaking allowed White Americans to feel more comfortable with him. He used a language variety that was familiarly White, which rightly or wrongly, did not “alienate” Whites in the way that Black Language sometimes does. Relatedly, his style of speaking was seen as “transcending” Blackness, with many describing him as “exceptionally articulate,” making (unintentional) racist links between “articulateness,” “Whiteness,” and “intelligence.17 ” Though some Americans noted that White, male mainstream ways of speaking English are problematically mapped onto “the language of politics” and “the language of success,” Black Americans highly regarded Obama’s proficiency in this style as well. Using positive terms, many respondents across racial lines described Obama’s ability to use “standard English,” “typical American English,” “normative English,” “standard American English,” “polished standard English”—and our personal (and hilarious) favorite, “a language literally born of the American educational system’s upper echelon.” Second, not only did Whites feel that Obama spoke familiarly White, many Black Americans felt that he spoke familiarly Black. While some Black women respondents noted that his “sounding Black” had to do with his “manly (deep) voice” or his “baritone” (he was also familiarly male in his sense), more often Blacks described Obama’s speech style in terms of “a Baptist preacher” or in the “tradition of the Black church.” (which is also overwhelmingly male). So, while responding positively to Obama’s command of “standard” English syntax, the real clincher for Black Americans was that Barack could speak in a style that was recognizable to the community as “something we do.” Rightly or wrongly, for many Black Americans, anything less than that might have aroused suspicion. This is because, sociolinguistically speaking, the way we use language often “hints at” our politics, indexing our (dis)alignment 17 The

phrase “exceptionally articulate” was actually used by one White American and “articulate” was used overwhelmingly by White respondents more than any other group. This led us to develop the idea of “articulate as an exceptionalizing discourse.” (please, see Chap. 2 of Articulate While Black, Alim and Smitherman (2012), for a much deeper analysis of “articulate”).

246

H. S. Alim and G. Smitherman

with particular groups or causes. We read into people’s words for clues, signs, anything that might help us figure out where they stand. In the case of Barack Obama, accurately or not, many Black Americans read his use of Black modes of discourse as indexing a political alignment with the Black community. Thirdly, Obama’s ability to bring together “White syntax” with “Black style” and to speak familiarly Black was not only important for the Black community. It was also critically important for the White community for at least two reasons. One, Whites have always found Black preacher style interesting, so long as preachers did not critique Whiteness too hard in that caustic, biting, damn-you-to-hell kind of way. The second and most critical reason why speaking familiarly Black was important for Whites is this: It made Obama both “American” and “Christian.” Not only are White Americans more familiar with a Black Christian identity, but due to the contentious history of the Nation of Islam and contemporary tensions with immigrant Muslims in post-9/11 America, many Whites also fear “(Black) Muslims.” Speaking familiarly Black made Obama familiarly American and familiarly Christian. To borrow from one Asian American respondent who wrote about forever feeling like a “foreigner” in the U.S., “Barack needed to not only be American; he needed to be 110% American.” After all, who can forget the lunacy of some White folks at the McCain-Palin rallies (“I, I, I don’t trust him—he’s A-A-Arab!”)? And the never-ending and overwhelmingly White “Birther Movement,” which even includes current 2016 Republican presidential candidates like Donald Trump? Growing up in Hawai’i and Indonesia with a Kenyan father and Muslim family roots was apparently too much for White Americans to handle. But White Americans are not the only ones who express xenophobia and antiMuslim bias. Sounding familiarly Black, and thus familiarly American and familiarly Christian, also won over those in the Black community who questioned Obama’s heritage (“He ain’t Black – he from Kenya!” or “Ain’t he a Mooozlim?”) or did not agree with what they saw as his appropriation of the Black American struggle (“He’s probably one of those Africans who doesn’t like us, but will use the label ‘African American’ to take advantage of affirmative action programs”). To sum up, Obama’s styles of speaking clinched his victory because he put most Americans at ease. Here was a Black male candidate for President whom Black Americans could trust because “he sounds White, but not too White,” and White Americans could trust because “he sounds Black, but not too Black.” Of course, it would be too simple to leave it there. The reality is that Whites too were happy with a Black man who “sounded White, but not too White.” His familiarly Black style Americanized and Christianized him, helping them get over their irrational fears of a “foreign Muslim” or a “socialist African.” It is also likely that Blacks too were happy with a Black man who “sounds Black, but not too Black.” Though not widely discussed, because of Black Language’s marginalized status in broader American society, some Black folks suffer a linguistic shame that hyper-criticizes any speech that sounds “too Black.” The stories of people “cringing” every time they hear Magic Johnson speak, for example, are all too common. In a similar way that Barack Obama’s familiarly Black style helped some White folks get over irrational fears of a “foreign/Black Muslim” or a “socialist African,” his familiarly White style

13 Amazing Grace: An Analysis of Barack Obama’s Raciolinguistic …

247

helped some Black folks get beyond irrational insecurities that “the whole race” would be deemed “ignorant” because of one Black person’s speech. Caught in between discriminatory discourses of language, citizenship, religion and race, Barack Obama’s language use, perhaps miraculously, managed to appeal to the majority of Americans. It did not matter how many times he repeated that he wasn’t a Muslim or how many times he presented his birth certificate. What mattered more to most Americans, even if subconsciously, was not what he said but how he said it. More than any other cultural symbol, Barack Obama’s multifaceted raciolinguistic performances allowed Americans to create linguistic links between him and famous African American male historical figures. These links served to simultaneously “Whiten,” “Blacken,” “Americanize,” “Christianize” and gender Obama as “Male” in the eyes/ears of both Black and White Americans. These raciolinguistic performances, to us, were delivered not only with amazing style, but also, with “amazing grace.”

References Alim, H. S. (2004). You know my Steez: AN ethnographic and sociolinguistic study of styleshifting in a Black American speech community. Durham, NC: Duke University. Alim, H. S. (2005). Hearing what’s not said and missing what is: Black Language in White public space. In S. Kiesling, & C. B. Paulston (Eds.), Intercultural discourse and communication: The essential readings (pp. 180–197). Malden, MA: Blackwell. Alim, H. S. (2009). Racing language, languaging race. Paper presented at the University of California, Los Angeles Symposium on Race & Ethnicity in Language, Interaction, and Culture, February 27th, 2009. Alim, H. S., & Reyes. (2011). Complicating race: Articulatory race across multiple social dimensions. Special Issue of Discours & Society, 22(4). MA: Blackwell. Alim, H. S., & Smitherman, G. (2012). Articulate wHite Black: Barack Obama, language and race in the U.S. New York: Oxford University Press. Alim, H. S., Rickford, J. R., & Ball A. (Eds.). (2016). Raciolinguistics: How language shapes our ideas about race. New York: Oxford University Press. Bucholtz, M. (2011). White kids: Language, race and styles of youth identity. Cambridge, UK: Cambridge University Press. Dick, H., & Wirtz, K. (Eds.). (2011). Racializing discourses. A Special Issue of the Journal of Linguistic Anthropology, 21(1). Ibrahim, A. (2003). Whassup, homeboy? Joining the African Diaspora: Black English as a symbolic site of identification and language learning. In S. Makoni, G. Smitherman, A. Spears, & A. Ball (Eds.), Black linguistics: Language, society and politics in Africa and the Americas. New York: Routledge. Ibrahim, A. (2014). The Rhizome of blackness: A critical ethnography of hip-hop culture, language, identity and the politics of becoming. New York: Peter Lang. Makoni, S., Smitherman, G., Ball, A. & Spears, A. (Eds.). (2003). Black linguistics: Language, society and politics in africa and the americas. New York/London: Routledge. Mendoza-Denton, N. (2008). Homegirls: Language and cultural practice among Latina youth. Malden, MA: Blackwell. Morgan, J. (2009). Black like Barack. In Sharpley-Whiting, T. D. (Ed.), The speech: Barack Obama’s ‘a more perfect union’. New York: Bloomsbury.

248

H. S. Alim and G. Smitherman

Reyes, A. (2007). The other Asian: Language, identity and stereotype among Southeast Asian American youth. Mahwah, NJ: Lawrence Erlbaum. Reyes, A., & Lo., A. (Eds.). (2009). Beyond yellow English: Toward a linguistic anthropology of Asian Pacific America. New York: Oxford University Press. Rosa, J. (2017). Looking like a language, sounding like a race: Inequality and ingenuity in the learning of Latina/o identities. New York: Oxford University Press. Roth-Gordon. (2016). Race and the Brazilian body: Blackness, whiteness, and everyday language in Rio de Janeiro. New York: Palgrave Macmillan. Spears, A. (1999). Race and ideology: Language, symbolism and popular culture. Detroit: Wayne State University Press. Urciuoli, B. (1996). Exposing prejudice: Porto Rican experiences of language, race and class. Boulder, CO: Westview Press. Zentella, A. C. (1997). Growing up bilingual. Malden, MA: Blackwell.

Part V

Syntax, Semantics and Cognition

Chapter 14

Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’Problem Michael Zock

To search for a word in a dictionary without a proper index is like looking for an address in a city without a decent map.

Abstract Dictionaries are repositories of knowledge concerning words. While readers are mostly concerned with meanings, writers are generally more concerned with word forms (lemma) expressing meanings. I will focus here on this latter task: building a tool to help authors to find the word they are looking for, word they may know but whose form is eluding them. Put differently, my goal is to build a resource helping authors to overcome the Tip-of-the-Tongue problem (ToT). Obviously, in order to access a word, it must be stored somewhere (brain, resource). Yet this is far from sufficient. Access may depend on many other factors than storage of word forms: organization of the dictionary (index), the user’s cognitive state, i.e. available knowledge at the onset of search, the distance between the source- and the targetword (direct neighbor or not) , the knowledge of the relationship between the two, etc. I will try to provide evidence for the claim that (word) storage does not guarantee access. To this end I will compare a well-known lexical resource, WordNet (WN), to an equivalent one, but bootstrapped from Wikipedia (WiPe). While both versions contain basically the same set of words, the latter contains many more (syntagmatic) links than WN. This is probably the reason why WiPe outperforms WN. In the last two sections I will explain under what conditions WN is suitable for word access, and what it might take to go beyond the limitations of this famous resource.

M. Zock (B) Aix-Marseille Université, CNRS, LIF (UMR 7279), 163, Avenue de Luminy, 13288 Marseille, France e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_14

251

252

M. Zock

14.1 Introduction Speaking a language can be a daunting task. Planning what to say (message) and how to say it (linguistic form) have to be carried out on the fly, that is, quasy simultaneously, and while finally producing the desired form (articulation) one may have to plan already the next stretch of discourse. Actually, speaking is quite a bit more complex than that, requiring the solution of at least half a dozen of problems: determine content, find suitable words and sentence frames, add function words, perform morphological operations (agreement) and articulate the resulting form(s). To get a better idea of the complexity of the task at hand let’s focus only on one of them, lexical access. Bear in mind though that speech is fast (2–5 words per second, Levelt 1993, 2001), and that words must be found in a resource containing, say, 100,000 words (Miller 1991). This raises a number of interesting questions: • how do people manage to access words so quickly in such a huge lexicon, succeeding most of the time while making very few mistakes? • can the insights gained by studying the mental lexicon (storage, organization, process) be transposed to an external resource? Put differently, can we use this knowledge to enhance paper- or electronic dictionaries (off-line processing)? • If ever the answer is negative. Are there ways to achieve something equivalent, at least in terms of precision, that is, can we build or enhance an existing resource (electronic dictionary) in such a way as to allow its users to find quickly the desired word? Having addressed the first two questions elsewhere (Zock et al. 2010), I will focus here mainly on the last problem, building a resource meant to help users to overcome the tip-of -the-tongue problem.1 Hence, functionally speaking I try to achieve something equivalent to the human brain, though in slow motion: help people to find the word they are looking for. Before showing the way how this can be done, let me say a few words concerning the word-access problem. When speaking or writing we encounter basically either of the following two situations: one where everything works automatically (Segalowitz 2000), somehow like magic, words popping up one after another as in a fountain spring, leading to a discourse where everything flows like in a quiet river (Levelt et al. 1999; Rapp and Goldrick 2006). The other situation is much less peaceful: discourse being hampered by hesitations, the author being blocked somewhere along the road, forcing her to look deliberately and often painstakingly for a specific, possibly known word (Zock et al. 2010; Abrams et al. 2007; Schwartz 2002; Brown 1991). I will be concerned 1 The

ToT problem is characterized by the fact that the author has only partial access to the word form s/he is looking for. The typically lacking parts are phonological (Aitchison 2003). The ToT problem is a bit like an incompleted puzzle, containing everything apart from some minor small parts (typically, syllables, phonemes). Alas, not knowing what the complete picture (target, puzzle) looks like, we cannot determine the lacking part(s). Indeed, we cannot assume to know the target, and claim at the same time to look for it or any of its elements. Actually, if we knew the target (word) there wouldn’t be a search problem to begin with, we would simply produce the desired form.

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

253

here with this latter situation. More specifically, I am concerned here with authors using an electronic dictionary to look for a word. While there are many kind of dictionaries, most of them are not very useful for the language producer. The great majority of them are semasiological, that is, words are organized alphabetically in the resource within which search takes place. Alas, this kind of organisation does not fit well the language producer whose starting points (input) are generally meanings2 or cue-words (primes) and only the end point (outputs) the corresponding target word. To be fair though, one must admit that great efforts have been made to improve the situation both with respect to lexical resources and electronic dictionaries. Since the invention of the thesaurus (Roget 1852) quite a few onomasiological dictionaries have been built (van Sterkenburg 2003; Casares 1942), even if nowadays they are not built by hand anymore (Dornseiff et al. 2004; Rundell and Fox 2002). Today we build lexical resources via corpora (Hanks 2012; Kilgarriff and Kossem 2012), crowdsourcing (Benjamin 2014), dictionary writing systems (Abel 2012) and corpus tools like Sketch Engine (Kilgarriff et al. 2004), MonoConc or WordSmith. Yet thesauri are not the only kind of onomasiological resources. There are analogical dictionaries (Boissière 1862; Robert et al. 1993), collocation dictionaries (Benson et al. 2010), reverse dictionaries (Bernstein 1975; Kahn 1989; Edmonds 1999), rhyme dictionaries (Fergusson and Fergusson 1985 ; Webster 2007), and network-based lexical resources: WordNet (Fellbaum 1998; Miller 1990), MindNet (Richardson et al. 1998), HowNet (Dong and Dong 2006), and Pathfinder (Schvaneveldt 1989). There are Longman’s Language Activator (Summers 1993) and OneLook 3 , which, akin to BabelNet (Navigli and Ponzetto 2012), combines a dictionary (WordNet) and an encyclopedia (Wikipedia). Besides all this there are interesting proposals coming from Fontenelle (1997), Sierra (2000), Moerdijk (2008), and Mel’çuk (Mel’çuk and Polguère 2007). Finally, there is MEDAL (Rundell and Fox 2002), a thesaurus produced with the help of Sketch Engine (Kilgarriff et al. 2004). In parallel to dictionary making a lot of progress has been made by psycholinguists who study the time course of lexical access (Levelt et al. 1999), word associations (de Deyne and Storms 2015) and the structure, i.e. organization of the mental lexicon (de Deyne et al. 2016).4 Clearly, a lot has happened during the last two decades, yet more can be done especially with respect to indexing (the organization of the data) and navigation. This paper is organized as follows. I start by providing evidence that storage does not guarantee access. That this holds for humans has been shown already 50 years ago (Tulving and Pearlstone 1966), in particular via Brown and Mc Neill’s (1966) seminal work devoted to the tip-of -the-tongue problem. I will show here that this can also hold for machines. The assumption that what is stored can also be accessed (anytime), is simply wrong. To illustrate this claim I will compare an extended version of WN (Mihalcea and Moldovan 2001) to an equivalent resource based on Wikipedia 2 More

or less well specified thoughts (concepts, elements of the word’s definition), or somehow related elements: collocations, i.e. associations (elephant: tusk, trunk, Africa). 3 https://www.onelook.com 4 For a short survey of some of this work see Zock et al. (2009), and Zock (2015a, b).

254

M. Zock

(WiPe). Next, I will discuss under what conditions WN is adequate for word access, and finally, I will sketch a roadmap describing the steps to be performed in order to go beyond this very popular resource. The goal is to build a navigational tool (index, association network) allowing authors to go from the word they know (word available when being in the ToT state) to the word they are looking for (target). Before doing so, I will present though my theory concerning the dialogue between the dictionary user and the lexical resource.

14.2 Storage Does Not Guarantee Access To test this claim let me describe here briefly an experiment carried out with a colleague of mine (Zock and Schwab 2011). We ran a small experiment, comparing an extended version of WN (henceforth, WN-x) and Wikipedia, which we converted into a lexical resource. Our goal was not so much to check the quality of WN or any of its extensions as to show, firstly, that storage does not guarantee access and, secondly, that access depends on a number of factors like (a) quality of the resource within which the search takes place (organisation, completeness), (b) index, and (c) type of the query (proximity to the target).5 Having two resources built with different foci, our goal was to check the efficiency of each one of them with respect to word access. For practical reasons we considered only direct neighbors. Hence, we defined a function called direct neighborhood, which, once applied to a given window (sentence/paragraph,6 produces all its co-occurences. Of course, what holds for direct associations (our case here), holds also for indirectly related words, that is, words whose distance >1 (mediated associations).

14.3 Comparisons of the Two Resources Table 14.1 shows the results produced by WN-x and WiPe for the following, randomly given inputs: ‘wine’, ‘harvest’ or their combination (‘wine + harvest’). Our goal was to find the word ‘vintage’. As the results show, ‘harvest’ is a better query term than ‘wine’ (488 vs. 30 hits), and their combination is better than either 5 To show the relative efficiency of a query, D. Schwab has developed a website in Java as a servlet.

Usage is quite straightforward: people add or delete a word from the current list, and the system produces some output. The output is an ordered list of words, whose order depends on the overall score [i.e. the number of co-occurrences between the input, i.e. ‘source word’ (Sw ) and the directly associated words, called ‘potential target word’ (PTw )]. For example, if the Sw ‘bunch’ co-occured five times with ‘wine’ and eight times with ‘harvest’, we would get an overall score or weight of 13: [(wine, harvest), bunch, 13]. Weights can be used for ranking (i.e. prioritizing words) and the selection of words to be presented, both of which may be desirable when the list becomes long. 6 Optimal size is an empirical question, which may vary with the text type (encyclopedia vs. raw text).

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem Table 14.1 Comparing two corpora with various inputs Input Output: WN-x

255

Output: WiPe

wine

488 hits grape, sweet, serve, France, small, fruit, dry, bottle, produce, red, bread, hold…

3045 hits name, lord characteristics, christian, grape, France, … vintage (81st), …

harvest

30 hits month, fish, grape, revolutionary, calendar, festival, butterfish, dollar, person, make, wine, first, …

4583 hits agriculture, spirituality, liberate, production, producing, …, vintage (112th), …

wine + harvest

6 hits make, grape, fish, someone, commemorate, person, …

353 hits grape, France, vintage (3rd), …

of them (6 hits). What is more interesting though is the fact that none of these terms allows us to access the target, eventhough it is contained in the database of WN-x, which clearly supports our claim that storage does not guarantee access not even in computers. Things are quite different for an index built on the basis of information contained in WiPe. The same input, ‘wine’ evokes many more words (3045 as opposed to 488, with ‘vintage’ in the 81st position). For ‘harvest’ we get 4583 hits instead of 30, ‘vintage’ occurring in position 112. Combining the two yields 353 hits, which pushes the target word to the third position, which is not bad at all. I hope that this example is clear enough to convince the reader that it makes sense to use real text (ideally, a well-balanced corpus) to extract from it the information needed (associations) in order to build an index allowing users to find the elusive word. One may wonder why we failed to access information contained in WN and why WiPe performed so much better. We believe that the relative failure of WN is mainly due to the following two facts: the size of the corpus (114,000 words as opposed to 3,550,000 for WiPe), and the number of syntagmatic links, both of which are fairly small compared to WiPe. Obviously, being an encyclopedia, WiPe contains many more syntagmatic links than WN. Of course, one could object that we did not use the latest release of WN which contains many more words (147,278 words, clustered into 117,659 synsets). True as it is, this would nevertheless not affect our line of reasoning or our conclusion. Even in this larger resource we may fail to find what we are looking for because of the lack of syntagmatic links.7 7 It should be noted though that serious efforts have been made to enrich WN by adding syntagmatic

links (Bentivogli and Pianta 2004) and various kinds of encyclopedic information: topic signatures (Agirre et al. 2001), domain-specific information (Boyd-Graber et al. 2006; Gliozzo and Strapparava 2008; Fernando 2013), etc., but none of them seems to be integrated in the version accessible via the web interface (http://wordnetweb.princeton.edu/perl/webwn). Yet this is the one accessed by the ordinary language user who is generally either not able or willing to spend time to write an algorithm to integrate the different resources.

256

M. Zock

Table 14.2 Comparing two corpora with various inputs Output: WN-x

Output: WiPe

ball

346 hits game, racket, player, court, volley, Wimbledon, championships, inflammation, …, tennis (15th), …

4891 words sport, league, football, hand, food, foot, win, run, game, …, tennis (27th), …

racket

114 hits break, headquarter, gangster, lieutenant, rival, kill, die, ambush, tennis (38th), …

2543 words death, kill, illegal, business, corrupt, …, tennis (72nd), …

ball + racket

11 hits game, tennis, (2nd), …

528 hits sport, strike, tennis (3rd), …

As mentioned already, the weak point is not so much the quantity of the data, as the quality of the index (the relative sparsity of links). Yet, in order to be fair towards WN, one must admit that, had we built our resource differently,—for example, by including in the list of related terms, not only the directly evoked words, i.e. potential target words, but all the words containing the source-word (wine) in their definition (Bordeaux, Retsina, Tokay),—then we would get ‘vintage’, as the term ‘wine’ is contained in its definition (‘vintage’: a season’s yield of ‘wine’ from a vineyard). Note that in such cases even Google works often quite well, but see also (Bilac et al. 2004; El-Kahlout and Oflazer 2004; Dutoit and Nugues 2002). Last but not least, success may vary quite dramatically, depending on the input (quality of the query). As you can see in Table 14.2, WN performs slightly better than WiPe for the words ‘ball’, ‘racket’ and ‘tennis’. Yet, WiPe does not lag much behind; additionally, it contains many other words possibly leading to the target words (“player, racket, court”, ranked, respectively in position 12, 18 and 20). Not being an encyclopedia, WN lacks most of them, though surprisingly, it contains named entities like ‘Seles’ and ‘Graf’, two great female tennis players of the past. Given the respective qualities of WN and WiPe one may well consider integrating the two by relying on a resource like BabelNet (Navigli and Ponzetto 2012).8 This could be done in the future. In the meantime let us take a closer look at WN and its qualities with respect to word look up.

14.4 Under What Condition Is WN Really Good for Consultation? It is a well-kown fact that WN is based on psycholinguistic principles (associations, network, hierarchical structure, …). What is less known though is the fact, that despite 8 http://lcl.uniroma1.it/babelnet/.

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

257

the original motivations of its creators, WN has never been built for consultation. It has been primarily conceived for usage by machines: “WordNet is an online lexical database designed for use under program control.” (Miller 1995, p. 39). This being said, WN can nevertheless be used for consultation, all the more as it is quite good at it under certain circumstances. Remains the question under what conditions WN actually is able to reveal the desired target word. I believe that it can do so perfectly well provided that the following three conditions are met: (a) the author knows the link holding between the source word (input, say ‘dog’) and the target, e.g. ([dog] + synonym  [? target]) → ([target = bitch]); ([dog] + hypernym  [? target]) → ([target = canine]); (b) the input (source word) and the target are direct neighbors in the resource. For example, [seat]-[leg] (meronym), or [talk]-[whisper] (troponym), … (c) the link is part of WN’s database, e.g. ‘hyponym/hypernym’, ‘meronym’, …

14.5 The Framework of a Navigational Tool for the Dictionary of the Future To access a word means basically to reduce the entire set of words stored in the resource (lexicon), to one (target). Obviously, this kind of reduction should be performed quickly and naturally, requiring as little time and effort (minimal number of steps) as possible on the users’ side. Note that this process is knowledge based, meaning that the user may have stored not only the elusive word but also other, somehow related words. This is a very important point, as in case of failure, the dictionary user may now start from any of these connected words. When I wrote that WN is quite successful with regard to word look-up under certain circumstances, I also implied that it is not so good when these conditions are not met. More precisely, this may happen when: (a) the source (input) and the target are only indirectly related, the distance between the two being greater than 1. This would be the case when the target (‘Steffi Graf’) cannot be found directly in reponse to some input (‘tennis player’), but only via an additional step, say, ‘tennis pro’—([tennis player] → [tennis pro])— given as input at the next cycle, in which case it will only then reveal the target.9 (b) the input (‘play’) and the target (‘tennis’) belong to different parts of speech (see ‘tennis problem’, Fellbaum 1998); 9 Note

that the situation described is a potential problem for any association network. Note also that, eventhough Named Entities (NEs) are generally not contained in a lexicon, some of them have made it into WN. This is the case for some famous tennis players, like Steffi Graf. Anyhow, since NEs are also words, the point we are trying to make holds for both. Hence, both can be organized as networks, and whether access is direct or indirect depends on the relative proximity of the input (prime) with respect to the target word.

258

M. Zock

(c) the prime and the target are linked via a syntagmatic association (‘smoke’‘cigar’). Since the majority of relations used by WN connect words from the same part of speech, word access is difficult if the output (target) belongs to a different part of speech than the input (prime)10 ; (d) the user ignores the link, he cannot name it, or the link is not part of WN’s repertory.11 Actually this holds true (at least) for nearly all syntagmatic associations. Let us see how to go beyond this. To this end I present here briefly the principles of the resource within which search takes place, as well as the required navigational aid (categorial tree) to allow authors to find quickly the word they are looking for. Yet, before doing so, let me clarify some differences between hierarchically structured dictionaries and my approach. While lexical ontologists (LO) try to integrate all words of a language into a neat subsumption hierarchy, we try to group them mainly in terms of direct neighborhood. More precisely, we try to build a lexical graph where all words are connected, regardless of whether we can name the link or not. Put differently, we try to build a hybrid association network whose elements (words) are connected via typed and untyped links. Both kinds of links are necessary for filtering, i.e. to ensure that the search space is neither too big (typed links), nor too small (untyped links). Knowledge of the relationship between the source and the target is an obvious asset, as it reduces considerably the search space. Yet, untyped links are a necessary evil: they address the fact that two words evoke each other. Hence, even if we cannot name the link, we should still include the connected word in the list within which search takes place. Otherwise, how can the user find it? Of course, untyped links can cause growth of the search space. Yet, in order to avoid this problem we could group by category the words devoid of a link (Fig. 14.1, step 2). Obviously, this approach yields a quite different network than WN. Hence it will also produce different results than WN for a given input (see Table 14.3). Suppose we started from a broad term like ‘food’. A lexical ontology like WN would produce the entire list of objects referring to ‘food’ (hyponyms), while an association network would only reveal typically evoked words {food, bread, noodles, rice, fish, meat, cook, eat, buy, starving, good, expensive, fork, chopsticks….}. This list contains, of course, a subset of the terms found in a LO (terms referring to ‘food’), but it also contains syntagmatically related words (origine: France; state: hungry, …). Compare the respective results obtained by WN and the Edinburgh Association Thesaurus.12 By taking a look at this second list one can see that it contains not only hyponyms, that is, specific kinds of food (meat, cheese, …), but also syntagmatically related words (cook, good, France, …), i.e. words typically co-occurring with the term ‘food’. Note that our list may lack items like ‘bagles’, ‘cheese’ or ‘olives’. This is 10 This being said, WN does have cross-POS

relations, i.e. “morphosemantic” links holding among semantically similar words: observe (V), observant (Adj) observation (N). 11 For example: ‘well-known_for’, ‘winner_of’, … 12 http://www.eat.rl.ac.uk (see also: http://rali.iro.umontreal.ca/word-associations/query/).

259

Fig. 14.1 Lexical access a four-step process for the user

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

260

M. Zock

Table 14.3 The respective outputs produced by a lexical ontology (here WN) as opposed to an association network, here, the Edinburgh Association Thesaurus (E.A.T) WN: hypernym: solid; part_holonym: nutrient; hyponyms: leftovers, fresh_food, convenience_food, chocolate, baked_goods, loaf, meat, pasta, health_ food, junk_food, breakfast_food, green_goods, green_groceries, coconut, coconut_meat, dika_bread, fish, seafood, butter, yoghourt, cheese, slop E.A.T: at, drink, good, thought, dinner, eating, hunger, salad, again, apple, baby, bacon, bread, breakfast, case, cheese, consumption, cook, firm, fish, France, goo, great, hungry, indian, kitchen, lamb, loot, meal, meat, mix, mouth, noah, nosy, of, pig, please, poison, rotten, sausage, steak, stomach, storage, store, stuff, time, water, yoghurt, yum

quite normal, if ever these words are not strongly associated with our input (food), which does not imply, of course, that we cannot activate or find them. Had we given ‘wine’ or ‘oil’ ‘green’ and ‘Greece’ as input, chances are that ‘cheese’ and ‘olives’ would pop up immediately, while they are burried deep down in the long list of food produced by a LO. Let us return to the problem of word access. Just as orientation in real world requires tools (map, compass) we need something equivalent for locating a word in a lexical resource. While the semantic map defines the territory within which search takes place, the lexical compass guides the user, helping her to reach the goal (target word). Obviously, the terms map and compass are but metaphors, as there are important differences between world maps and lexical graphs (see below) on one hand, and compasses sailors use and the tool an information seeker is relying on (human brain) on the other. The map I have in mind is basically an association network. It is a fully connected graph encoding all directly associated words given some input. This kind of graph has many redundancies, and the links are not necessarily labeled. In this respect it is very different from WN and even more so from the maps we use when traveling in real world. Also, when using a world map the user generally knows more or less precisely the destination or the relative location of the place he is looking for, for example, south of Florence. He may also be able to deduce its approximate location, eventhough she is not able to produce its name (Rome). This does not hold in the case of a user resorting to a lexical resource (map) based on associations. While the user may know the starting point (knowledge available when trying to find the target, the elusive word), he cannot name the destination (target), as if he could, there would be no search problem to begin with. The user is either able to activate the word (in which case the problem is solved), or not. In this latter case all he can do is to rely on available knowledge concerning the target, an assumption I clearly make here. For example, users often know a related word, and they know how it relates to the target: (part of the) meaning, sound, collocational, etc. Knowledge is often fragmentary. Yet, incomplete as it may be, this kind of information may allow us to help them to find the target, guiding him in a reduced, clearly marked search space (details here below). To get back to navigation in real world. In the case of spatial navigation it suffices to know that ‘Rome’ is south of ‘Florence’, which is part of ‘Lazio’, and that it can be

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

261

reached by car in about 2 h. Having this kind of knowledge we could initiate search in the area of ‘Lazio’, since ‘Lazio’ is an area south of ‘Tuscany’, the area containing ‘Florence’. While this strategy works fine in the case of spatial navigation, it will not work with lexical graphs. In this kind of network terms are related in many ways and their strength may vary considerably. Hence, it is reasonable to show a term only if it is above a certain threshold. For example, a term A (Espresso) being connected to term B (coffee) may be shown only if it is sufficiently often evoked by B. Note that eventhough words are organized in terms of neighborhood, the link between them (explicited or not) may be of many other kinds than a spatial relation. In sum, the links connecting words in an associative network are much more diverse than the ones typically found in a lexical ontology. As mentioned already, humans using world maps usually know the name of their destination, whereas people being in the ToT state do not. Yet, even if they did, they would not be able to locate it on the map. Lexical graphs are simply too big to be shown entirely on a small screen.13 In sum, we need a different approach: search must be performed stepwise, taking place in a very confined space, composed of the input and the direct neighbors (directly associated words). It is like a small window moved by the user from one part of the graph to the next. If there are differences between world maps and association networks (lexical graphs), there are also important differences between a conventional compass and our navigational tool. While the former automatically points to the north, letting the user compute the path between his current location and the desired goal (destination, target), the latter (brain) assumes the user to know, the goal, i.e. target word,14 or its direction (even if one does not know its precise location). While the user cannot name the goal—he has only passive knowledge of it,—the system cannot guess it. However it can make valuable suggestions. In other words, eventhough the system can only make suggestions concerning the target or the directions to go (which word to use as input for the next cycle), it is the user who finally decides whether the list contains the target or not, and if so, in what direction to go. He is the only one to know which suggestion corresponds best to the target (the word he has in mind) or which one of them is the most closely connected to it. Of course, the user may go wrong, but as experience shows his intuitions are generally quite good. Before sketching a roadmap concerning the scenario of word access via the stillto-be-built resource (association network), let me quickly provide some background information concerning the users’ knowledge, a critical component in this kind of dialogue.

13 Associative networks contain many redundancies and are potentially endless, since they contain loops. For example, an input, say ‘Rome’ may well appear to be the direct neighbor of one of its outputs, ‘Italy’: ([Rome] → {[capital], [Italy], [city]}); ([Italy] → {[country], [France], [Rome]}). 14 It has been shown over and over again that people being in the ToT state are able to identify immediately, and without making any mistakes the target word if it is shown to them, eventhough they could not name it. This is passive knowledge.

262

M. Zock

14.6 Navigation, a Fundamentally Cognitive Process As I will show in this section, navigation in a lexical resource is above all a knowledgebased process. Before being able to access a word, we must have acquired it. It is only then that it has become part of our knowledge. Yet, storage does not guarantee access (Zock and Schwab 2011). This fact has not received the attention it deserves by lexicographers. Note also that there are several kinds of knowledge: declarative, meta-knowledge (not necessarily linguistic) and knowledge states. • Declarative knowledge is what we acquire when learning words (meaning, form, spelling, usage), and this is the information generally encoded in dictionaries. Obviously, in order to find a word or to find the information associated with it, they must be stored, though this is not enough. • Next, there is meta-knowledge, which also needs to be acquired. Being generally unavailable for in(tro)spection, meta-knowledge reveals itself in various ways. For example, via the information available when we fail to access a word (Schwartz 2006), or via the query we provide at the moment of launching a search. As word association experiments have shown (Aitchison 2003) words always evoke something. Since this is true for all words one can conclude that all words are connected in our mind, which implies that all words are accessible from anywhere like in a fully connected graph.15 All we have to do is to provide some input (source word, available information) and follow then the path linking this input to the output (target). Interestingly, people hardly ever start from words remotely related to the target. Quite to the contrary, the words they give at the input (source words) tend to be more or less direct neighbors of the target, requiring generally only one or two steps for the solution, that is, they are hardly ever further away than the distance of two (steps).16 Also, dictionary users often know the type of relationship holding between the input (prime) and the target, otherwise, why would lexicographers build thesauri, synonym-or collocation dictionaries? All these observations lend support to our intuition that people have a considerable amount of (meta-) knowledge concerning the organization of words in their mind, i.e. their mental lexicon. The notion of relationship has been nicely exploited by WN, which due to this feature keeps the search space, i.e. a set of candidates among which the user has to choose, quite small. The idea of relatedness has led lexicographers to build thesauri, collocation- and synonym dictionaries. Obviously an input consisting only of a 15 Note that this does not hold for WN, as WN is not a single network, but a set of networks. There are 25 for nouns, and at least one for all the other parts of speech. 16 This is probably one of the reasons why we would feel estranged if someone provided as cue ‘computer’, while his target were ‘mocha’. The two are definitely not directly connected, though, there is a path between them, eventhough it is not obvious (The chosen elements are always underlined.): computer → (Java, Perl, Prolog; mouse, printer; Mac, PC); (1) Java → (island, programming language); (2) Java (island) → (coffee; Kawa Igen); (3) coffee → (cappucino, mocha, latte). Note that ‘Java’ could activate ‘Java beans’, a notion inherent to java, the programming language. In this case it would lead the user directly to the class (hypernym) containing the desired target word (mocha).

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

263

simple word is hard to interpret. Is the user looking for a more general/specific word, a synonym or antonym? Is the input semantically or phonetically related to the target, or is it part of the target word’s definition (dog-animal)? In each case the user is expecting a different word (or set of words) as output. Hence, in order to enable a system to properly interpret the users’ goals we need this kind of metalinguistic information (neighbor of the target, i.e. source word + relation to the target) at the input.17 If ever the user cannot provide it, the system is condemned to make a rough guess, presenting all directly connected words. Obviously, such a list can become quite large. This being so, it makes sense to provide the system with this kind of information, so that it can produce then the right set of words, while keeping the search space small. • Knowledge states, refer to the knowledge activated at a given point in time, for example, when launching a search. What has been primed? What is available in the user’s mind? Not all information stored in our mind is equally available or prominent anytime. The fact that peoples’ knowledge states vary is important, as it co-determines the way a user proceeds in order to find the information he is looking for. This being so, it is important to be taken into consideration by the system designer. In conclusion, all this knowledge must be taken into account as it allows us to determine the search space, reducing its scope, which otherwise is the entire lexicon. The example here below illustrates to some extent these facts with regard to wordfinding in an electronic resource. Suppose you are looking for a word conveying the idea of a large black-and-white herbivorous mammal of China. Yet, for some reason you fail to retrieve the intended form panda, even though you do know a lot concerning the target. People being in this state, called the ToT-problem, would definitely appreciate if the information they are able to access could be used to help them find the target. Figure 14.1 illustrates the process of getting from a visual stimulus to its corresponding linguistic output (word, expression) via a lexical resource. Given an external stimulus (A) our brain activates a set of features (B) that ideally allow us to retrieve the target form. If our brain fails, we use a fallback strategy and give part of the activated information to a lexical resource (C) expecting it to filter its base (D) in the hope to find the target (panda) or a somehow related word (E). As one can see, we consider look-up basically as a two-step process. At step one the user provides some input (current knowledge) to which the system answers with a set of candidates, at step two the user scans this list to make her choice (Table 14.4).

17 This has of course consequences with respect to the resource. To be able to satisfy the different user

needs (goals, stratgies) we probably need to create different databases: Obviously, to find a target on the basis of sound (rhymes), meanings (meaning-fragments) or related words (co-occurrences), requires networks encoding different kinds of information.

264

M. Zock

Table 14.4 Lexical access a two-step process mediated by the brain and an external resource (lexicon) A: Perceptual B: Associated C: Input to D: Lexical E: Output of input, i.e. target features in the lexical resource resource lexical resource mental lexicon (brain) Type: bear Lives_in: China Features: black patches Diet: eats bamboo

bear China

aardvark … … panda … … theorem … zygote

panda polar bear

14.7 The Roadmap Since alphabetically organized dictionaries are not very useful for onomasiological search (language production), we follow WN by organizing words in terms of neighborhood. All words are connected, and if possible, the links are given names. This being so, we have a map supporting navigation. The user can enter the graph at any point, to follow the links until having reached the target. Obviously, the kind of links, as well as the presence/absence of link names (see below) are important with respect to the search space (see below). A network devoid of link names yields many more hits for a given input than a network containing the same set of words but whose links are named. Imagine the number of possible outputs for [‘tree’] compared to [(‘tree’) + (‘synonym’/‘hypernym’]. Let us now see quickly how to make all this work. Imagine an author wishing to convey the name of a beverage commonly found in coffee shops (target: ‘mocha’). Failing to do so, he reaches for a lexicon. Since dictionaries are too huge to be scanned from cover (letter A) to cover (Z), I suggest to create a dialog between the user and the computer to reduce incrementally the search space. The user provides the input,18 —word coming to his/her mind (source) when trying to access the target,—and the system produces as output a list of potential target words, in the hope to enable the user to find the elusive word. Note that, concerning the source and the target, there are basically three cases: (A) the two are directly related, and the user knows their relationship; (B) the user knows a direct neighbor, but he ignores the name of their relationship; (C) the source and the target are not directly related, they are only indirect neighbors. Since the first case (A) is quite well handled by WN, I will illustrate here only ‘B’, ignoring the last case (C), as it can be solved indirectly by applying recursively the procedures

18 This latter can be a single word—‘coffee’ in the case of target ‘mocha’—or a set of words, which in a normal communicative setting would yield a sentence, where the information seeker asks someone else to help him to find the elusive word.

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

265

proposed in ‘A’ and ‘B’.19 To convey as simply as possible the rationale underlying my approach let us make the following assumptions: (A) the user’s input is a single word, here ‘coffee’ (step-1, Fig. 14.1); (B) the target, i.e. the elusive form, is ‘mocha’; (C) the two are directly related in the resource in which search takes places; (D) the nature of their relationship has not been specified by the user.20 Because of this last point—the relationship between the source and the target not being given with the input—search space may grow considerably. To avoid this problem I propose to present in a clustered and labeled form (categorial tree) all direct associates of the input (step-2, Fig. 14.1).21 The user navigates in this tree, deciding on the category within which to search, hoping to find the target, and if he cannot find it in any of them, in what direction to go. If he could find the target, search stops, otherwise the user will pick one of the associated terms or provide an entirely new word, and the whole process iterates. The system will then respond with a new set of proposals. Two points, one concerning step-1, the other step-2. Ideally, the searchspace determined at step-1 should contain the target word. As we have seen in the experiment described here above (Sect. 14.3), WN failed to reveal a candidate though it was stored it in its database. This failure was due to a lack of syntagmatic associations, that is, it was a side-effect of a design choice of how to connect words, or, which words to connect. Hence, the search space proposed by WN in response to some input was too small. At least it did not contain the target while its competitor (WiPe) did. Obviously, this is something we would like to avoid. The categories of our tree (step-2) resemble somehow those of Roget’s thesaurus. Actually, they are not quite the same, but this is not really the point. What I would like to stress is the fact that both the categories and the words composing the search-space are dynamically computed in our case, while they are frozen, i.e. determined once and for all in Roget. Hence, the set of words (search space) to be presented in the categorial tree will be considerably smaller in our case than the ones displayed by Roget. I believe that this kind of flexibility is a desirable feature as we cannot predict the user’s input or his goals, we can only comply with them. 19 This kind of wording can be generalized to a pattern for asking the following question: “What is the word for ‘[X] that [Y]?”, where [X] is usually a hypernym and [Y] a stereotypical, possibly partial functional/relational/case description (action) of the target word. A similar pattern could be used for namefinding. For example, asking “What is the name of the of ?” could yield ‘Pizarro’ or ‘Cortés’, depending on the value of the empire (Inca/Aztec). As one can see, the processes underlying wordfinding and namefinding are not very different. 20 Note, that in order to determine properly the initial search space (step-1), we must have already well understood the input [mouse1 /mouse2 (rodent/device)], as otherwise our list will contain a lot of noise, presenting ‘cat, cheese’ together with ‘computer, mouse pad’ {cat, cheese, computer, mouse pad}, which is not quite what we want, since some of these candidates are irrelevant, i.e. beyond the scope of the user’s goal. 21 This labeling is obligatory to allow for realistic navigation, as the list produced in response to the input may be very long and the words being of the same kind may be far apart from each other in the list. Hence it makes sense to structure words into groups by giving them appropriate (i.e. understandable) names so that the user, rather than looking up the entire list of words, searches only within a specific bag labeled by a category.

266

M. Zock

To get back to our roadmap. As one can see, the proposed method is quite straightforward, reducing considerably time and space needed for navigation and search. Suppose that you had to locate a word in a resource of 50,000 words. If your input triggered 100 direct associates, one of them being the target, then we would have reduced in a single step the search space by 99.8%, limiting navigation and search to a very small list. Suppose that our hundred words were evenly spread over 5 groups, than search would consist in spotting the target in a list of 25 items: 5 being category names and 20 being words within the chosen group. A small note concerning the 2nd step. Step-2 yields a tree whose leaves are potential target words and whose nodes are categories, which, while being also words, are not at all the goal of the search. They are only the means to reach the goal. Put differently, their function is orientational, guiding the user during his search. Words at the leave-level are potential target words, while the ones at the intermediate level (category names; preterminal nodes) are meant to reduce the number of words among which to perform search, and to help the user to decide on the direction to go. Hence, category names are reductionist and orientational (signposts), grouping terminal nodes into a bag, signaling via their name not only the bag’s content, but also the direction to go. While the system knows the content of a bag, it is only the user who can decide which of the bags is likely to contain the elusive word. Because, eventhough he cannot name the target, he is the only one to know the target, be it only passively and in fairly abstract terms. This is where the categoy names have their role to play. In sum, it is not the system that decides on the direction to go next, but the user. Seeing the names of the categories she can make reasonable guesses concerning their content. In conclusion, categories act somehow like signposts signaling the user the kind of words he is likely to find choosing one bag rather than another. Indeed, knowing the name of a category (fruit, animal), the user can guess the kind of words contained in each bag (kiwi vs. crocodile). Assuming that the user knows the category of the searched word,22 she should be able to look in the right bag and take the best turn. Navigating in a categorial tree, the user can search at a fairly high level (class) rather than at the level of words (instances). This reduces not only the cognitive load, but it increases also chances of finding the target, while speeding up search, i.e. the time needed to find a word. While step-1 is mainly a matter of relatedness (‘wine’ and ‘red’ being different in nature, they are nevertheless somehow related), step-2 deals with similarity: there are more commonalities between ‘dogs’ and ‘cats’ than between ‘dogs’ and ‘trees’. Put differently, the first two terms are more similar in kind than the last two. The solution of the second step is certainly more of a challenge than the one of step-1 which is largely solved, eventhough there is an issue of relevance: not all co-occurences

22 A fact which has been systematically observed for people being in the ToT state who may tell the listener that they are looking for the name of a “fruit typically found in a ”, say, New Zealand, in order to get ‘kiwi’.

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

267

are really useful.23 To put words into clusters is one thing, to give them names an ordinary dictionary user can understand is quite another.24 Yet, arguably building this categorial tree is a crucial step, as it allows the user to navigate on this basis. Of course, one could question the very need of labels, and perhaps this is not too much of an issue if we have only say, 3–4 categories. We are nevertheless strongly convinced that the problem is real, as soon as the number of categories (hence the words to be classified) grows. To conclude, I believe it is fair to say that the 1st stage seems to within reach, while the automatic construction of the categorical tree remains a true challenge despite some existing tools (word2vec) and the vast literature devoted to this specific or otherwise strongly related problems (Zhang et al. 2012; Biemann 2012; Everitt et al. 2011). One last point: to be truly useful, the user should provide as input not only a word, but also a clue concerning the relationship between this input and his goal (target word). Does he look for a semantically, formally (sound) or otherwise related word with respect to the input? Since inputs can be interpreted in many ways, we need additional information. Given some input, what is the user’s goal? Is he looking for a synonym, hypernym or a similarly sounding word? Obviously, different goals yield different searchspaces. This is a very important point. Authors searching for a sound-related word to ‘right’ expect a different set of candidates (write, wright, rite), from authors looking for its antonym (‘wrong’). Table 14.5 shows some possible links between some input (prime) and its directly associated output (possible target). WN takes advantage of this fact eventhough only a subset of the links mentioned here below are actually implemented. Note that these links are of different sort: some are conceptual (1–13), others are formal, i.e. they concern linguistic forms (15–19), and what we dubbed ‘free association’ (14) concerns both. For more details concerning links or relations, take a look at (Evens 2009; Green et al. 2002; Miller and Fellbaum 1992; Murphy 2003; Nastase et al. 2013; Storjohann 2010).

14.8 Conclusion Obviously, the human brain is a complex object and so is the process of accessing words in the mental lexicon. My goal was not so much to address the problem of complexity, i.e. the topology of the map of the mental lexicon. My goal was rather to describe a method, that, once implemented, should help people to overcome the 23 Take for example the Wikipedia page devoted to ‘panda’, and check which of the co-occurrences are those typically evoked when looking for this particular lexical concept. 24 For example, while the sequence of hypernyms listed by WN for horse captures much of the phylogenetic detail a biologist would want to see recorded (horse → equine → odd-toed ungulate → ungulate → placental mammal → mammal → vertebrate → chordate → animal → organism → entity), most of these terms mean next to nothing to an ordinary dictionary user.

268

M. Zock

Table 14.5 Possible links or associations between an input (cue/prime) and the target Type of relation Description of the relation Prime-target 1

Hypernym

A more general word

pie-pastry

2

Hyponym

A more specific word

fruit-nut

3a

Meronym_substance

A concept being a substance of another concept blood-body

3b

Meronym_part_of

A concept being part of another concept

3c

Meronym_member_of A concept being a member of another concept

4a

Holonym_substance

A concept having another concept as substance sea-salt

4b

Holonym_part_of

A concept having another concept as part

4c

Holonym_member_of A concept having another concept as member

team-player

5

Cause to

A verb expressing the cause of a result

kill-die

6

Entailment

A verb expressing an unavoidable result

buy-have

7

Troponym

A specific way to perform an action

drink-sip

8

Part_of_meaning

Part of the target word‘s definition

butter-milk

ship-fleet kid-family tree-leave

9

Quality

Typical quality, or inherent feature

snow-cold

10

Co-occurrence

Two concepts occurring frequently together

blue-sky

11

Topically related

Two concepts related by topic

sea-tide

12

Used_for

Instrumentally related words

fork-eating

13

Made_of

Substance or element used to make

glass-sand

14

Free association

Can be any kind of link between two words

door-open

15

Synonym

Word expressing basically the same meaning

cup-mug

16

Antonym

A word meaning the opposite

dry-wet

17

Sound (rhyme)

Two similar sounding words

bad-mad/sad

18

Homophones

Words sounding alike, but spelled differently

right-write

19

Anagrams

Composed of same or similar components

cheater-teacher

ToT-problem. The method is radically knowledge-based, that is to say, it takes into account knowledge users may have at the onset of consultation (see below). I have started the paper by observing that word access remains a problem for dictionary builders (Thumb 2004) and users alike, in particular for those being in the production mode (Zock 2015a, b; Zock and Tesfaye 2015). Next I have shown that word storage does not guarantee its access, even if the target is stored in a computer. I have then analyzed some of the reasons why even a psycholinguistically motivated resource like WN often fails to reveal the word authors are looking for. Finally, I have recasted the problem within a cognitive framework, presenting a roadmap of how to overcome the ToT-problem. The idea is to build a navigational tool (hybrid association network) to help humans to find the word they are looking for. The user provides the information coming to his mind when failing to access the target word (input), and the resource produces a list of potential target words (output). If the user can provide a (direct) neighbor of the target and its link, the

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

269

answer is generally straightforward. In the opposite case I suggest to present the candidates in a labeled cluster-form (categorial tree) rather than as a huge, flat list. While the system‘s task is search-space reduction in step-1 (Fig. 14.1), its function in step-2 (building of the categorial tree) is to support navigation. Just as it is unreasonable to perform search in the entire lexicon, is it cumbersome to drill down huge lists. This is why I suggested to cluster and label the outputs produced in response to the query. After all, we want users to find the target quickly and naturally, rather than drown them under a huge, unstructured (or poorly structured) list of words. Note that there is at least one study supporting the idea that knowledge of link names is useful. Nikolova et al. (2010) could show that word-finding is enhanced when wordnets contain syntagmatic links. They describe a study where people struck by aphasia used their resource, showing that retrieval was significantly better in this case than when relying on a resource devoid of this information. They conceded though that finding the first word to start communication with was still a problem. For other related work see (Ferret 2015; Zock and Biemann 2016). One last point: the success of the (yet-to-be-built) resource hinges critically on three kinds of knowledge: (a) factual knowledge: to find a word it must exist, i.e. it must be stored; (b) metaknowledge: to allow for word access, words must also be wellorganized, and the user must have some knowledge concerning this organization. This amounts to knowing at least some of the words connected to the target, and the relationship between some input (currently available word) and the goal (target word). Put differently, in order to be able to provide a decent input (typically a relatively close neighbor of the target), one must have at least a local view of the organization of the mental lexicon; (c) cognitive states: they are revealed by the word(s) coming to our mind when we search for a (word) form that we know, but cannot access. Solving this problem is what this paper has been about. So far this is only a concept, but I hope to be able to provide one day evidence of its feasibility, as having such a tool would be extremely precious for dictionary users being in the production mode.

References Abrams, L., Trunk, D. L., & Margolin, S. J. (2007). Resolving tip-of-the-tongue states in young and older adults: The role of phonology. In L. O. Randal (Ed.), Aging and the elderly: Psychology, sociology, and health (pp. 1–41). Hauppauge, NY: Nova Science Publishers Inc. Abel, A. (2012). Dictionary writing systems and beyond. In S. Granger, & M. Paquot (Eds.), Electronic lexicography (pp. 83–106). Oxford: Oxford University Press. Agirre, E., Ansa, O., Hovy, E., & Martinez, D. (2001). Enriching WordNet concepts with topic signatures. Retrieved from http://arxiv.org/abs/cs.CL/0109031. Aitchison, J. (2003). Words in the mind: An introduction to the mental lexicon. Oxford: Blackwell. Benson, M., Benson, E., & Ilson, R. (2010). The BBI combinatory dictionary of English. Philadelphia: John Benjamins. Bentivogli, L., & Pianta, E. (2004). Extending WordNet with syntagmatic information. In P. Sojka, K. Pala, P. Smrz, C. Fellbaum, & P. Vossen (Eds.), GlobalWor(l)dNet Conference, Proceedings (pp. 47–53). Brno: Masaryk University.

270

M. Zock

Benjamin, M. (2014). Collaboration in the production of a massively multilingual lexicon. In LREC Conference Proceedings (pp. 211–215). Reykjavik. Bernstein, T. (1975). Bernstein’s reverse dictionary. New York: Crown. Biemann, C. (2012). Structure discovery in natural language. Berlin: Springer. Bilac, S., Watanabe, W., Hashimoto, T., Tokunaga, T., & Tanaka, H. (2004). Dictionary search based on the target word description. In Proceedings of the Tenth Annual Meeting of The Association for Natural Language Processing (pp. 556–559). Tokyo. Boissière, P. (1862). Dictionnaire analogique de la langue française: Répertoire complet des mots par les idées et des idées par les mots. Paris: Auguste Boyer. Boyd-Graber, J., Fellbaum, C., Osherson, D., & Schapire, R. (2006). Adding dense, weighted, connections to WordNet. In P. Sojka, Ks. Choi, C. Fellbaum, & P. Vossen (Eds.), Proceedings of the Global WordNet Conference 2006 (pp. 29–35). Brno: Masaryk University. Brown, A. S. (1991). The tip of the tongue experience: A review and evaluation. Psychological Bulletin, 10, 204–223. Brown, R., & Mc Neill, D. (1966). The tip of the tongue phenomenon. Journal of Verbal Learning and Verbal Behaviour, 5, 325–337. Casares, J. (1942). Diccionario ideológico de la lengua española. Barcelona: Gustavo Gili. de Deyne, S., & Storms, G. (2015). Word associations. In J. R. Taylor (Ed.), The Oxford handbook of the word. Oxford, UK: Oxford University Press. de Deyne, S., Verheyen, S., & Storms, G. (2016). Structure and organization of the mental lexicon: A network approach derived from syntactic dependency relations and word associations. In Towards a theoretical framework for analyzing complex linguistic networks (pp. 47–79). Berlin: Springer. Dong, Z., & Dong, Q. (2006). HowNet and the computation of meaning. London: World Scientific. Dornseiff, F., Wiegand, H. E., & Quasthoff, U. (2004). Der deutsche Wortschatz nach Sachgruppen. Berlin/New York: de Gruyter. Dutoit, D., & Nugues, P. (2002). A lexical network and an algorithm to find words from definitions. In F. van Harmelen (Ed.), Proceedings of the 15th European Conference on Artificial Intelligence (pp. 450–454). Amsterdam: IOS Press. Edmonds, D. (Ed.). (1999). The Oxford reverse dictionary. Oxford: Oxford University Press. El-Kahlout, I. D., & Oflazer, K. (2004). Use of WordNet for retrieving words from their meanings. In P. Sojka, K. Pala, P. Smrž, C. Fellbaum, & P. Vossen (Eds.), Proceedings of the Global Wordnet Conference (pp. 118–123). Brno: Masaryk University. Evens, M. W. (2009). Relational models of the lexicon: Representing knowledge in semantic networks. Cambridge: Cambridge University Press. Everitt, B., Landau, S., Leese, M., & Stahl, D. (2011). Cluster analysis. Chichester: John Wiley & Sons, Ltd. Fellbaum, C. (Ed.). (1998). WordNet: An electronic lexical database and some of its applications. Cambridge: MIT Press. Fergusson, R., & Fergusson, R. (1985). The Penguin rhyming dictionary. London: Penguin. Fernando, S. (2013). Enriching lexical knowledge bases with encyclopedic relations (Doctoral dissertation). University of Sheffield. Ferret, O. (2015). Typing relations in distributional thesauri. In N. Gala, R. Rapp & G. Bel-Enguix, (Eds.), Language Production, Cognition, and the Lexicon. Springer, 113–134. Fontenelle, T. (1997). Turning a bilingual dictionary into a lexical-semantic database. Tübingen: Max Niemeyer. Gliozzo, A., & Strapparava, C. (2008). Semantic domains in computational linguistics. Berlin: Springer. Granger, S., & Paquot, M. (Eds.). (2012). Electronic lexicography. Oxford: Oxford University Press. Green, R., Bean, C. A., & Myaeng, S. H. (2002). The semantics of relationships. Dordrecht: Kluwer. Hanks, P. (2012). Corpus evidence and electronic lexicography. In S. Granger & M. Paquot, (Eds.), Electronic lexicography (pp. 57–82). Oxford: Oxford University Press. Kahn, J. (1989). Reader’s Digest reverse dictionary. London: Reader’s Digest.

14 Eureka! A Simple Solution to the Complex ‘Tip-of-the-Tongue’-Problem

271

Kilgarriff, A., Rychlý, P., Smrž, P., & Tugwell, D. (2004). The sketch engine. In G. Williams & S. Vessier (Eds.), Proceedings of the Eleventh EURALEX International Congress (pp. 105–116). Lorient, France: UBS. Kilgarriff, A., & Kosem, I. (2012). Corpus tools for lexicographers. In S. Granger & M. Paquot (Eds.), Electronic Lexicography (pp. 31–56). Oxford: Oxford University Press. Levelt, W., Roelofs, A., & Meyer, A. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22, 1–75. Levelt, W. J. (1993). Speaking: From intention to articulation. Cambridge: MIT Press. Levelt, W. J. (2001). Spoken word production: A theory of lexical access. Proceedings of the National Academy of Sciences, 98(23), 13464–13471. Mel’ˇcuk, I., & Polguère, A. (2007). Lexique actif du français: l’apprentissage du vocabulaire fondé sur 20 000 dérivations sémantiques et collocations du français. Champs linguistiques. Bruxelles: De Boeck. Mihalcea, R., & Moldovan, D. (2001). Extended WordNet: Progress report. In NAACL 2001—Workshop on WordNet and Other Lexical Resources (pp. 95–100). Pittsburgh, USA. Miller, G. (1991). The science of words. Scientific American Library. New York: W H Freeman & Co. Miller, G. A. (1995). WordNet: A lexical database for English. Communications of the ACM, 38(11), 39–41. Miller, G. A., Beckwith, R., Fellbaum, C., Gross, D., & Miller, K. J. (1990). Introduction to WordNet: An on-line lexical database. International Journal of Lexicography, 3(4), 235–244. Miller, G. A., & Fellbaum, C. (1992). Semantic networks of English. In B. Levin & S. Pinker (Eds.), Lexical and conceptual semantics (pp. 197–229). Cambridge and Oxford, England: Blackwell. Moerdijk, F. (2008). Frames and semagrams. Meaning description in the General Dutch Dictionary. In Proceedings of the Thirteenth Euralex International Congress (pp. 561–570). Barcelona: EURALEX. Murphy, M. L. (2003). Semantic relations and the lexicon: Antonymy, synonymy and other paradigms. Cambridge: Cambridge University Press. Nastase, V., Nakov, P., Seaghdha, D. O., & Szpakowicz, S. (2013). Semantic relations between nominals. Synthesis Lectures on Human Language Technologies, 6(1), 1–119. Navigli, R., & Ponzetto, S. (2012). BabelNet: The automatic construction, evaluation and application of a wide-coverage multilingual semantic network. Artificial Intelligence, 193, 217–250. Nikolova, S., Tremaine, M., & Cook, P. R. (2010). Click on bake to get cookies: Guiding wordfinding with semantic associations. In Proceedings of the 12th International. ACM SIGACCESS Conference on Computers and Accessibility (pp. 155–162). New York: ACM. Rapp, B., & Goldrick, M. (2006). Speaking words: Contributions of cognitive neuropsychological research. Cognitive Neuropsychology, 23(1), 39–73. Richardson, S., Dolan, W., & Vanderwende, L. (1998). Mindnet: Acquiring and structuring semantic information from text. In Proceedings of the 17th international conference on Computational linguistics, ACL-COLING’98 (pp. 1098–1102). Montréal. Robert, P., Rey, A., & Rey-Debove, J. (1993). Dictionnaire alphabetique et analogique de la Langue Française. Paris: Le Robert. Roget, P. (1852). Thesaurus of English words and phrases. London: Longman. Rundell, M., & Fox, G. (Eds.). (2002). Macmillan English dictionary for advanced learners. Oxford: Macmillan. Schvaneveldt, R. (Ed.). (1989). Pathfinder associative networks: Studies in knowledge organization. Norwood, New Jersey, US: Ablex. Schwartz, B. L. (2002). Tip-of-the-tongue states: Phenomenology, mechanism, and lexical. Mahwah, NJ: Lawrence Erlbaum Associates. Schwartz, B. L. (2006) Tip-of-the-tongue states as metacognition. Metacognition and Learning. 1(2), 149–158. Segalowitz, N. (2000). Automaticity and attentional skill in fluent performance. In H. Riggenbach (Ed.), Perspectives on fluency (pp. 200–219). Ann Arbor, MI: University of Michigan Press.

272

M. Zock

Sierra, G. (2000). The onomasiological dictionary: A gap in lexicography. In U. Heid, S. Evert, E. Lehmann, & C. Rohrer (Eds.), Proceedings of the Ninth Euralex International Congress (pp. 223–235). Stuttgart: IMS, Universität Stuttgart. Storjohann, P. (Ed.). (2010). Lexical-semantic relations: Theoretical and practical perspectives. Amsterdam: John Benjamins Publishing. Summers, D. (1993). Language Activator: The world’s first production dictionary. London: Longman. Thumb, J. (2004). Dictionary look-up strategies and the bilingualised learner’s dictionary. A thinkaloud study. Tübingen: Max Niemeyer Verlag. Tulving, E., & Pearlstone, Z. (1966). Availability versus accessibility of information in memory for words. Journal of Verbal Learning and Verbal Behavior, 5, 381–391. van Sterkenburg, P. (2003). Onomasiological specifications and a concise history of onomasiological dictionaries. In P. van Sterkenburg (Ed.), A practical guide to lexicography (pp. 127–143). Amsterdam: John Benjamins Publishing. Webster, M. (2007). Merriam Webster’s rhyming dictionary. Merriam-Webster, Inc. Springfield, Massachusetts. Zhang, Z., Gentile, A., & Ciravegna, F. (2012). Recent advances in methods of lexical semantic relatedness – A survey. Journal of Natural Language Engineering, 19(4), 411–479. Zock, M. (2015a). ‘Errare humanum est’. Refusing to ‘appreciate’ this fact could be a big mistake! In G. Adda, M. Adda-Decker, J. Mariani, V. Barbu Mititelu, D. Tufis, & I. Vasilescu (Eds.), Errors by Humans and Machines in multimedia, multimodal and multilingual data processing. Proceedings of ERRARE 2015. Bucharest: Romanian Academy Publishing House. Zock, M. (2015b). Introduction to the special issue of ‘cognitive aspects of natural language processing’ (Words in books, computers and the human mind). Journal of Cognitive Science, 16(4), 355–378. Institute for Cognitive Science, Seoul National University (http://j-cs.org/gnuboard/ bbs/board.php?bo_table=__vol016i4). Zock, M., & Biemann, C. (2016). Towards a resource based on users’ knowledge to overcome the Tip-of-the-Tongue problem. In Proceedings of the COLING Workshop ’Cognitive Aspects of the Lexicon’ (CogALex-V) (pp. 57–68) Osaka, Japan. Zock, M., Ferret, O., & Schwab, D. (2010). Deliberate word access: An intuition, a roadmap and some preliminary empirical results. International Journal of Speech Technology, 13(4), 107–117. Zock, M., & Schwab, D. (2011). Storage does not guarantee access. The problem of organizing and accessing words in a speaker’s lexicon. Journal of Cognitive Science, 12(3), 233–258. Institute for Cognitive Science, Seoul National University. Zock, M., & Tesfaye, D. (2015). Automatic creation of a semantic network encoding part_of relations. Journal of Cognitive Science, 16(4), 431–491. Institute for Cognitive Science, Seoul National University. Zock, M., Wandmacher, T., & Ovchinnikova, E. (2009). Are vector-based approaches a feasible solution to the « tip-of-the-tongue » problem? S. Granger & M. Paquot (Eds.), eLexicography in the 21st century: New challenges, new applications (pp. 355–366). Louvain-la-Neuve.

Chapter 15

The Emergence of Hubs in Complex Syntactic Networks and the DP Hypothesis: The Relevance of a Linguistic Analysis Lluís Barceló-Coblijn, Maia Duguine and Aritz Irurtzun Abstract A series of analyses of linguistic corpora of L1 acquisition of different languages (Catalan, Basque, Dutch, Italian, German, French, Spanish and English) showed that functional words emerge as hubs of the network. Such emergence always takes place late and abruptly, in coincidence with that of a new topology of the network: the small-world network. This kind of network is the middle stage between completely random and completely regular networks. In earlier analyses, determiners stood out among functional words as hubs in all languages, regardless of their linguistic phylogeny. But there are two different ways to analyze the syntactic relationship between determiners and nouns: determiners could either be “governors” or “dependents” of nouns. Here we explore the two possible analyses and argue that the first one should be preferred over the second one, in line with contemporary syntactic theorizing.

15.1 Introduction Psycholinguistics offers several tools to examine the syntactic ability of an individual: for example, some techniques make the participants repeat a series of sentences; or the researcher may ask the parents about the child’s syntactic ability with the help of some model such as the Words and Sentences Scale of the MCD (Fenson et al. 2007). A classical metric is the so-called Mean Length of Utterance (MLU), based on countL. Barceló-Coblijn (B) Department of Catalan Philology and General Linguistics, University of the Balearic Islands, Carretera de Valldemossa, km 7.5, 07122 Palma, Spain e-mail: [email protected] M. Duguine · A. Irurtzun Centre National de la Recherche Scientifique, IKER (UMR 5478), Château Neuf, 15 Place Paul Bert, 64100 Bayonne, France e-mail: [email protected] A. Irurtzun e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_15

273

274

L. Barceló-Coblijn et al.

ing the words of 100 sentences produced by the child and dividing the total number by 100. This method provides a MLU coefficient. A more sophisticated tool is the so called Index of Productive Syntax (IPS), which is used to track syntactic development (Scarborough 1990). Importantly, IPS is implicitly committed to a linear view of the ontogeny of syntax, but seems to address some weaknesses of the MLU. The IPS score is obtained from a corpus of 100 utterances, within which 56 specific language structures must be found. IPS has then evolved into a new tool (Sagae et al. 2010) by combining it with NLP techniques. In particular the notion of structural dependency has been introduced to identify grammatical relationships (Hudson 1990). In science a particular approach is always adapted to a level: either micro-, meso-, or macroscopic views are possible, though a particular tool tends to be optimal for a particular level. Thus, from the above mentioned methodologies it is clear that there is a variety of tools today that could help pointing at the syntactic development of a speaker. Crucially, these methodologies focus on the level of morphemes (sometimes words)1 or on particular syntactic structures—up to the sentence level. We identify the first within the microscopic scale (morphemes and to some extent “words”) and the latter within the mesoscopic scale (phrases and sentences). In this work, we want to test the more advanced technique of complex networks against two purely theoretical linguistic hypotheses. The technique of syntactic complex networks2 has been developed in a relatively recent series of works (CorominasMurtra et al. 2009; Barceló-Coblijn et al. 2012; Barceló-Coblijn et al., submitted) and the present article represents a further step in refining this technique, putting to the test a particular linguistic hypothesis. The structure of the article is the following. The second section summarizes the main hits of the development of this technique, pointing out a particular fact of syntactic networks: the emergence of hubs and the lexical categories that tend to become hubs in typically developing (TD) conditions. When children develop their syntactic capacity to combine words they undergo three phases. In the second and third phases hubs emerge abruptly and most of them represent functional words: determiners, adpositions, etc. This is examined in the third section, raising the following question: since scientists model the network on the basis of linguistic theory, does linguistic theory influence network creation in such a way that two competing linguistic hypotheses would produce two qualitatively different results? The paradigmatic case of the Determiner Phrase (DP) hypothesis (cf. i.a. Abney 1987; Longobardi 1994) is explained in contrast to the more traditional Noun Phrase (NP) hypothesis: the first would imply that determiners have two edges, while with the latter, the determiner would only have one edge, a consequence of its postulated dependence with respect 1 We

would like to note that the term «word» is highly ambiguous and therefore not really suitable as a term for a unit. The boundaries of «words» can vary dramatically from one language to another (e.g., Catalan vs. Basque, or Frisian vs. West Greenlandic) and from one module to the other (cf. words in morphology vs. in phonology). 2 The term c-network is used in order to make clear the radical difference between the networks with which we work here and other networks used in cognitive science such as Neural Networks or Connectionism. Complex networks have no relation whatsoever with the networks developed by connectionists, the word «network» being the only point in common.

15 The Emergence of Hubs in Complex Syntactic Networks …

275

to the noun, the head of the NP. These two hypotheses are put to the test by confronting the results of the hypotheses about the structure of the sentences that will conform the network. As it is explained in the fourth section, the procedure of these analyses is grounded on the Netlang program (Barceló-Coblijn et al. 2017), a brand new software specifically developed for this technique. The present publication represents the first in which this software has been put to use. The final section discusses the results of the analysis, along with some arguments supporting the adoption of the DP hypothesis in c-network analyses.

15.2 Syntactic Complex Networks and Cognitive Science The use of networks in the study of linguistic behavior is not new. One can find already in the sixties applications of complex networks to English language. But network science underwent an important twist when Watts and Strogatz (1998) proved that within the wide range of networks there are those that are completely regular at one end, those that are completely random at the other end, and between these two extremes, there is the class of scale-free networks. The latter is characterized by an average number of edges per node ranging from 1 to 3, with a very reduced number of nodes concentrating a high number of edges. These latter highly connected nodes are called hubs. Vitevitch (2008) applied network science to the study of sound organization, showing that there are some aspects of human phonology that can only be detected thanks to the use of networks. The topic of that work was the mental lexicon and hence a key concept for the analysis was shown to be neighborhood density, which accounts for the number of words that sound as a target word.3 Arbesman et al. (2010) found similarities in this regards between Spanish, Mandarin, Hawaiian and Basque (suggesting a psycholinguistic mechanism shared across the architecture of the mental lexicon). Regarding the interaction of lexical items, Ferrer-i-Cancho and Solé (2001) and Corominas-Murtra et al. (2009) represented several language corpora in graphs of word interactions, observing that these large linguistic samples also exhibit a complex behavior. For instance, Solé et al. (2010) show that the novel Moby Dick has the connectivity of a small-world network. Other analyses have explored child language development, observing that network science can also be applied to track the longitudinal stages in linguistic development (Corominas-Murtra et al. 2009). Most network approaches to syntax have used word co-occurrence to account ˇ for syntactic behavior, (there are notable exceptions, like for example Cech et al. 2011; Liu and Hu 2008). However, this changed since Corominas-Murtra et al. (2009) who introduced by hand a collection of syntactic analyses into the networks. The procedure starts by syntactically analyzing all the sentences uttered by a child in a sample of conversation. These analyses followed the basic lines of Dependency Grammar 3 A word is said to be a phonological neighbor of a target word if the substitution, addition or deletion

of a single phoneme in any position in that word converts it to the target word.

276

L. Barceló-Coblijn et al.

(Hudson 1990). Corominas-Murtra used to that end a program especially created for syntactic annotation (Popescu 2003) and transmitted the annotated sentences to a script able to save this information in a format that was readable for the network program: dependency relations—including the arrow direction—were saved in that new format. The network program read that information representing it under the form of a graph: for each lexical item, a node; for each syntactic relation, an edge between two nodes. With a graph from each chronologically ordered piece of corpus, one can follow the development of the child’s ability to combine words. Corominas-Murtra et al. (2009) analyzed two children acquiring English, studying ten chronologically ordered files per child. This macroscopic perspective demonstrated that languagedevelopment undergoes sharp transitions between kinds of networks. While in the first phase the largest structure within a graph was a tree-like network, there was a subsequent abrupt transition into a scale-free network, small-world network along with the concomitant emergence of hubs. Interestingly, hubs were functional words: the, a, that and it. Barceló-Coblijn et al. (2012) reproduced this protocol but this time analyzing three children, each of them acquiring a different first language: Dutch, German and Spanish. In this work seventeen chronologically ordered files per child/language were analyzed. The increased number of files allowed them to obtain a more detailed picture of each child’s development: three different developmental stages, each of them identified by a type of network: phase (1) with a tree-like network, phase (2) with a scale-free network and phase (3) with a scale-free with the characteristic of small-world network. Each phase lasted for several weeks—suggesting linear development—until an abrupt transition started the new phase—suggesting a nonlinear process. Crucially, functional words emerged abruptly as hubs in the second phase (scale-free network) in all three languages. In a further study (Barceló-Coblijn et al., submitted), new longitudinal corpora of new languages have been analyzed following the same protocol: Catalan, French and Italian, plus two additional smaller corpora of Basque (in order to assess the suitability of this technique for agglutinative languages); these three corpora were then compared to 32 linguistic samples of speakers with Down syndrome (DS) (20 Dutch speakers and 12 English speakers). The most important observation was that DS speakers developed in a divergent way, following a developmental path that moved them away from TD children (a rather homogeneous group, regardless of the language developed). C-network analyses in DS speakers revealed that functional words were severely affected by their atypical development: where TD English and Dutch speakers had 30, 40 or 60 edges per hub, DS speakers of both Dutch and English showed a much lower node degree (n of edges of a node), sometimes zero as in the case of the pronoun it.

15 The Emergence of Hubs in Complex Syntactic Networks …

277

15.3 Does Linguistic Theory Affect Complex Networks? 15.3.1 The Central Question All the studies on language ontogeny combining syntactic Dependency-Grammar with complex networks followed the same protocol and reached similar results. In addition to a similar protocol, these works also adopted similar assumptions on the nature of the syntactic relations between words in the speakers’ production. It is true that in certain cases, whatever the theoretical approach to syntax one adopts, the basic assumptions of the relationship between the syntactic objects in a sentence will be comparable. For instance, it is generally accepted that the arguments of a verb are dependents of the verb (even though this may be formalized very differently in each approach). But in other cases, the dependency relation between two connected objects in a sentence will be completely inverse in different approaches. This is the case of the relation between the noun and the determiner in a phrase like the boy, where certain approaches characterize the noun as a dependent of the determiner while others characterize the determiner as a dependent of the noun. How to analyze the syntactic relationship between a determiner and a noun could be non-trivial. In those works where syntactic analyses and complex networks have been combined, determiners have stand out as hubs in all languages so far analyzed. If a different hypothesis is adopted regarding the hierarchy between nouns and determiners, it could potentially yield a different node degree for determiners. This raises a question as to whether linguistic theory determines the emergence of hubs within networks: are determiners generally the hubs of the network (see Sect. 15.2) because of the hypothesis adopted regarding their syntactic nature? Besides, are there arguments that favor one of the two hypotheses? In the next subsection, we briefly overview some of the arguments employed in favor of the shift from the classical “NP-hypothesis” to the “DP-hypothesis” in generative theoretical syntax, and in Sect. 15.4 we report our case-study comparing both syntactic hypotheses.

15.3.2 DP Hypothesis Versus NP Hypothesis As is well-known, generative syntactic theory changed substantially in the eighties and nineties, and an important factor in the change was the adoption of the “DP-hypothesis” (cf. i.a. Abney (1987); Stowell (1989); Longobardi (1994, 2001); Zamparelli (2000); Bernstein (2001)). The classical analysis of arguments (such as the subject or object of the verb) was that they were Nominal Phrases (NP), projections of the noun with articles in the specifier position. The proposal in Abney (1987) is that instead of NPs (1a), we should analyze arguments as richer structures involving a functional layer at the top: DPs (1b).

278

L. Barceló-Coblijn et al.

(1)

The NPh in (1a) assumes that in a phrase such as the little boy from the city, the noun (N) is the head of the structure, the determiner (D) the specifier, the preposition phrase (PP) its complement and the adjective (Adj) an adjunct. On the other hand, following the DPh of (1b) the D stands out as the head that takes the NP as its complement. There are a number of reasons for this move from the NPh to the DPh; a major argument of cohesion is the fact that functional elements constitute closed lexical classes and so do determiners. Besides, a range of morphosyntactic and semantic arguments have been made in favor of the idea that determiners behave as heads of nominal expressions. A main empirical argument is that there are strong similarities between clauses and DPs (cf. “Nero destroyed the city” vs. “Nero’s destruction of the city”). In fact, several languages like Yup’ik, Hungarian or Mayan display morphological agreement in the nominal domain between the possessor and the head noun, as illustrated with the Yup’ik data in (2) and (3) (cf. Abney (1987: 39–42)).4 (2)

(3)

The paradigm in (2) and (3) shows that nouns agree with their possessors (2b)–(3b), and that the agreement morpheme has precisely the same exponent as subject-verb agreement. Besides, the possessor surfaces with ergative case, which is the case of subjects of transitive predicates. This parallelism in agreement and case-assignment is naturally explained, Abney (1987) argues, if we assume parallel structures for clauses (4a) and nominal expressions (4b), where the lexical head kiputaa—(4a)/kuiga—(4b) raises to join to AGR (see also Szabolcsi (1983, 1994)): 4 “OM”

stands for “Object Agreement Marker” and “SM” for “Subject Agreement Marker”.

15 The Emergence of Hubs in Complex Syntactic Networks …

279

(4)

Under the DPh, the structure of the nominal projection is assimilated to that of the clausal projection since in both cases a lexical projection (VP or NP) is dominated by a functional projection (TP and DP respectively). Further evidence for the DPh is found in Romance languages. For instance, in Italian bare nominal (i.e. determinerless) expressions are grammatical in a number of constructions (cf. for instance the vocatives in (5), the predicates in (6), or the exclamatives in (7) (data from Longobardi (1994: 612)): (5)

(6)

(7)

However, in the absence of determiners, argument NPs are ungrammatical, as shown in (8). That is, it is the presence of a D functional layer that turns a nominal expression into an argument (cf. Longobardi (1994); Bernstein (2001) among others):

280

L. Barceló-Coblijn et al.

(8)

Besides, Longobardi argues, a number of constructions illustrate movement from inside NP to higher positions, which can only be captured with a DPh. First, notice that adjectives (and possessive adjectives like mio) in Italian may never occur in the position preceding the determiner: (9)

Now, consider the famous Italian paradigm with proper names and possessive adjectives in (10) due to Longobardi (1994: 623): (10)

Even if both Adj-N (10a) and N-Adj (10c) word orders are possible, the lack of the article forces a N-initial order (hence, the ungrammaticality of (10b)). The adjective may occur in a prenominal position between D and N (10a), or in a postnominal position (10b). However, when there is no overt article, only postnominal positions are available for adjectives (10c). Longobardi shows that this gap can be naturally accounted for with the DPh, by assuming that articleless proper names in Italian undergo raising from N0 (the head of the NP) to D0 (the head of the DP), thus crossing over the prenominal adjective position (Spec-NP): (11)

15 The Emergence of Hubs in Complex Syntactic Networks …

281

Fig. 15.1 Syntactic annotations in Netlang of the dependency relations in (12). The DPh (12a) is illustrated on the left, the NPh (12b) on the right

This brief overview presented just a few pieces of evidence in favor of the DPh, but the literature of the last twenty years is full of morphological, syntactic and semantic evidence and arguments supporting a rich functional structure above nominal phrases (see Bernstein (2001) for an overview). Now, regarding network-analysis, the choice of the DPh over the NPh is crucial, since the number of connections could differ substantially. For example, if the determiner is the head of a nominal expression (12a), it will participate in more dependency-relations than if it is the noun that heads the structure, with the determiner being its mere specifier (12b). Figure 15.1 illustrates this: (12) John prepares the hamburger a. John ⇒ prepares ⇐ the ⇐ hamburger [DPh] b. John ⇒ prepares ⇐ hamburger ⇐ the [NPh] In (12a) the determiner the has two connections, whereas in (12b) it has only one. Therefore, the decision about the nature and dependencies of this kind of syntactic items within the structure of the utterance can affect in a crucial way the final network.

15.4 Procedure and Results In order to test whether there are significant differences between the DPh and the NPh, three corpora were studied in this work, from three different languages with different typological properties and belonging to different families: Dutch (Germanic), French (Romance) and Basque (language isolate). The corpora were analyzed twice: once assuming the DPh and once assuming the NPh (hence, two files—two networks—were obtained from each corpus). Each linguistic expression produced by the child was analyzed using the Netlang software, a new tool created overcome some problems detected in previous c-network analyses (Fig. 15.2). Netlang allows to analyze linguistic expressions one by one. Crucially, it incorporates the possibility of selecting the particular relationship existing between two lexical items (e.g., subject, object, complement, modifier, determiner; it also makes

282

L. Barceló-Coblijn et al.

Fig. 15.2 Screenshot of the Netlang software used for the syntactic analyses. On the left: (bottom) transcription of the speech corpus to be analyzed and (top) historic of analyzed sentences; in the middle: (bottom) grammar of the program and (top) table with the origin nodes, the syntactic relationships and the goal nodes; on the right: graphic representation of the network. In front, the window of the annotator with an example of syntactic analysis

it possible to add new ones). This allows the researcher to analyze the data through particular assumptions. In this case, we analyzed Ns as “complements” of Ds (DPh) versus Ds as “determiners” of Ns (NPh). All three files were downloaded from the CHILDES data base (MacWhinney 2000) from a series of longitudinal studies of children acquiring a first language. The Dutch file belongs to the folder Groningen, and corresponds to a Dutch boy (Daan) at age 03;00.15 (Wijnen and Verrips 1998). The Basque file belongs to the folder Soto, and reproduces the speech of a Basque boy (Ander) at age 2;09.23—(Soto Valle 2012). Finally, the French file belongs to the folder Lyon and the girl’s name is Anaïs at age 2;05.25—(Demuth and Tremblay 2008).

15.4.1 The Analysis The first phase of our analysis involved the syntactic annotation of the speech corpora. We followed a protocol developed to this end. The protocol moves away from Corominas-Murtra’s (2007) structural criteria. For example, in comparison to previous works, several relationships have been eliminated: for example, for this study a child’s utterance like «car there» is considered to involve predication between the items, but not a syntactic dependency (following Moro (1997, 2000) and others, both items are taken to be merged in a symmetric small clause dominated by a copular

15 The Emergence of Hubs in Complex Syntactic Networks …

283

head that will break the symmetry and establish syntactic dependencies). Hence all expressions containing this kind of structures—«X there»—have not been annotated as involving syntactic relationships (productions like these at the two-word stage are taken to be the outcome of a specific linguistic constraint (cf. Berk and Lillo-Martin (2012)). In sum, the application of this protocol has produced a reduction of nodes and edges, in comparison to previous studies, that has rendered a drastic reduction in some cases. Because the present study does not target phonetic variation, words produced under several phonetic forms have been unified, in order to minimize the appearance of multiple nodes resulting from children’s speech errors. This criterion has also been applied to standard phonetic variability (e.g., the Dutch determiner het is sometimes realized as ’t, and hence both were unified under the form het). Once a corpus was unified/standardized and syntactically annotated, the .netlang file was exported to a .sif file, one of the formats accepted by the network program Cytoscape 2.8.3 (Shannon et al. 2003), which was employed for the statistical analyses of the networks. Then the degree k (n of edges) of determiners and pronouns was recovered and compared (see Tables 15.1, 15.2 and 15.3). It is crucial to adequately disambiguate homophonous items. In doing this, we have followed Barceló-Coblijn et al. (2012). Given that some determiners are ambiguous with pronouns, the two functions must be differentiated (Table 15.1). For example, Dutch het can be a determiner (e.g., het [determiner] raam, «the window»), but it can also be a pronoun (e.g., ik heb het [pronoun] gezien, «I have seen it»). Disambiguation was done manually, on the basis of the analysis of the expression’s syntactic context. As can be seen in Fig. 15.3, from each conversation file a graph was obtained. A graph can contain one or more components (networks having one or more nodes). The largest network is the focus of our attention, although the whole graph is also

Table 15.1 Degree (n of edges) of each determiner and pronoun in the annotated corpora (Dutch)

Dutch

DPh

NPh

de het [determiner]

20 1

11 1

het [pronoun] een[determiner] een[pronoun] deze[determiner]

6

5

30

21

7

7

21

11

deze[pronoun]

5

5

die[determiner]

10

5

die[pronoun]

17

13

dit [determiner]

2

1

dit [pronoun]

3

3

dat [determiner]

5

3

dat [pronoun]

5

5

284 Table 15.2 Degree (n of edges) of each determiner and pronoun in the annotated corpora (Basque)

Table 15.3 Degree (n of edges) of each determiner and pronoun in the annotated corpora (French)

L. Barceló-Coblijn et al. Basque

DPh

NPh

-a[determiner]

41

26

-ak [determiner]

14

7

bat [determiner] , batzuk [determiner]

10, 1

7, 1

hau[demonstrative/pronoun] , hauek [demonstrative/pronoun]

4, 1

4, 1

hori[demonstrative/pronoun] , horiek [demonstrative/pronoun]

4, 1

4, 1

hura[demonstrative/pronoun] , haiek [demonstrative/pronoun]

–, –

–, –

French

DPh

NPh

la[determiner] , le[determiner] , les[determiner]

5, 7, (3)

4, 6, (3)

un[determiner] , une[determiner]

14, 2

12, 2

ce[determiner] , ça[demonstrative/pronoun] , ceci[demonstrative/pronoun]

–

–

celle-là[demonstrative/pronoun]

2

2

celui-là[demonstrative/pronoun]

1

1

important for some measures. The largest network is also known as the Giant Connected Component (GCC). Other indicators were also recovered in order to see whether the option selected could affect more generally the structure of the network. These indicators and their scores were gathered in Table 15.4. The clustering coefficient C is a measure of cohesion. C indicates how many nodes that connect to a node are also connected to each other (thus, forming a triangle). The more “triangles”, the higher the level of cohesion.5 Then, the path length L indicates how many “steps” are necessary in order to go from one randomly selected node to another randomly selected node. Next, it is interesting to see, on average, how many edges a node has. This is indicated by the average degree of the GCC k. And finally, RS or the Relative Growth of the GCC shows the growth of the largest network within the graph, the Giant Connected Component, in comparison with the whole graph.

5 An

example typically used to illustrate C is, thinking in nodes as people and edges as friendship relationships, “how many of my friends are also friends of each other?”.

15 The Emergence of Hubs in Complex Syntactic Networks …

285

Fig. 15.3 The networks of the Dutch sample after the syntactic analysis. Top left, the analysis adopting the Determiner Phrase (DP) hypothesis, bottom left the analysis adopting the Noun Phrase (NP) hypothesis. Figures on the right show a zoom of each network, highlighting the Dutch determiner de with the kind of relationships that this determiner has with other words Table 15.4 Clustering coefficient C, path length L, average n of edges per node k, Relative Size of the Giant component of the graphs. There are two columns per language, corresponding to the DPh and the NPh Dutch French Basque DPh

NPh

DPh

NPh

DPh

NPh

C L k

0.048 3.393 3.7

0.096 3.496 3.933

0.057 3.933 2.495

0.032 4.122 2.455

0.077 3.251 3.314

0.051 3.596 3.471

RS

0.8780

0.8774

0.6733

0.6733

0.8869

0.8869

15.5 Discussion The three corpora were analyzed in two different ways: first applying the DPh and then applying the NPh. According to the DPh, the determiner is the head, and the

286

L. Barceló-Coblijn et al.

NP its complement, whereas according to the NPh the noun is the head and the determiner its specifier (see Sect. 15.3.2 above). First of all, the degree k is much higher in determiners than in the average degree of the rest of the nodes within the network (between 3.1 and 3.5 in all languages analyzed here), regardless of the hypothesis endorsed. This confirms the relevance of determiners in the macro-syntactic structure of discourse. Determiners are much more highly connected than the rest of items in both conditions. Then, as expected, there are differences in the degree k of determiners, depending on the hypothesis adopted. The difference in the degree k of the whole network does not reach levels of statistical significance. However, significant differences were observed between the NPh and the DPh: C (clustering coefficient) is significantly higher under the DPh in both French and Basque (though not in Dutch). The rest of network indicators analyzed here did not show significant differences; but still, the tendencies are clear: under the DPh determiners show more connectivity than under the NPh. If we compare these measures with those of other functional elements such as adpositions, we see that the DPh provides a clearer pairing between both categories (for instance, French preposition comme (like) and Basque inessive postposition -n also show a high degree of connectivity, which makes them emerge as hubs). As a result, we have a clearer match between determiners and adpositions under the DPh than under the NPh, a further reason for adopting the DPh in c-network studies. The existence of all these patterns suggests that the fact that not all differences are statistically significant is to be attributed to the small size of the corpora. Besides, it is clear that the new protocol of syntactic analysis has had a more significant impact than the application of one of the two previous hypotheses. The new protocol is much more strict regarding the kind of syntactic relationships to be sanctioned and, crucially, it represents a set of procedural rules much more detailed and integrated within current linguistic theory. To date, most network approaches barely take into account what are considered significant tenets in syntactic theory—clearly, the application of co-occurrence simply neglects such theoretical advances. We believe that future studies involving larger corpora will show significant results in what we observed here as meaningful tendencies. The present work represents a significant advance in the application of Network Science (Newman (2010)) to language. Firstly, it has been able to test the significance of the adoption of competing theories in linguistics, showing which aspects affect more profoundly the macroscopic pattern depicted by the syntactic network. Moreover, it is the first application of a new software exclusively created to reconcile linguistic analysis and its representation into networks. We think that this work will serve as a firm bridge between linguistic theory and network science, a combination of scientific disciplines that, we believe, will render interesting results for cognitive science in general.

15 The Emergence of Hubs in Complex Syntactic Networks …

287

Acknowledgements This research benefited from the following grants: EC FP7/SSH-20131 AThEME 613465 (European Commission), IT769-13 (Eusko Jaurlaritza), TIN2016-80347R (MICINN), FFI2016-78034-C2-2-P and FEDER, FFI2017-87140-C4-1-P, FFI2014-53675-P (MINECO).

References Abney, S. P. (1987). The English noun phrase in its sentential perspective. Doctoral dissertation. Cambridge: Massachusetts Institute of Technology. Retrieved form http://www.ai.mit.edu/ projects/dm/theses/abney87.pdf. Arbesman, S., Strogatz, S. H., & Vitevitch, M. S. (2010). The structure of phonological networks across multiple languages. International Journal of Bifurcation and Chaos, 20, 679–685. https:// doi.org/10.1142/s021812741002596x. Barceló-Coblijn, L., Corominas-Murtra, B., & Gomila, A. (2012). Syntactic trees and small-world networks: Syntactic development as a dynamical process. Adaptive Behavior, 20, 427–442. https:// doi.org/10.1177/1059712312455439. Barceló-Coblijn, L., Real Puigdollers, C., Irurtzun, A., López-Navarro, E. & Gomila, A. (2018). How children develop their ability to combine words: A network based approach. Manuscript submitted for publication. Barceló-Coblijn, L., Serna Salazar, D., Isaza, G., Castillo Ossa, L. F., & Bedia, M. G. (2017). Netlang: A software for the linguistic analysis of corpora by means of complex networks. PLoS ONE, 12(8), e0181341. https://doi.org/10.1371/journal.pone.0181341. Berk, S., & Lillo-Martin, D. (2012). The two-word stage: motivated by linguistic or cognitive constraints? Cognitive Psychology, 65, 118–140. https://doi.org/10.1016/j.cogpsych.2012.02.002. Bernstein, J. B. (2001). The DP hypothesis: Identifying clausal properties in the nominal domain. In M. Baltin & C. Collins (Eds.), The Handbook of Contemporary Syntactic Theory (pp. 536–561). Oxford: Blackwell Publishers Ltd. ˇ Cech, R., Maˇcutek, J., & Žabokrtský, Z. (2011). The role of syntax in complex networks: Local and global importance of verbs in a syntactic dependency network. Physica A: Statistical Mechanics and its Applications, 390, 3614–3623. https://doi.org/10.1016/j.physa.2011.05.027. Corominas-Murtra, B. (2007, April 30). Network statistics on early English Syntax: Structural criteria. Retrieved from https://arxiv.org/pdf/0704.3708.pdf. Corominas-Murtra, B., Valverde, S., & Solé, R. V. (2009). The ontogeny of scale-free syntax networks: Phase transitions in early language acquisition. Advances in Complex Systems (ACS), 12, 371–392. https://doi.org/10.1142/s0219525909002192. Demuth, K., & Tremblay, A. (2008). Prosodically-conditioned variability in children’s production of French determiners. Journal of Child Language, 35, 99–127. https://doi.org/10.1017/ s0305000907008276. Fenson, L., Marchman, V., Thal, D., Dale, P., Reznick, J., & Bates, E. (2007). The MacArthur-Bates communicative development inventories user’s guide and technical manual (2nd ed.). Baltimore: Brookes Publishing. Ferrer-i-Cancho, R., & Solé, R. V. (2001). The small world of human language. Proceedings Biological Sciences/The Royal Society, 268, 2261–2265. https://doi.org/10.1098/rspb.2001.1800. Hudson, R. A. (1990). English word grammar. Oxford: Blackwell. Liu, H., & Hu, F. (2008). What role does syntax play in a language network? Europhysics Letters, 83: 18002 Longobardi, G. (1994). Reference and proper names: A theory of N-movement in syntax and logical form. Linguistic Inquiry, 25, 609–665.

288

L. Barceló-Coblijn et al.

Longobardi, G. (2001). The structure of DPs: Some principles, parameters, and problems. In M. Baltin & C. Collins (Eds.), The handbook of contemporary syntactic theory (pp. 562-603). Oxford: Blackwell. MacWhinney, B. (2000). The childes project: The database. New York & London: Psychology Press. Moro, A. (1997). The raising of predicates: Predicative noun phrases and the theory of clause structure. Cambridge: Cambridge University Press. Moro, A. (2000). Dynamic antisymmetry. Cambridge: MIT Press. Newman, M. (2010). Networks: An introduction. Oxford: Oxford University Press. Popescu, M. (2003). Dependency grammar anotator. In F. Hristea & M. Popescu (Eds.), Building awareness in language technology (pp. 17–34). Bucharest: Editura Universit˘a¸tii din Bucure¸sti. Ramos, J. R. (1992). Introducció a la sintaxi. València: Tandem Edicions. Sagae, K., Davis, E., Lavie, A., Macwhinney, B., & Wintner, S. (2010). Morphosyntactic annotation of CHILDES transcripts. Journal of Child Language, 37, 705–729. https://doi.org/10.1017/ s0305000909990407. Scarborough, H. S. (1990). Index of productive syntax. Applied Psycholinguistics, 11, 1–22. https:// doi.org/10.1017/s0142716400008262. Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D., et al. (2003). Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Research, 13, 2498–2504. https://doi.org/10.1101/gr.1239303. Solé, R. V., Corominas-Murtra, B., Valverde, S., & Steels, L. (2010). Language networks: Their structure, function, and evolution. Complexity, 15, 20–26. https://doi.org/10.1002/cplx.20305. Soto Valle, R. (2012). La adquisición del euskera: Aproximación a una secuencia típica del desarrollo morfosintáctico de 21/2 a 5 años. Doctoral dissertation. University of the Basque Country UPV/EHU. Stowell, T. (1989). Subjects, specifiers, and X-bar theory. In M. R. Baltin & A. Kroch (Eds.), Alternative conceptions of phrase structure (pp. 232–262). Chicago: University of Chicago Press. Szabolcsi, A. (1983). The possessor that run away from home. The Linguistic Review, 3, 89–102. https://doi.org/10.1515/tlir.1983.3.1.89. Szabolcsi, A. (1994). The noun phrase. In F. Kiefer & K. É. Kiss (Eds.), The syntactic structure of Hungarian, volume 27 (pp. 179–274). New York: Academic Press. Vitevitch, M. S. (2008). What can graph theory tell us about word learning and lexical retrieval? Journal of speech, language, and hearing research, 51, 408–422. https://doi.org/10.1044/10924388(2008/030). Watts, D. J., & Strogatz, S. H. (1998). Collective dynamics of “small-world” networks. Nature, 393, 440–442. https://doi.org/10.1038/30918. Wijnen, F., & Verrips, M. (1998). The acquisition of Dutch syntax. In S. Gillis & A. De Houwer (Eds.), The acquisition of Dutch (pp. 223–300). Amsterdam/Baltimore: John Benjamins. Zamparelli, R. (2000). Layers in the Determiner Phrase. London & New York: Routledge.

Chapter 16

The World Color Survey: Data Analysis and Simulations Peter Lewinski, Michal Lukasik, Konrad Kurdej, Filip Leonarski, Natalia Bielczyk, Franciszek Rakowski and Dariusz Plewczynski

Abstract The distribution of colors in the environment shapes local peoples’ perceptions of those colors, a phenomenon observable across all types of environments. We analyzed color categorization data from each of the 107 languages in the World Color Survey (WCS) database. Next, we grouped the WCS languages according to their geographic location, with reference to the seven terrestrial habitats (biomes) classified by the World Wildlife Fund (WWF). We developed a computer algorithm to establish the most frequently occurring colors in each environment based on the color distribution extracted from National Geographic natural images of the respective biomes. We then compared the average standardized value of the mode (i.e., most frequently occurring answers) for each group of WCS languages; we followed The original version of this chapter was revised: Co-author name has been removed. The correction to this chapter is available at https://doi.org/10.1007/978-3-030-04598-2_18 Dariusz Plewczynski—Senior author. Peter Lewinski, Michal Lukasik, Konrad Kurdej and Dariusz Plewczynski have contributed equally to this work. P. Lewinski Faculty of Law and Saïd Business School, University of Oxford, Oxford OX1 2JD, UK e-mail: [email protected] M. Lukasik Google, Brandschenkestrasse 110, 8002 Zurich, Switzerland K. Kurdej Faculty of Mathematics, Informatics and Mechanics, University of Warsaw, Stefana Banacha 2c, 02-097 Warsaw, Poland e-mail: [email protected] F. Leonarski Faculty of Chemistry, Interdisciplinary Centre for Mathematical and Computational Modelling, University of Warsaw, Stefana Banacha 2c, 02-097 Warsaw, Poland e-mail: [email protected]; [email protected] N. Bielczyk Polish Academy of Sciences, Radboud University Nijmegen Medical Centre, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, Netherlands e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_16

289

290

P. Lewinski et al.

the same procedure for the most frequently occurring colors as well as the remaining colors. Results indicated statistically significant lower values of the average mode answers for the most frequently occurring colors. These results support our hypothesis that the environment type shapes color category boundaries. Further, we follow Steels and Belpaeme’s (Behav Brain Sci 28:469–489, 2005) model, which allows for computer simulations of the cultural emergence of color categories. An agent-based model of the cultural emergence of color categories shows that boundaries might be seen as a product of agent’s communication in a given environment. We propose the extension of this generic agent-based modeling framework to include a culturally driven emergence of color categories. We therefore underscore external constraints on cognition: the structure of the environment in which a system evolves and learns, and the learning capacities of individual agents. Finally, we discuss the methodological issues related to real data characterization (World Color Survey), as well as to the process of modeling the emergence of perceptual categories in human subjects.

16.1 Introduction Does the environment in which people live shape their cognition and perception? There is strong evidence that Chinese people conceptualize time vertically, not horizontally, as is typical of Western nations. One possible explanation of this phenomenon is that in China, characters are written from top to bottom, whereas in the Latin alphabet, letters are written from left to right (Boroditsky 2001). The Inuit people have in their linguistic repertoire many more names and categories for snow than people from different—less snow—environments (see for review: Martin 1986; Pullum 1989). Therefore, the question to ask is whether language usage influences perception. On the one hand, a relativist approach (Roberson et al. 2000, 2005; Roberson and Davidoff 2000) and the Sapir-Whorf hypothesis states that “the structure of anyone’s native language strongly influences or fully determines the world-view he will acquire as he learns the language” (Brown as cited in Kay and Kempton 1984, p. 2). On the other hand, a universalist approach (Kay and Regier 2003; Cook et al. 2005), strongly advocated by Berlin and Kay in their book “Basic Color Terms: Their Universality and Evolution” (1969), states that the number of names for colors in a culture can predict the number of basic color terms in that culture. In addition, Berlin and Kay’s model suggests a strict chronological sequence (seven stages) of basic color term development in any culture. F. Rakowski Interdisciplinary Centre for Mathematical and Computational Modelling, University of Warsaw, Stefana Banacha 2c, 02-097 Warsaw, Poland e-mail: [email protected] D. Plewczynski (B) Centre of New Technologies, University of Warsaw, Stefana Banacha 2c, 02-097 Warsaw, Poland e-mail: [email protected]; [email protected]

16 The World Color Survey: Data Analysis and Simulations

291

Research on colors, which involves comparing linguistics categories (and therefore the structure of language), is commonly used to answer many fundamental questions about language and perception in general. The physical stimuli of colors (hue, saturation and value) are easily measureable; moreover, people from all cultures are exposed to colors, so we are able to compare them across languages. Likewise, the categorization process is a well-defined phenomenon and an important part of human perception, as well as being effectively measurable (see e.g.: Belapeme and Bley 2005; Baronchelli et al. 2010). Humans speak thousands of languages, all of which have varying numbers of names for colors. People from different cultures perceive color category boundaries in different ways. Many universalistic tendencies exist on the general level, such as frequently encountered patterns within the seven stage acquisition model (Berlin and Kay 1969). There is also significant variability on the local level (Heider 1972; Kay and Regier 2006), namely the huge variability in the number of color names and their exact boundaries. Research confirms that the surroundings in which one is raised influences the perception of colors. For instance, early surroundings shape the development of neuronal pathways in the brain, which are responsible for the perception of certain stimuli (e.g., orientation of lines and edges in the visual field) later in life (Hein et al. 1970; Blakemore and Cooper 1970). Leventhal and Hirsch (1975) demonstrated this effect in cats; when brought up in an environment without diagonal lines, cats could not see such lines later in life. More recently, these theories regarding the categorization and naming of colors have been simulated in computer models (see Steels and Belpaeme 2005). Using a modification of one such model, Puglisi et al. (2008) have shown how specially trained computer agents develop categories for colors depending on the distribution of colors in the agents’ “environment.” As it happens, the number of linguistic labels of the categories agreed upon jointly by the computer agents depends on the frequency of occurrence of that particular color. Uniform distribution of colors and two pictures (warm colors vs. cold colors) were presented to the agents with the distribution of colors in the environment. Researchers found that more labels (and therefore categories) were created for more frequently occurring colors. The simulations exploring the categorization of colors are focused on the central problem for cognitive science, i.e., how representations should be modeled. The symbol-grounding (SG) problem originates in the long-term discussion of how to build a functional representation of the external world using some internal representation of a cognitive system, a problem encapsulated in the semiotic triad (ST). The semiotic triad is express in the relation between the following three elements: an object (external, real-world entity with which a cognitive system interacts); a concept (some kind of action, or the set of features that are applicable to the object); and a name (symbolic internal representation for the class of objects characterized by similar features or similar actions). Peirce (1839–1914) introduced the idea of the semiotic triad to link real-world objects to the symbols human observers used to describe them. Searle (1994) presented the idea that semiotics is not only a linguistic theory, but moreover a theory of how to build meaning out of sets of names or objects. Steels (1998) proposed a

292

P. Lewinski et al.

set of principles that may explain how language and meaning originate in a group of physically grounded cognitive agents. He further observed the emergence of distinctions, a lexicon, and primitive syntactic structures. Steels also focused on the issue of how symbols are linked to things in the external world; what embodiment means for the language; and under what conditions cognition is truly embodied (Steels and Kaplan 1999; Steels 2001, 2002, 2003, 2006; Steels and Belpaeme 2005; Wellens et al. 2008). On a similar trajectory, Gärdenfors (2000) proposed a geometric model of external, “real” objects based on conceptual spaces (Gärdenfors 2000). This metatheory is located between the symbolic and connectionist modes of representation that dominate the field of cognitive science. The proposed geometric representation of information is useful for modeling concept formation, semantics, non-monotonic inferences and inductive reasoning (Gärdenfors 2000). In most agent-based models of communication, symbols are treated in the traditional manner—as entities that can be mapped to external objects. These models assume that semantics can be unequivocally ascribed to a symbol. Steels and Belpaeme (2005) analyzed the cultural emergence of color categories using their original modeling framework. Based on that framework, we propose here the generic agent-based modeling framework of cultural emergence of color categories shows that boundaries might be seen as a product of agent’s communication in a given environment. We discuss the methodological issues related to real data characterization (World Color Survey), as well as to the process of modeling the emergence of perceptual color categories in human subjects. Afterwards, we present a generic model of a cognitive system, where a symbol links to dynamic behavior of two cognitive systems, therefore constraining its function. We compare the real-life data on the categorization of colors in a multitude of cultures, with computational approaches and mathematical methods to verify the hypothesis that environment type shapes the color categorization process. Based on previous research findings (Heider 1972; Puglisi et al. 2008) and commonsense assumptions, we reasoned that the more of a particular color in the environment, the more agreement on the category’s boundaries for that color will be present between the speakers of the culture inhabiting those surroundings. In other words, the more people are exposed to a color, the better they will communicate about those colors to each other. By contrast, they will be less proficient in discriminating between colors with which they do not have much contact. For example, indigenous people from the African Savannah should communicate better and more readily agree on the color category boundaries between each other with respect to the colors they are most exposed to on a daily basis—hypothetically, shades of yellow and brown. However, Inuit peoples should not be as accomplished in discriminating different hues of yellow and brown, therefore agreeing less on those two color category boundaries. More agreement on color category boundaries for colors that are most frequent would indirectly indicate that people from the same culture perceive those colors in a similar manner and can communicate about them in a faster and more reliable way. In our research, we highlight external constraints on cognition, namely the structure of the environment. We conducted statistical analyses of the influence of the environment on people’s and agents’ ability to reach consensus. We also propose two

16 The World Color Survey: Data Analysis and Simulations

293

methodological innovations in studying color perception: (i) an approach for determining color frequency by collecting and analyzing representative natural images; and (ii), a cognitive architecture that allows for simulation of semiotic triad formation.

16.2 Methods The main idea behind computational approaches is that the categorization process can be used for grounding artificial symbols in real-world objects. Typically, categorization is performed using either clustering algorithms or machine learning methods. We propose here a novel cognitive architecture allowing for simulation of semiotic triad formation process using an ensemble of machine learning and clustering algorithms. The semiotic triad is a dynamic collection of learning algorithms, enabling the assignment of names to given objects, therefore translating them into underlying categories or concepts that are linked to the objects. Following Pierce and Steels’ definition, the semiotic triad is the set of three elements: a concept (i.e., an implicit category of objects), a symbol (i.e., an abstract/symbolic category name) and a mapping between subsets of objects belonging to a corresponding category and the concept itself. Therefore, the semiotic triad marries a category with the subsets of real objects (training examples for a given category) and the symbol denoting the category. In other words, the method constrains the use of a particular symbol to objects belonging to a given category. Furthermore, studying cultures and languages native to various environments and assessing which colors are most frequently encountered in those environments requires: • • • •

a precise definition and common measure of “environment” a database of color categorization from languages representing all environments a list of colors most frequently encountered in the environments a statistical way of comparing color category boundaries.

In the method section, each of these steps is operationalized and defined. Color naming data comes from WCS (http://www.icsi.berkeley.edu/wcs/); for the definition of “environment,” we adopt the concept of biomes (Olson et al. 2001). The frequency of colors in each biome is calculated using natural images from National Geographic magazine (2001). The color category boundaries are compared to each other using ANOVA analysis of the average standardized value of the mode (i.e., most frequently occurring answers) between the colors (Munsell chips) that were established to be most frequent and the rest of the colors.

294

P. Lewinski et al.

16.2.1 World Color Survey The World Color Survey is a data archive created by Cook, Kay and Regier (http:// www.icsi.berkeley.edu/wcs/) in which researchers asked people from 110 unwritten, mostly pre-industrial cultures how they perceive colors. On average, 24 native speakers representing each culture were asked to identify 330 Munsell chips (cards) from the associated chart (See Table A.1 in Annex: Supplementary Materials). A Munsell chart consists of 330 colored cards, where the cards are arranged vertically by lightness and horizontally by hue (see Fig. 16.1). The Munsell chip is therefore a standardized stimulus for presenting a color. The speakers were shown the chips in constant, random order, and the entire procedure was as controlled as possible. All fieldworkers were provided with identical instructions for the color-naming task.

16.3 Natural Biomes To group together languages that are observable in similar types of environment, we needed a common classification of the term “environment.” We used the notion of terrestrial biomes, carefully explained by Olson et al. (2001). Biomes are biogeographic regions that rather neatly divide Earth’s terrestrial biodiversity. The biome concept was originally developed by the World Wildlife Fund (WWF) (2011a) in their Conservation Science Program (CSP). Essentially, this classification is based on differences between various regions in terms of climate, fauna, flora, landscape, habitat,

Fig. 16.1 Cognitive systems simulation framework. Picture Authors’ own

16 The World Color Survey: Data Analysis and Simulations

295

biodiversity, human activity, etc. According to the WWF classification, Earth’s land surface includes 14 main biomes (see Table 16.1), including Forests (e.g., Tropical, Temperate, Broadleaf, Mediterranean), Grasslands, Shrublands, Tundra, Deserts and Mangroves.

16.3.1 Environmental Distribution of Colors Establishing which colors are most frequently encountered in biomes was the most difficult part of our study, as to our knowledge no objective approach has been developed to assess the frequency of occurrence of colors in the environment. Nonetheless, we adhered as much as possible to strict methodological rules in the creation of our own novel method. We elected to extract the frequency of colors by collecting a set of the most representative natural images of the biomes encountered in the WCS languages and employing a computer algorithm to measure its color qualities. The rationale behind using natural images for establishing to which colors people are most exposed is that there is perhaps no better approximation than this method. No technology exists to “scan” the images people see in front of their eyes. However, one can capture many pictures of peoples’ natural environment and feed those images to a computer program capable of analyzing the distribution of colors for that particular habitat. Step one was to prepare enough natural images. We decided against the more popular means of acquiring pictures, like Google Images or a similar web-based provider, because we could not ensure those pictures would faithfully represent the eco-regions in question. Additionally, the differences in technical parameters between images would render the estimation process less reliable. Fortunately, we discovered a large database of high-quality pictures classified in accordance with the

Table 16.1 WWF classification of world biomes ID Name of biome ID 1

5

Tropical and subtropical moist broadleaf forests Tropical and subtropical dry broadleaf forests Tropical and subtrobical coniferous forests Temperate broadleaf and mixed forests Temperate conifer forests

6

Boreal forests/taiga

7

Tropical and subtropical grasslands, savannas and shrublands

2 3 4

8

Name of biome

9

Temperate grasslands, savannas and shrublands Looded grasslands and savannas

10

Montane grasslands and shrublands

11

Tundra

12 13

Mediterranean forests, woodlands and scrub Deserts and xeric shrublands

14

Mangroves

296

P. Lewinski et al.

WWF eco-region (and therefore biome) classifications: National Geographic (2001). We manually downloaded all available pictures for biomes present in WCS (that is, biomes number 1, 2, 3, 4, 6, 7 and 13), and we performed the following steps to ensure that the pictures represented as accurately as possible what people living in each of these biomes see every day: • Filtering out pictures not showing the landscape (i.e., not using pictures of animals and plants). • Cutting out the sky in all remaining pictures. We carried out these modifications to the original set of pictures so as not to receive artificially high frequencies of some colors (especially colors that might be present on such pictures because they are exotic, e.g., a rarely seen red frog). Similarly, we cropped some pictures to remove the sky, as we worried the presence of sky in the picture would exaggerate the frequency of blue in the color distribution. Altogether, we analyzed 143 pictures. The number of pictures used per biome is shown below in Table 16.2. The number of pictures per biome varies because we collected all natural images available for that biome from National Geographic’s database, and the number of pictures per biome varied. At the same time, our filtering process contributed to rejecting a different number of pictures per biome.

16.4 A Code that Associates Pixels on a Photograph with Munsell Chips In the original WCS data set, colors were distributed uniformly in 3-dimensional Lab space aiming at best approximating human vision, where “L” stands for lightness, “a” denotes the red-green scale, and “b” denotes the yellow-blue scale. To associate any given pixel from the data with appropriate Munsell chip, we first represented the RGB color value of that pixel in Lab coordinates and then found a Munsell chip whose value was the closest to the obtained value in Euclidean metrics.

Table 16.2 Number of natural images per biome

Biome ID

#pictures

Biome ID

#pictures

1 2 3 4

20 16 8 28

6 7 13

10 22 39

16 The World Color Survey: Data Analysis and Simulations

297

16.5 Approximation of the Color Distribution Across Biomes This problem was approached in three steps: • Creating a histogram of Munsell chips represented by pixels of a given photograph and filtering out the 5% of pixels corresponding to the rarest chips. The purpose here is to exclude accidental details in the photographs; • Adding a weight to all pixels in a given photo, so that a product of a number of pixels and a weight for every photograph is the same for a given biome; • Making a joint distribution of Munsell chips represented by the aggregate set of pixels from all photographs in a given biome with respect to their weights, and then once more removing the outlying 5% of pixels representing the rarest chips. This avoids the possibility of having an unusual image that stands out from the rest, e.g., 50 largely similar pictures of green jungle with a single picture of uncommon yellow flowers. The resulting sets of colors are considered representative for a given biome. To obtain a universal set of colors, we conducted an identical procedure with the total set of pictures for all biomes. The next step involved extracting color frequency distribution. Two of our researchers skilled in physics and computer science developed an algorithm to accomplish this task. The exact procedure implemented falls outside the scope of this article. Counting the distribution of colors in the natural images from National Geographic allowed us to create a Munsell charts with the most frequently occurring colors (Munsell chips) indicated for each of the biomes present in WCS (Fig. 16.2). The frequency criterion was defined in the computer algorithm as those colors that, after adding them together, accounted for 95% of the color distribution present in the respective set of biome pictures.

Fig. 16.2 Munsell chart. Letter coordinates (A…J) denote 10 degrees of Munsell value (lightness); numeral coordinates denote 40 equally space Munsell hues. All chips at maximum Chroma for the (Hue, Value) pair. http://www.icsi.berkeley.edu/wcs/data.html

298

P. Lewinski et al.

16.6 Clustering of Colors Names We compared the average standardized value of the mode (i.e., most frequently occurring answers) between the Munsell chips, which were established to be the most frequent Munsell chips along with the remainders, or non-frequent colors. The results were standardized to “1,” according to the number of speakers in each culture. The mean and the standard deviation values are presented in Table 16.3 (Fig. 16.3).

Table 16.3 The average standardized value of the mode Frequent M SD Non-frequent colors (chips) colors (chips) Biome 1 Biome 2 Biome 3 Biome 4 Biome 6 Biome 7 Biome 13

0.6757 0.6249 0.6169 0.5758 0.3568 0.6368 0.6035

0.1031 0.1031 0.0800 0.1085 0.3384 0.0795 0.1126

Fig. 16.3 Distribution of colors in biomes

Biome 1 Biome 2 Biome 3 Biome 4 Biome 6 Biome 7 Biome 13

M

SD

0.6669 0.6708 0.6706 0.6127 0.3717 0.6186 0.6709

0.1171 0.1299 0.0954 0.2041 0.3354 0.0925 0.1481

16 The World Color Survey: Data Analysis and Simulations

299

16.7 Results 16.7.1 Data Analysis Repeated measure two-way ANOVA analysis was carried out to establish any differences between the standardized average mode values of the Munsell chip types (frequent and non-frequent) and the type of biome (6 biomes). A significant difference was found between the types of the Munsell chip, F(1,102)  6.66, p < 0.05, partial eta2  0.06, however, there was also a significant, unexpected difference between the types of biome, F(6,102)  2.94, p < 0.05, partial eta2  1.48. Bonferroni post hoc test revealed this result was due to unusually high difference between biome number six and the rest of the biomes. The two languages from biome number six were dropped and once more a significant difference between the type of chip (frequent and non-frequent) was found, F(1,101)  8.27, p < 0.01, partial eta2  0.08, however, there was no longer a significant difference between the types of biome. Analysis also showed a significant interaction between the type of the Munsell chips and the type of biome, F(5,101)  4.04, p < 0.01, partial eta2  0.17 (Fig. 16.4). Through analysis of individual cultures, we show that for 22 of 110 in the WCS study, mode for colors with higher frequency is significantly different from other colors. This effect is strongest in biome 13. This study was accompanied by a general analysis of patterns of colors that have the highest or lowest mode. Manual analysis of these data for cultures with a similar number of color categories and high quality answers showed that two large clusters can be determined. The first one are cultures with a better agreement on categories for green and blue colors and the second one being cultures with a better agreement on red and yellow colors. The results contradict our hypothesis. The average mode values for the most frequent Munsell chips are lower (less agreement) than the average mode values for the rest of the Munsell chips. So our finding does not support the notion that the more of a particular color exists in the environment, the more agreement on the categories’ boundaries for that color among the people living there. The opposite seems to be true—more of the color equals less agreement, while less of the color leads to more agreement (Fig. 16.5). These findings are not what we expected. Nonetheless, it’s enlightening that there are significant differences between the colors that are encountered frequently and the rest of the colors. Apparently, the categorization process does depend on the physical features of the environment. However, that people agree more on the color category boundaries within the less frequent colors is striking. We suggest a few possible explanation of this finding: • Lower average mode values for the most frequent colors may indicate more linguistic labels (categories) for those colors. More linguistic labels for the most frequent colors could be predicted from our first assumptions about the role of the environment in the categorization process. In other words, people are better able to discriminate between colors, noticing even subtle differences between hues.

300

P. Lewinski et al.

Fig. 16.4 a ANOVA: Differences between the standardized average mode values of the most frequent colors (95%) and the rest of the colors in the biomes. b ANOVA: Differences between the standardized average mode values of the most frequent colors (80%) and the rest of the colors in the biomes

• Lower agreement for the most frequent colors is logical from a certain perspective. When speakers (presumably) communicate with each other more about the most frequently occurring colors, the constant use of those colors means the entire categorization process is constantly under review, flexible, adaptable and prone

16 The World Color Survey: Data Analysis and Simulations

301

Fig. 16.5 Munsell chart of colors, mode values for biome 1 and measures of agreement (mode) and disagreement (color standard deviation)

to change. If one variable is more prominent than others, then it perhaps evolves faster and is subject to more interpretation. For example, a popular poem by a wellknown artist would be subject to many more interpretations than a poem written by a high school student, which has been seen and interpreted only by the student and the student’s teacher. The more something is presented, the harder it might be to reach agreement on what it “is.” Likewise, the more of a particular color one sees in the environment, the less agreement on what exactly it is. • More frequently occurring colors may have different linguistic labels in different contexts. For example, adjectives associated with the color “white” will be quite different when describing a piece of jewelry as opposed to describing snow. However, it must be noted that in most of the biomes, there is merely a tendency for the non-frequent colors to have higher average standardized modes. Overall, there is a significant difference between the frequent and non-frequent chips (colors). There is a general tendency for the non-frequent colors to have higher means, but only three biomes show a clear pattern.

16.7.2 Simulation Study In the following modeling study, a population of autonomous agents is able to develop a repertoire of perceptually grounded categories that is sufficiently shared within the population to allow successful communication (Plewczynski et al. 2014). A shared repertoire of categories and a lexicon to express them is observed to emerge within the group. The categorical sharing is sufficient for successful communication; moreover, the semiotic dynamics can be observed and analyzed during computational simulations.

302

P. Lewinski et al.

We employed the agent-based model to analyze how a population of agents reacts to unequal distribution of perceptual data. We wanted to test whether regions that are more frequent yield higher average modes. We divided the Munsell palette into 2 subgroups (A, B) of stimuli. Afterwards, we considered environments with varying frequencies of stimuli from each of the parts. The options are with stimuli in group A being 10 times more frequent, 10 times less frequent or equally frequent as stimuli in group B. We applied a t-test for comparing the average modes from the groups with varying frequency of stimuli. We noticed no significant difference when comparing average modes on groups A and B when stimuli are equally frequent p  0.611. On the other hand, when the groups differed, we observed p  2.2e−16 < 0.01 and p  0.0022 < 0.01, indicating statistically significant differences. We conclude that regions with more frequency allow agents to reach better consensus with respect to naming. We observed a direct correlation between the frequency of the region in perceptual space and the average mode obtained by agents’ population during the simulation. This is the reverse of what was revealed in the study of human subjects. The modeling framework offers possibilities of further studies of how various distributions of stimuli across the environment can influence learning processes (for examples, see Fig. 16.6). Perhaps patterns with greater complexity can yield more phenomena in better agreement with human-based studies. We are currently working in this direction.

Fig. 16.6 Various stimuli applied to cognitive systems simulations, where different distributions of stimuli give rise to slightly modified categorization patterns for agent based modeling. Example complex colour distributions on WCS pallette. Brown colour denotes regions of frequent colours (e.g. 10 times more frequent than the blue background)

16 The World Color Survey: Data Analysis and Simulations

303

16.7.3 Discussion In this paper, we provide researchers with two novel methodological approaches to investigate color categorization. We also provide two sets of preliminary results, which could serve in the future as an inspiration for follow-up research. A limitation of our analysis is that results from human and computer frameworks are seemingly contradictory. A variety of reasons are offered to account for this contradiction, such as: (i) not enough testing of variable parameters of the machine learning algorithms—unorthodox selection and pre-processing of the natural images; (ii) qualitative differences in artificial (computers) and natural (human) categorization systems. In conclusion, we hope to have provided new tools with some preliminary results to investigate this topic further. An encouraging outcome of this study is that no difference was found between the types of biomes in their average modes of the most frequently occurring answers. The modes differ only within the type of the Munsell chips, which gives us confidence about the strength of our results. Such findings might support the notion that there are more universalistic tendencies on the general level (between the environments) and more relativistic tendencies on the local level (within the environments). However, there are additional features of our study indicating the results are not only significant in the statistical sense, but also methodologically solid. It must be noted that the significant differences we obtained were found even though the methods used were at best a rough approximation. The World Color Survey database is certainly not an ideal tool for establishing color category boundaries, but it’s the best researchers have at this moment. Additionally, the process of gathering and editing pictures is a gross estimation of the actual, real distribution of environmental colors, where we arbitrary choose the cut-off points at the 95 or 80% level. In theory, one could find another mathematical way of comparing color category boundaries, perhaps superior to ours (e.g., comparing the standardized average mode values of the type of Munsell chips), and such methods are currently a subject of scientific investigation. Comparing such complex distributions is not simple, and each time the method must be developed from scratches to answer questions of whether there are any differences or similarities between the distributions. Finding a statistically significant difference using such imperfect tools only adds to the strength of the results and the soundness of the methods used. Acknowledgements This work was supported by grants from the Polish National Science Centre (Grants number 2014/15/B/ST6/05082 and UMO-2013/09/B/NZ2/00121), and the European Cooperation in Science and Technology (COST BM1405 and BM1408). Michal Lukasik was supported by research fellowships within “Information technologies: research and their interdisciplinary applications” agreement POKL.04.01.01-00-051/10-00.

304

P. Lewinski et al.

Annex: Supplementary Materials See Table A.1 Table A.1 Characteristics of the cultures from WCS data archive on how they perceive colors No. Language name No. of Coordinates Biome Continent Ecoregion and location subjects 1

Abidji Ivory Coast Agarabi Papua New Guinea Agta Philippines

25

25 25

9

Aguacatec Guatemala Amarakaeri Peru Ampeeli Papua New Guinea Amuzgo Mexico Angaatiha Papua New Guinea Apinaye Brazil

10

Arabela Peru

2 3 4 5 6 7 8

11

12 13 14 15 16 17 18

5° 40 N, 4° 35 W 6° 10 S, 146°E

1

2

AT0111

1

3

AA0105

17° 58 N, 121° 50 E

1

1

IM0123

15° 25 N, 91° 20 W 12° 30 S, 70° 30 W 6° 45 S, 146° 5 E 16° 50 N, 98°W 7° 13 S, 146° 15 E

1

4

NT0303

1

5

NT0166

1

3

AA0105

2

4

NT0230

1

3

AA0120

30

5° 30 S, 48°W

1

5

NT0140

25

2°S, 75° 10 W

1

5

NT0142

142°

1

3

AA0115

24

6 27 25 25

35 S,

Bahinemo 25 Papua New Guinea Bauzi 25 Indonesia Berik Indonesia 25 (Irian Jaya)

4° 50 E

2° 30 S, 137° 30 E 2° 15 S, 138° 50 E

1

1

AA0116

1

1

AA0115

Bete Ivory Coast Bhili India

1

2

AT0111

25

6° 15 N, 6° 15 W 22°N, 73°E

2

1

IM0206

25

30 N,

1

4

NT0130

3

4

NT0303

1

5

NT0166

Buglere Panama Cakchiquel Guatemala Campa Peru

25

30 25

8° 81° 15 W 14° 30 N, 91°W 12°S, 74°W

(continued)

16 The World Color Survey: Data Analysis and Simulations Table A.1 (continued) No. Language name No. of and location subjects 19 20 21 22 23 24

25 26 27 28 29 30 31

305

Coordinates

Biome

Continent

Ecoregion

25

1° 10 N, 77°W

1

5

NT0121

11

4°10 S, 77°W

1

5

NT0142

Cavineña Bolivia Cayapa Ecuador Chácobo Bolivia Chavacano (Zamboangueño) Philippines

25

20 S,

7

5

NT0702

1

5

NT0178

Chayahuita Peru Chinantec Mexico Chiquitano Bolivia Chumburu Ghana Cofán Ecuador

Camsa Columbia Candoshi Peru

Colorado Ecuador Cree Canada

24

13° 66° 30 W 0° 40 N, 79°W

1

5

NT0166

25

12° 10 S, 66° 45 W 7°N, 122° 5 E

1

1

IM0129

25

5° 30 S, 77°W

1

5

NT0174

25

17° 50 N, 96° 1 30 W 17° 30 S, 60°W 2

4

NT0154

5

NT0212

2

AT0707

25

25

5

NT0142

25

7° 45 N, 0° 15 7 E 0° 10 N, 77° 1 10 W 1° S, 79°20 W 1

5

NT0178

25

25 20

54° N,74° W

6

6

NA0606

25

6°S, 70° 30 W

1

5

NT0166

33

Culina Peru, Brazil Didinga Sudan

25

1

2

AT0108

34

Djuka Surinam

25

4° 30 N, 33° 30 E 5°N, 54° 30 W

35

Dyimini Ivory Coast Ejagam Nigeria, Cameroon Ese Ejja Bolivia

25

32

36

37

25

25

1

5

NT0125

8° 25 N, 4° 25 W 5° 25 N, 8° 40 E

7

2

AT0707

1

2

AT0107

11°S, 66°W

1

5

NT0166 (continued)

306

P. Lewinski et al.

Table A.1 (continued) No. Language name No. of and location subjects 38

39 40 41 42 43 44 45 46 47 48

49

50 51 52

53

Coordinates

Biome

Continent

Ecoregion

Garifuna (Black Carib) Guatemala Guahibo Colombia Guambiano Columbia Guarijío Mexico Guaymí (Nga¨ bere) Panama

28

15° 40 N, 88°W

1

4

NT0111

25

5°N, 69°W

7

5

NT0709

27

2° 30 N, 76° 40 W 27° 45 N, 108° 40 W 8° 40 N, 82°W

1

5

NT0109

2

4

NA0302

1

4

NT0167

Gunu Cameroon Halbi India

25

7

2

AT0712

Huastec Mexico Huave Mexico

25 25

25 25 25

4° 35 N, 11° 15 E 21°N, 81°E

2

1

IM0201

22° 5 N, 99° 20 W 16° 13 N, 95°W 9° 20 S, 150° 15 E 16° 50 N, 121° 5 E

1

4

NA0303

2

4

NT0230

1

3

AA0125

1

1

IM0123

Iduna Papua New Guinea Ifugao (Keley-i) Philippines

25

Iwam (Sepik) Papua New Guinea Jicaque Honduras Kalam Papua New Guinea Kamano-Kafe Papua New Guinea Karajá Brazil

25

4° 20 S, 142°E

1

3

AA0115

10

14° 40 N, 87°W 5° 15 S, 144° 35 E 6° 15 S, 145° 40 E

3

4

NT0303

1

3

AA0105

1

3

AA0105

10°S, 50° 20 W 2° 40 S, 140° 20 E

1

5

NT0704

1

1

AA0115

20° 45 N, 73° 30 E

1

1

IM0134

25

25 25

19

54

Kemtuik 25 Indonesia (Irian Jaya)

55

Kokni (Kokoni) 25 India

(continued)

16 The World Color Survey: Data Analysis and Simulations Table A.1 (continued) No. Language name No. of and location subjects 56 57 58 59

Konkomba Ghana Kriol Australia Kuku-Yalanji Australia Kuna Panama

307

Coordinates

Biome

Continent

Ecoregion

25

10°N, 0° 5 E

7

2

AT0722

25

7

7

AA0704

20

14° 50 S, 135°E 16°S, 145°E

1

7

AA0705

25

8°N, 77° 20 W

1

4

NT0115

1

1

AA0115

7

2

AT0705

7

2

AT0722

1

3

AA0105

13

7

AA1304

1

1

IM0134

1

5

NT0166

3

4

NA1302

15 S,

60

Kwerba 25 Indonesia (Irian Jaya)

2° 30 E

61

Lele Chad

15

62

24

65

Mampruli Ghana Maring Papua New Guinea Martu Wangka Australia Mawchi India

66

Mayoruna Peru 25

67

1

4

NT0146

25

1

3

AA0120

25

45°N, 63°W

4

6

NA0410

25

32°N, 85°W

4

6

NA0413

72

Mazahua Mexico Mazatec Mexico Meny(a/e) Papua New Guinea Micmac Canada Mikasuki United States Mixtec Mexico

9° 5 N, 15° 35 E 10° 20 N, 0° 40 W 5° 30 S, 144° 40 E 21° 30 S, 126°E 21° 20 N, 73° 40 E 5° 25 S, 73° 15 W 19° 25 N, 99° 55 W 18° 15 N, 96° 50 W 7° 15 S, 146°E

25

3

4

NT0309

73

Mundu Sudan

18

7

2

AT0712

74

Múra Pirahá Brazil Murle Sudan

25

17° 15 N, 98° 35 W 4° 20 N, 30° 29 E 7°S, 62°W

1

5

NT0135

7

2

AT0705/AT0905

63 64

68 69

70 71

75

25 25 25

25 25

25

138°

6° 30 N, 33° 30 E

(continued)

308

P. Lewinski et al.

Table A.1 (continued) No. Language name No. of and location subjects 76 77 78 79 80

81 82 83 84 85

Murrinh-Patha Australia Nafaanra Ghana Nahuatl Mexico Ocaina Peru Papago (O’odham) United States, Mexcio Patep Papua New Guinea Paya Honduras Podopa Papua New Guinea Saramaccan Surinam Seri Mexico

25 29 6 25 25

24 20 14 25 25

Biome

Continent

Ecoregion

14° 40 S, 129° 40 E 8°N, 2° 35 W

7

3

AA0706

1

2

AT0111

3

4

NT0310

1

5

NT0132

13

6

NA1310

1

3

AA0120

1

4

NT0303

1

3

AA0122

1

5

NT0125

13

4

NA1310

19°N, 98° 15 W 2° 45 S, 71° 45 W 32°N, 112°W

6° 55 S, 146° 35 E 15°N, 85° 30 W 7°S, 144° 30 E 4° 30 N, 55° 30 W 29°N, 112°W

Shipibo Peru

1

5

NT0166

87

Sirionó Bolivia 25

15° 35 S, 64°W 7

5

NT0702

88

Slave Canada

67°N, 125°W

6

6

NA1113/NA0614

89

Sursurunga 26 Papua New Guinea Tabla Indonesia 25 (Irian Jaya)

4°S, 152° 46 E

1

3

AA0111

2° 27 S, 140° 25 E

1

1

AA0116

Tacana Bolivia

13° 30 S, 68°W 1

5

NT0166

3

4

NA0302

91

24

8

7°

30 S,

86

90

25

Coordinates

75°W

30 N,

92

Tarahumara 9 (Central dialect) Mexico

27° 30 W

93

Tarahumara 6 (Western dialect) Mexico

27° 30 N, 108°W

3

4

NA0302

94

Tboli Philippines

6° 10 N, 124° 30 E

1

1

IM0129

25

107°

(continued)

16 The World Color Survey: Data Analysis and Simulations Table A.1 (continued) No. Language name No. of and location subjects 95

Teribe Panama

26

96

Ticuna Peru

25

97

100

Tifal Papua New Guinea Tlapanec Mexico Tucano Colombia Vagla Ghana

25

101

Vasavi India

25

102

Waorani 25 (Auca) Ecuador

103

Walpiri Australia Wobé Ivory Coast Yacouba Ivory Coast Yakan Philippines

25

Yaminahua Peru Yucuna Colombia Yupik United States Zapotec Mexico

98 99

104 105 106 107 108 109 110

309

Coordinates

Biome

Continent

Ecoregion

9° 20 N, 82° 40 W 4°S, 70° 30 W

1

4

NT0167

1

5

NT0166

E 1

3

AA0105

20

25

5°S, 141°

25

17° 5 N, 99°W

3

4

NT0309

25

0° 30 N, 69° 10 W 9° 25 N, 2° 25 W 24° 30 N, 71° 30 E 1°S, 76° 30 W

1

5

NT0132

7

2

AT0722

13

1

IM1303

1

5

NT0142

13

7

AA1304

1

2

AT0130

27

20°S, 132° 20 E 7° 25 N, 7° 20 W 7° 30 N, 8°W

1

2

AT0114

25

6° 30 N, 122°E 1

1

IM0129

25

8°S, 73°W

1

5

NT0166

25

0°45 S, 71°W

1

5

NT0107

25

65°N, 173°W

11

6

PA1104

25

16° 25 N, 95°W

2

4

NT0230

25

Legend • No. number of the WCS language • Language name and location language name and which country it comes from • No. of subjects number of speakers doing color naming task in WCS study • Coordinates geographical coordinates, latitude and longitude • Biome see Table A.2 • Continent see Table A.3 • Ecoregion Ecoregion classification based on Olson et al. (2001)

310

P. Lewinski et al.

Table A.2 World Wildlife Fund classification of world biomes ID Name of biome ID Name of biome 1

5

Tropical and Subtropical Moist Broadleaf Forests Tropical and Subtropical Dry Broadleaf Forests Tropical and Subtropical Coniferous Forests Temperate Broadleaf and Mixed Forests Temperate Conifer Forests

6

Boreal Forests/Taiga

13

7

Tropical and Subtropical Grasslands, 14 Savannas and Shrublands

2 3 4

8 9

Temperate Grasslands, Savannas and Shrublands Flooded Grasslands and Savannas

10

Montane Grasslands and Shrublands

11

Tundra

12

Mediterranean Forests, Woodlands and Scrub Deserts and Xeric Shrublands

Table A.3 Continent which WCS language comes from ID Continent ID 1 2 3 4

Asia Africa Oceania Central America

5 6 7

Mangroves

Continent South America North America Australia

References Baronchelli, A., Gong, T., Puglisi, A., & Loreto, V. (2010). Modeling the emergence of universality in color naming patterns. Proceedings of the National Academy of Sciences, 107(6), 2403–2407. Berlin, B., & Kay, P. (1969). Basic color terms: their universality and evolution. Berkeley, CA: University of California Press. Blakemore, C., & Cooper, G. F. (1970). Development of the brain depends on the visual environment. Nature, 228(5270), 477–478. Boroditsky, L. (2001). Does language shape thought?: Mandarin and English speakers’ conceptions of time. Cognitive Psychology, 43(1), 1–22. Cook, R. S., Kay, P., & Regier, T. (2005). The world color survey Database: history and use. In H. Cohen & C. Lefebvre (Eds.), Handbook of categorisation in the cognitive sciences (pp. 223–242). Amsterdam and London: Elsevier. Dryer, M. S., & Haspelmath, M. (2011). The world Atlas of language structures online. Munich: Max Planck Digital Library. Retrieved from http://wals.info/. Gärdenfors, P. (2000). Conceptual spaces: The geometry of thought. Cambridge: The MIT Press. Hein, A., Held, R., & Gower, E. C. (1970). Development and segmentation of visually controlled movement by selective exposure during rearing. Journal of Comparative and Physiological Psychology, 73(2), 181–187. Heider, E. A. (1972). Probabilities, sampling and ethnographic method: The case of Dani colour names. Man, 7, 448–466. Kay, P., & Kempton, W. (1984). What is the sapir-whorf hypothesis? American Anthropologist, 86(1), 65–79.

16 The World Color Survey: Data Analysis and Simulations

311

Kay, P., & Regier, T. (2003). Resolving the question of color naming universals. Proceedings of the National Academy of Sciences, 100, 9085–9089. Kay, P., & Regier, T. (2006). Color naming universals: The case of berinmo. Cognition, 102(2). https://doi.org/10.1016/j.cognition.2005.12.008. Laboratory for Anthropogenic Landscape Ecology. (2010). Terrestrial biomes [Data file]. Retrieved from http://ecotope.org/anthromes/maps. Leventhal, A. G., & Hirsch, H. V. (1975). Cortical effect of early selective exposure to diagonal lines. Science, 190(4217), 902–909. Martin, L. (1986). “Eskimo words for snow”: A case study in the genesis and decay of an anthropological example. American Anthropologist, 88(2), 418–423. National Geographic. (2001). [Natural image of the terrestrial ecoregions of the world for all 867 land-based ecoregions on the planet]. National Geographic Data Access from nationalgeographic.com Wild World in collaboration with World Wild Fund (WWF). Retrieved from http://www.nationalgeographic.com/wildworld/profiles/photos/. Olson, D. M., Dinerstein, E., Wikramanayake, E. D., Burgess, N. D., Powell, G. V. N., Underwood, E. C., et al. (2001). Terrestrial ecoregions of the world: A new map of life on earth. BioScience, 51(11), 933–938. Plewczynski, D., Łukasik, M., Kurdej, K., Zubek, J., Rakowski, F., & R˛aczaszek-Leonardi, J. (2014). Generic framework for simulation of cognitive systems: A case study of color category boundaries. Man-Machine Interactions special issue of Advances in Intelligent Systems and Computing, 242, 385–393. https://doi.org/10.1007/978-3-319-02309-0_42. Pullum, G. K. (1989). The great eskimo vocabulary hoax. Natural Language & Linguistic Theory, 7(2). Retrieved from http://www.jstor.org/pss/4047733. Roberson, D., & Davidoff, J. (2000). The categorical perception of colors and facial expressions: The effect of verbal interference. Memory and Cognition, 28, 977–986. Roberson, D., Davidoff, J., Davies, I. R. L., & Shapiro, L. R. (2005). Color categories: Evidence for the cultural relativity hypothesis. Cognitive Psychology, 50, 378–411. Roberson, D., Davies, I. R. L., & Davidoff, J. (2000). Color categories are not universal: Replications & new evidence from a Stone-age culture. Journal of Experimental Psychology: General, 129, 369–398. Searle, L. (1994). Pierce, Charles Sander. In M. Groden & M. Kreiswirth (Eds.), The John Hopkins guide to literature theory and criticism (pp. 560–562). Baltimore and London: John Hopkins University Press. Steels, L. (1998). The origins of syntax in visually grounded robotic agents. Artificial Intelligence, 103(1–2), 133–156. Steels, L. (2001). Language games for autonomous robots. IEEE Intelligent Systems, 16(5), 16–22. Steels, L. (2002). Language games for emergent semantics. IEEE Intelligent Systems, 17(1), 83–85. Steels, L. (2003). Evolving grounded communication for robots. Trends in Cognitive Sciences, 7(7), 308–312. Steels, L. (2006). Semiotic dynamics for embodied agents. IEEE Intelligent Systems, 21(3), 32–38. Steels, L., & Belpaeme, T. (2005). Coordinating perceptually grounded categories through language: a case study for colour. Behavioural and Brain Sciences, 28(4), 469–489; discussion 489–529. Steels, L., & Kaplan, F. (1999). Collective learning and semiotic dynamics. Advances in Artificial Life, Proceedings, 1674, 679–688. Wellens, P., Loetzsch, M., & Steels, L. (2008). Flexible word meaning in embodied agents. Connection Science, 20(2–3), 173–191. https://doi.org/10.1080/09540090802091966. World Color Survey. (2003). WCS data archives [Data file]. Retrieved from http://www.icsi. berkeley.edu/wcs/data.html. World Wildlife Fund. (2011a). The terrestrial ecoregions database [Data file]. Retrieved from http:// www.worldwildlife.org/science/ecoregions/item1267.html. World Wildlife Fund. (2011b). Terrestrial ecoregions base global dataset [Data file]. Retrieved from http://www.worldwildlife.org/science/data/item1874.html.

Chapter 17

Cognitive Meaning: Review of the Concepts of Imagination, Image Schema and Mental Image and Consequences on the Conceptualization of Emotions Maria Antònia Font Fernández

Abstract Contrary to what was advocated during the last century, today we know that cognition is not atomistic but has gestalt and imaginative properties. Central nervous system is not involved in preparing the mind evenly, but the various regions involved have some guidelines that interconnect very complex signals. Particularly, knowledge is understood as a journey that uses neural imagination as a basis and is sophisticated in mental imagination. The first stage of knowledge is linked to pre-linguistic cognition and perception (Psychology), neural images (Neurology), image schemes (Philosophy and Psycholinguistics) and emotional themes (Psychology). However, the second stage of knowledge is linked to linguistic cognition and conceptualization (Psychology and Linguistics), mental images (Psychology and Linguistics) and cognitive-emotional variants (Psychology). In other words, reality, experience and knowledge are conceptualized as an emotional journey (theme) that are gradually formed until the feeling (variation). Thus, imagination can be understood as the cognitive-emotional process that, from image schemes and neural images (perception) leads to mental images (conceptualization).

17.1 A Little Introduction. Imagination Is not the Same that Fantasy Nowadays, Romanticism still has a consolidated residue in Western cultures: the idea or image that comes to mind when we evoke the concept of author is ‘the writer who has published at least one original work, usually fiction’. We might wonder if a work can be oral, collective, about investigation, unpublished, and how much must it be M. A. Font Fernández (B) Universitat de Barcelona, Barcelona, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_17

313

314

M. A. Font Fernández

original (not so much in the sense of ‘peculiar’ as in the sense of ‘not copied’): the bond between contemporary production and copyright are issues that we inherited from the 19th. In the previous paragraph, I probably have evoked the concept of imagination. It is said that someone has a lot of imagination to refer to the ability of that person to invent fantastic stories, often with pejorative sense (remember that, etymologically, fables are ‘lies’). Despite that, creativity and imagination constitute only one aspect of imagination, so there is a metonymical connection between imagination and fantasy. Precisely, there are several capacities implicated by nervous central system: sensitivity, motion, cognition, memory, attention and, in effect, imagination and fantasy. Therefore, fantasy must not be confused with imagination: fantasy shaped reality that is replaced intentionally. However, imagination is a quality of human cognition (2000: Chap. 6, note 15), as Gestalt Psychology advocated. Castilla del Pino extracts a sample of language that illustrates that people interpret reality shaping images: «Por la cara que trae, me figuro que […]» (2000: Chap. 9, note 18).

17.2 The Philosophy Approach: Sensitive Knowledge Sensitive and Intellectual Knowledge In The Body in the Mind, Mark Johnson noted that there was no theory about speakers understand as imagination. That is, the ability to shape the experience, to extract something new from old or to project comprehensively in the position of another (1987: 223). However, a few centuries of philosophical speculation have been to pass to reach the cognitive notions of imagination and image. Especially because they could become essential in linguistic conceptualization. Plato set two stages of knowledge. In the sensitive world, there were individual and mutable demonstrations of concepts. At this stage, common for all animals, knowledge was sensitive, so that it was learned through sense organs. On the contrary, in the world of ideas there were universal, rational and paradigmatic ideas. Those are independent from sensitive world and are exclusive to human. In the Platonic tradition, imagination generated only sensitive knowledge: that knowledge was considered a delusion of senses, since the reality (objective) is constantly changing. True knowledge, therefore, forced to fight appearances to capture immutable ideas or concepts, a process that involved mental ability. Later, Aristotle replied to his master’s theory arguing that postulating two separate worlds unnecessarily duplicated reality. For him, knowledge was given the other way round and in an unique process: people are not born with a tabula rasa, but we infer concepts. About imagination, Aristotle abandoned the idea of Plato, according to which it was a chaotic and animal capacity. So, it had to be controlled; otherwise, imagina-

17 Cognitive Meaning: Review of the Concepts of Imagination, Image …

315

tion constituted a cognitive operation that linked sensory observation (perceptions) with reasoning language (concepts). This operation could store sensory impressions produced by objects in memory (which is another cognitive ability). From this point of view, knowledge consisted precisely on obtaining a mental impression of objects of reality. So, we say that we are affected when we are extremely moved by an experience and becomes momentarily fixed in the consciousness. In the 17th, Descartes still maintained the duality by which body and mind develop distinct functions. He also argued that the body had no role in human reasoning, so the mind was not embodied, and knowledge consists on capturing mental representations. The problem was that if we know only own ideas, we can not be sure of achieving the truth: the unconvincing solution of this philosopher was claiming that God ensures the connection between the mind and objective reality. In the 18th, Kant reviewed the rigid dichotomy that enclosed the material and perceptive component (the sensitive world of Plato) from the formal and conceptual component (the world of ideas of Plato). In other words, between corporeal processes and comprehension. And that had to be intellectual and categorical. As for imagination, it was understood by Kant as the ability to construct mental images (concepts) from image schemes (sensory impressions). However, Kant failed to overcome the weight of the tradition: although he realized that imagination had a dual nature, so that was both sensitive and intellectual, theory’s Kant was not successful.

17.3 The Psychology Approach. The Contribution of the Gestalt School The strictly cognitive notion of image schema has its origins in the Critique of Pure Reason write by Kant (1781). Particularly in the faculty of imagination, that is, in the in-formation (literally, ‘form inside’). For this philosopher, image schemes were the structures that connected perceptions with the related concepts. Thus, Kant understood, for example, that sense of balance is part of perceptual knowledge. For Cognitivism, image schemes are perceptual structures (gestalts) underlying conceptualization. However, there is no agreed definition: on the contrary, as it is a fundamental notion, there are much discussion around this question. Thus, to cite some authors, Mandler (1992) worked from Evolutionary Psychology on what he called image schema. He started from the assumption that both the evolution of the species and the individual evolution takes part in the construction of knowledge. In 2004, Mandler published a work that applies image schemes to the development of linguistic conceptualization. Also Kövecses has contributed to elucidate the relation between image schemas and the construction of knowledge. The research of Kövecses consists on finding the prototypical structure of emotional concepts, focusing on metaphorical operations.

316

M. A. Font Fernández

Especially, he studies fear and anger, emotional categories that are more prototypical than for instance sadness. Moreover, according to Lakoff, image schemes are the more basic idealized cognitive model: his definition recalls the Gestalt Psychology (which must be pointed out, means ‘form’ in German): «Paths, bounded regions and contact are cognitive topological concepts, commonly referred to as image schemes […]. (Image schemes) are topological in the sense that they generalize over geometry in virtue of preserving neighborhood relations» (1989: 114). However in cognitive linguistics the more accepted and used definition of image schema goes on being the Johnson’s proposal: «A recurring, dynamic pattern of our perceptual interactions and motor programs that gives coherence and structure to our experience […]. These patterns are embodied and give coherence, meaningful structure to our physical experience at a pre-linguistic level» (1987: xiv, 13). Johnson postulated the following 27 image schemes, which I am not going to deal: container, blockage, enablement, path, cycle, part-whole, fullempty, iteration, surface, balance, counterforce, attraction, link, nearfar, merging, matching, contact, object, compulsion, restraint removal, mass-count, center-periphery, scale, splitting, superimposition, process and collection. And from the definitions of Lakoff and Johnson, Peña (2003: 54) assumed that image schemes are organized hierarchically. This observation can quite remember, in cognitive linguistics, the idea of a category with prototypically members and the idea of a radial category with prototypical effects. So, Peña proposed an integrated definition of image schema: «An image-schema is a recurring pattern of experience which is abstract and topological in nature. Moreover, it can also be an eventive pattern in the sense that it takes place in space but is not necessarily identified with space itself» (2003: 42). In any case, the characteristics of image schemes are: (1) Image schemes belong to pre-linguistic cognition: they appeal to earlier experiences of the evolution and the cognitive-emotional development. (2) Image schemes are not propositional (unlike cognitive domains) or linguistics (unlike expressions), but they are previous to mental images and verbalization; another issue is that labels are needed to refer to that. They are neither symbolic abstractions (unlike mathematical symbols) or generic concepts (unlike cause or result) or specific and complete images: we must not confuse experience with the construction of the experiential reality. (3) As surfaced from direct interaction with reality experiential, image schemes are significant per se, and not vehicles of meaning. That means, image schemes constitute perceptions, the most basic meanings, also known by animals. Therefore, image schemes are corporeal: these are patterns with which people structure experiential reality from perceptual, motor and sensory experiences. In other words, from movements in relation to space and with the manipulation of objects in the environment.

17 Cognitive Meaning: Review of the Concepts of Imagination, Image …

317

(4) Image schemes are structured like a gestalt: people do not discriminate consciously the parts of a whole: they are so closely related that it is always perceived as a unit. In this sense, it can be remembered the fractal dimension: there are many objects with a similar structure. A fractal is «a geometric shape that has a fine structure at every scale» (Massip 2012: 38). Moreover, as a complex system, an image schema is not compositional: «hierarchical complex systems contain more intense interactions within a subsystem than between subsystems» (íd). (5) Simplicity is a continuous parameter. For this reason there are schemes more general (for example, object) and other more complex (for example, fusion). The gradual nature of simplicity explains there are experiences that involve the joined and simultaneous perception. Not only the parts of a whole, but even several image schemes at the same time (actually, time is an important element for complexity). Johnson (1987) called this phenomenon superimposition of image schemes, and Hampe (2005) adopted the term composition of image schema. An applied study of this phenomenon is the work of Langacker (1987) which approach the merger of three schemes: object, path and container. (6) Finally: image schemes are dynamic. They take place in time and over time, as any complex matter. They are also recurrent, so as they let us apprehend a variety of different experiences that start from the same pattern: meanings are known or ascribed from basic outlines of image. For example, abovebelow expresses position, whereas up-down expresses movement, as cognitive linguistics defended in the dynamic conceptualization.

17.4 The Neurology Approach. The Metonymical Basis of Emotional Cognition The process of attribution of meaning does not start from conceptualization but from neural imagination which considers the representation of the neurophysiological experience in the brain. Thus, the purpose of neural image is reflecting in the brain the interaction of signals with which pre-linguistic cognition is formed. Neural stimulus or sensor and motor nerves carry an enormous amount of data. Those are susceptible to become basic meanings. That is what saving a few distances Neurology calls neural images and Psychology image schemes. That means that «Complex systems can range from the social level to the neurological level from astronomy to biology» (Massip 2012: 36). Damásio (1994 and 2010) investigated how image schemes operate into the brain. It is enough to say that Turner (1996: 22) had already formulated the same principle but noted the difficulty of locating image schemes in the brain. So he came to the conclusion that it was not clear how could be related to neural images. The neural theory departed from the hypothesis that neural images constitute what Damásio calls dynamic topographical representations, which people are able to perceive at the same time experiences such as the vertical and the quantity (see § 3).

318

M. A. Font Fernández

So, today it is believed that neural activity is not located on a rigid or static map but in areas relatively separate but coordinated: the brain has always a representation or image of what happens within the boundaries of the body either an injury or the order to execute a movement (for example scratching). In the process of neural imagination, peripheral nervous system moves sensory and motor nerves from any part in the body, joint or internal organ to the somatic nervous system. Which integrates the volunteer processes, with dermal conductance as the preeminent response. So, when peripheral nerves reach the autonomic nervous system, they send signals in order to do the needed state of the body. And very often skeletal muscles complete a cognitive response through the appropriate facial expression and a body posture also communicative. As Cognitivism advocates, most of neural (and chemical) interactions occur within the limits of the body. In the theory of conceptual metaphor, Lakoff and Johnson believe that what they call basic experiences and primary metaphors are two different cognitive domains, but the conventional use has fixed words that are associated with both experiences. This is the case of the adjective high. Grady (1997) formulated the Reversibility Principle (1997). By this principle, the relation between two contiguous and simultaneous variables can be interpreted in both directions of cause-effect. I deduce some conclusions: first, if people can perceive two image schemes at the same time, it must be possible that two experiences can be noticed together. Second, all cognitive functions (basic and superior) work at the same time. In other words, the attribution of basic meaning (neural imagination and cognitive-emotional themes) and the attribution of complex meaning (mental imagination and cognitive-emotional variants) are partially inseparable. And the thing is that experiences such as verticality and quantity are perceived at the same time thanks to neural images. So, people can understand there are more sand in a place because it takes up more space or that sand takes up more space because there is more quantity of it. In the same manner, from mental cognitivism, anger is defined by the conceptual metaphor anger is fire. However, neural cognitivism conceptualizes anger (understood as a cognitive-emotional experience) from its physiological experience (fire). This last process is identified as a metonymy. The neural or sensory and motor stimulus meanings through perception: the verb perceive comes from the Latin per capio, which means give meaning to a stimulus. That is, attributing meaning to a sensation, which becomes perception. Then, perception can be defined as the cognitive process or function in which the thalamus, through attention, assesses sensory stimulus that penetrate in the experiential reality: perceiving consists on imaging, in other words, on shaping. However, not all the sensations become perceptions: in order to prevent that stimulus collapse cognitive resources, thalamus secretes those sensations that are potentially useful to allow living beings to inform the immediate experiential reality. And in this way to get the guidelines to respond: reality is neither objective nor subjective, so that what we know is what we perceive. And what we perceive is the image or representation that the nervous system informs from the perception.

17 Cognitive Meaning: Review of the Concepts of Imagination, Image …

319

To sum up, attention gives prominence to some stimulus detriment others. For example, in pre-linguistic cognition, heavy and sudden noises, so as moving objects to achieve a significance that allows them to function as figures against a background of silence or stillness respectively. That is why, from my point of view, the basis of cognition is metonymical: not all the sensations become perceptions. So, if sensations constitute a whole, only a part, that are the stimulus filtrated by the thalamus, become perceptions, neural images, image schemes. And, as images, perception involves the segregation of stimulus in a figure (part) and a background (whole). In next table, it is summarized the dynamic relation between basic meaning and complex meaning from the hypothesis that the basis of language is metonymical: Sensation WHOLE

>

Perception

basic meaning

PART

Perception WHOLE

>

Concept

complex meaning

PART

17.5 The Mental Image of Psychology and Linguistics. The Metonymical Basis in Linguistic Conceptualization Talmy, who thinks that language is a window that allows us to discover the human conceptual systems, adapted the notions of figure and background from Gestalt Psychology to cognitive linguistics (2000). And Langacker (1985) defined a mental image as the result of understanding a meaning and applying it to an experience, which can be conceptualized in diverse ways. Therefore, each perspective is a mental image. Returning to the idea of perceiving, Johnson (1987) defined this process as the cognitive operation that consists on a selection, in a conceptual structure, of a figure and a background. That means people do not form any mental image in which the figure and the background are neutral, since attention, also in linguistic cognition, gives prominence to some aspects over others. That is, in the essentially dynamic conceptualization accepted nowadays, people emphasize images from an experiential reality that is always moving. And, at this stage, attention is responsible for discriminating some components from conceptual or linguistic experience that are more significant not to themselves but compared to other semantic components. In my opinion, the linguistic symbol is not reality (experiential), but a way to reference, this is a metonymy; but it is also a representation, and thus a metaphor. In

320

M. A. Font Fernández

short, it seems that linguistic conceptualization is metonymical, while language, as a code or a set of symbols, is metaphorical. We can say that, behind cognition, there is a metaphor the representation for the action; that means, which occurs is an indirect representation: if perception involves a neural image, this image must appear in figure and ground, like mental images, and I understand that this is a part-whole relationship, a metonymy. However, in the same way that the meaning of the perceptions is in the image that the brain informs from neural images, the meaning of the concepts is in the mental image represented by us. What is conceptualized is not the cause of the constructed cognitive-emotional process but the associated meaning. Thus, in pre-linguistic cognition, perception and attention work together to form basic senses or meanings. Otherwise, in linguistic cognition, the patterns of shaping mental images (imagery), that are also called semantic construal, reflect the higher cognitive ability to conceive an experience from different perspectives, that is, applying attention in diverse ways. The results are different concepts for the same idealized cognitive model. In other words, patterns of mental images shaping allow us to form experiential reality, that is, in-form reality distributing attention in varied ways. These patterns must be associated with the human capacity to form different images (as Gestalt Psychology had demonstrated) and to generate mental images from image-schemes.

17.6 Emotional Themes and Emotional Changes. Relationship Diagrams and Neural Image Pictures and Mental Images Traditionally, it has been distinguished between primary emotions (which have similarities with the basic or pre-linguistic meanings) and secondary emotions or feelings (which have points of contact with conceptual or linguistic meanings). It is accepted that primary or basic emotions are happiness, sadness, anger, disgust, fear and surprise. The question is: that are concepts or words? Concepts and words are different things, despite various labels can be used. I understand a feeling as the part of a cognitive-emotional process that is subdued by superior cognitive functions (that is, conceptualization and language) and it is labeled. The problem is that human subjectivity and sociocultural variation block biunivocal correspondences between emotional categories, linguistic concepts and specific terms of a language. We must distinguish an experience of sadness (which belongs to attribution of basic meaning) from the labeling of experiences of sadness (which belongs to attribution of complex meaning). So, emotions can be labeled with different grades of precision: because of human subjectivity and sociocultural conventions, some but not all concepts are labeled. Moreover, there are variables between languages: for example, gratitud (grattitude) describes a recognition of a

17 Cognitive Meaning: Review of the Concepts of Imagination, Image …

321

favor, whereas ingratitud (ingrattitude) is referred to an evaluation about the behavior of somebody (Marina 1999: 253 and 325). Because the adaptation of the species is a result of evolution, Ekman (2004 [2003]: 44–46) thought about what evolution is and if it is a biological or an anthropological matter. The idea was that each emotion is associated with a basic and hereditary schema (emotional theme) enriched with the sociocultural context and the evocation and the rational learning (the variation or cognitive-emotional variants). In fact, it is accepted by complexity perspectives that genes and cultural are a link between complex and related parts. Meanwhile, what Ekman treated as themes and variations or variants, Damásio (1999 [1994]) explains it in terms of pre-organized emotions and organized emotions. Like Ekman, this neurologist argues that the first ones share a neurophysiological action program and are part of innate knowledge established in the genome and may not be linguistically defined (2010: 196). Therefore, we can say that the basic schema or the theme of an emotion consists in a substrate that is innate, physiological, sensory, motor and expressive. That substrate is developed with the maturation of the neurophysiological structures of the individual. Note that this definition coincides with cognitive specialists (like Piaget) have proposed for image schema. Thus, sadness (as a theme) focuses on a whole family of variants, including sadness and experiences such as grief, despair and sorrow. This inclusive connection (a variant respect to one or more themes) recalls the relationship postulated by Plutchik between a basic and a secondary emotion or feeling, with the difference that compositional emotions admitted an analysis in terms of necessary and sufficient properties, what structural and generative semantics do. I think that the definition of image schema of Johnson (1987 (see § 3)) has resonances of the concept of emotional theme or organized emotion: «A recurring, dynamic pattern of our perceptual interactions and motor programs that gives coherence and structure to our experience […]. These patterns are embodied and give coherence, meaningful structure to our physical experience at a pre-linguistic level» (Johnson 1987: xiv, 13). Moreover, «Emotions are complex programs of actions, usually automatic, elaborated by evolution. Actions are supplemented with a cognitive program, but the world of the emotions tends to be a world of actions that we carry out in our bodies, from facial expressions and postures to changes in entrails» (Damásio 2010: 175). Thus, the Aristotle’s position, for whom knowledge does not start from a tabula rasa but from the apprehension of specific data (see § 2) is consistent with the fact that cognitive-emotional variants (the traditional feelings) are not entities completely differentiated from emotional themes (the traditional emotions), but themes and combinations of themes generate conceptual and cultural variations. In addition, the fact that image schemes belong to pre-linguistic cognition, so, they appeal to the earliest experiences of evolution and development, coincides with the fact that the matter of schemes is the biological and affective knowledge underlying emotional. Even more, the sensitive matter of Philosophy coincides with what Psychology calls pre-linguistic cognition, while the world of ideas is consistent with the result

322

M. A. Font Fernández

of linguistic cognition. And what Kant understood as sensory impressions or image schema coincides with what we understand today as neural imaging. Then, rather than corporeal, image schemes are neural images, and from known neural images we can attribute basic meanings: we may say that image schemes constitute a frame whereas neural images constitute the content or substance of image schemes. Furthermore, in the 17 Th, Spinoza had already found that, in what today we call neural images, increases happiness and moves and accelerates the reactions or responses while sadness operates inversely. That made that Spinoza understood sadness as any negative modification of the state of the experimenter subject, as he postulated only three basic emotions: happiness, sadness and desire (Marina 1999: 286). The table below summarizes the contrast between the cognitive-emotional process of sadness with the process typically antonym, the happiness, illustrates it better (Font 2013: 88): Velocity of reasoning and of neural imagination

Wealth of reasoning and of neural imagination

Dysfunction: euphoria

Cognition is fast and reduces the attention dedicated to the stimulus or the cause. Positive values of attribution of meaning are prevalent

The associative process is quite rich. Profusion links to exploratory behaviors are not efficient but fast

Dysfunction: depression

Cognition is slow and benefits repetitive ideas that turn around the stimulus or cause. Negative values of attraction of meaning are prevalent

Reasoning is inefficient, with only a few inferences and the concentration on some obsessive reasons, which tend to promote negative reactions

17.7 Conclusions All nature is moving, so the whole experience is emotion, and all the knowledge or cognition has an emotional basis. Emotion is not an abstract matter but the matter that moves cognition. The key is distinguishing physiological change from neural change. In other words, people do not need externalize changes to produce an emotion: emotion and cognition are the same program, the same dynamic process and not all the time that knowledge is (re)created there are visible physiological changes. There are, however, neural and often chemical changes: emotion and knowledge are moving and the more elementary agents of motion are nerves. More precisely, cognition is imaginative, both in the pre-linguistic stage and in the linguistic stage too. On the one hand, the first Cognitivism, that was mental, started

17 Cognitive Meaning: Review of the Concepts of Imagination, Image …

323

from the idea that the basis of cognition is corporeal, which gave primacy to mental images. But, nowadays, the neural Cognitivism, argues that the basis of imagination is neural. An idea that is reinforced by neural images from Neurology. On the second hand, the linguistic stage: imagination is mental although it necessarily has a neural and corporeal basis. And in both stages, the faculties of attention and perception are vital since they ensure the segregation of stimulus in figure and background. As Gestalt Psychology had already advocated (although in the linguistic stage linguists usually refer to conceptualization). At the same time, knowledge in the pre-linguistic stadium is basic, biological and affective, and related to traditional emotion. Whereas knowledge in the linguistic stadium is more complex and with the presence of sociocultural and learning factors; in this case, the basic of emotion is specialized in what I have called cognitiveemotional variants, following Ekman (2004 [2003]).

References Castilla del Pino, C. (2000). Teoría de los sentimientos. Barcelona: Editorial Tusquets. Damásio, A. (1994). Descartes’ Error. New York: Editorial Putnam Book. (Used the translate into Spanish El error de Descartes (1999). Barcelona: Editorial Crítica). Damásio, A. (2010). Set comes to mind: constructing the conscious mind. New York: Panteon Books. Used the translate into Spanish Y el cerebro creó al hombre: ¿cómo pudo el cerebro generar emociones, sentimientos, ideas y el yo? (2010). (Barcelona: Editorial Destino). Ekman, P. (2003). Emotions revealed: Recognizing faces and feelings to improve communication and emotional life. New York: Times Books. Used the translate into Spanish ¿Qué dice ese gesto? (2004). Barcelona: RBA. Font, M. A. (2013). Aspectes de semàntica cognitiva en la construcció dels procesos emocionals, Llengua, societat i comunicació, 11. (Barcelona: Universitat de Barcelona). Grady, J. (1997). Foundations of meaning: Primary metaphors and primary scenes. Berkeley: University of Califòrnia. Hampe, B. (2005). From perception to meaning: Image schemes in cognitive linguistics. Cognitive Linguistics Research, 29. (Mouton de Gruyter). Johnson, M. (1987). The bodily basis of meaning, imagination and reason. Madrid: Editorial Debate. Lakoff, G. (1989). More than cool reason: A field guide to poetic metaphor. Chicago: University of Chicago Press. Langacker, R. (1985). Foundations of cognitive grammar. Indiana: Indiana University. Langacker, R. (1987). Foundations of cognitive grammar. Standford: Stanford University Press. Mandler, J. M. (1992). «Toward a theory of consciousness». In H. G. Geissler et al. (Eds.), Cognition, information processin and psychophysics: basic issues (pp. 43–66). Hillsdale: Lawrence Erlbaum Associates. Marina, J. A. (1999). Diccionario de los sentimientos. Barcelona: Anagrama. Massip, À. (2012). «Language as a complex adaptative system: Towards an integrative linguistics». Complexity perspectives on language, communication and society (pp. 35–60). Massip, Àngels and Bastardas, Albert. Berlín: Springer. Peña, M. S. (2003). «The image-schemestic basis of the event structure metaphor». Annual Review of Cognitive Linguistics, 2(1), 127–158. Talmy, L. (2000). Towards a cognitive semantics. Cambridge: The Mitt Press. Turner. (1996). The literary mind. New York: Oxford University Press.

Correction to: The World Color Survey: Data Analysis and Simulations Peter Lewinski, Michal Lukasik, Konrad Kurdej, Filip Leonarski, Natalia Bielczyk, Franciszek Rakowski and Dariusz Plewczynski

Correction to: Chapter 16 in: À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_16 In the original version of the book, the belated correction from author to remove the co-author name “Joanna Raczaszek-Leonradii” from Chapter 16 should be incorporated. The correction chapter and the book have been updated with the change.

The updated version of this chapter can be found at https://doi.org/10.1007/978-3-030-04598-2_16 © Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2_18

C1

Index

A Accusative, 100, 101 Action, (human) action, 3, 18, 19, 47, 49–51, 58, 60, 77–83, 85–89, 101, 171, 173–176, 183, 189, 191, 193, 194, 197–202, 221, 236, 242, 243, 246, 265, 268, 291, 321 Action-perception, 17–21, 24, 27 Activity theory, 207 Adaptation/adaptive, 2, 7, 79, 81, 82, 127, 167, 168 Adaptive systems, 18, 26 Agent, 19, 33, 38–40, 56, 79, 80, 84–88, 101, 107, 109–111, 116, 122, 123, 214, 290, 292, 302 Agent-based models, 56, 122, 123, 290, 292, 302 Agent demotion, 101 Algorithm, 109, 152, 159, 162, 166–168, 255, 289, 295, 297 Anthropological linguistics, 321 Argumentation theory, 187, 192 Arguments types of arguments cause-effect argument, 201 pragmatic argumentative conclusion (or pragmatic argument), 201 Artificial language, 47, 48 Artificial societies, 38, 39, 41 Association, 60, 85, 254, 257, 258, 260–262, 267, 268 Associative network, 261 Attractor, 24, 26–28, 80, 87, 94, 211

Australian languages, 97 Autopoiesis, 31, 32 B Basavi-Kaluli, 96, 97 Basque, 131, 274–276, 281, 282, 284–286 Belief, 9, 43, 53, 57–62, 65, 70, 102, 167, 171, 174, 175, 180, 216 Bilingualisation, 119, 122, 124, 127, 128, 130, 132 Bilingualism, 147, 172, 180 Biological-cultural object, 94 Brain, 5–8, 10, 19, 50, 77, 84, 85, 120, 180, 189, 197, 198, 251, 252, 260, 261, 263, 264, 291, 317, 318, 320 C Catalan, 78, 81, 89, 90, 127–129, 171, 198, 274, 276 Catalonia, 4, 128–132, 198 Categories, 7, 10, 60, 97, 141, 158, 167, 193, 213, 214, 221, 235, 265–267, 274, 286, 290–293, 299, 301, 316, 320 Causality, 5, 120 Cellular automata, 31, 33, 39, 56, 121, 122 Change, 5, 7–9, 25, 28, 32, 35, 36, 41, 44, 45, 47, 49, 77–83, 85, 87–90, 93–98, 107, 108, 110, 122, 123, 126, 128, 129, 131, 132, 167, 168, 171, 173–175, 180, 188, 192, 197, 200, 210, 219, 233, 234, 236, 241, 277, 301, 322 Change, external, 96 Change, internal, 19

© Springer Nature Switzerland AG 2019 À. Massip-Bonet et al. (eds.), Complexity Applications in Language and Communication Sciences, https://doi.org/10.1007/978-3-030-04598-2

325

326 Clusters of beliefs, 43, 62, 63, 65 Code, 6, 36, 37, 41, 59, 122, 127, 172–174, 176, 177, 180, 182, 183, 196, 240, 296, 320 Cognition, 18 Cognition, (embodied) cognition, 3, 8, 10, 17–21, 33, 51, 59, 77–79, 84, 94, 103, 189, 191, 192, 196, 198, 290, 292, 313, 314, 317, 319–323 Cognitive framework, 58, 192, 197, 201, 268 Cognitive operations, 51, 64 Cognitive system, 17–21, 26, 29, 32, 37, 291, 292 Coherence, 17, 23, 24, 27–29, 219, 316, 321 Collective cognition, 18, 19, 94 Collective intelligence, 87, 88 Common knowledge, 9, 171, 176, 178 Common Routine (CR), 171 Communication, 4, 5, 17, 22, 24–28, 35, 36, 38, 47, 48, 58, 59, 83, 85, 98, 107, 109, 110, 112, 123, 173–181, 183, 190, 191, 193–196, 200–202, 212, 222, 236, 237, 244, 269, 290, 292, 301 Communicative functions, 191 Communicative interactions, 47, 198 Communicative situations, 207 Competence, 126, 172–177, 181, 191 Complex Adaptive System (CAS), 7, 9, 31–33, 77, 79, 86, 87, 108, 119, 133 Complexics, 2, 4–6, 89, 119, 123, 133, 134 Complexity, 1–9, 31–34, 36–41, 43–46, 50, 52, 53, 55–57, 60, 62–64, 70, 71, 77–79, 81, 82, 85, 89, 93, 94, 96, 120–122, 124, 132–134, 146, 152, 162, 168, 187, 188, 192–194, 208–211, 252, 302, 317, 321 Complexity and discourse analysis, 9, 62, 63, 187, 193 Complexity, general, 8, 31–34, 37, 38, 41, 119, 122, 123, 132 Complexity of thinking, 2, 44, 50 Complexity, restricted, 8, 31–34, 41, 122, 133 Complexity theory, 3, 4, 39, 79, 93, 133, 139–141, 168, 187, 201, 207, 211 Complex open systems, 98 Complex problems, 44, 50, 52–57, 70, 71 Complex systems, 3, 4, 8, 9, 33, 38, 45, 50, 52, 55, 56, 63, 64, 66, 70, 78, 82, 86–89, 107, 121, 133, 173, 209, 210, 317 Complex thought, 50, 52, 57, 63, 64, 70, 71, 211 Conflict, 7, 38, 39, 86, 221, 222, 241, 242 Connectionism, 274 Constraint, 80, 86, 87, 283

Index Construction, 4, 32, 44, 51, 53, 54, 58, 60, 63–65, 70, 71, 89, 99, 101, 122, 123, 147, 192, 193, 195, 196, 198, 201, 202, 230, 267, 315, 316 Constructive approach, 54, 55 Constructivism, 46, 89, 167, 187, 192, 195 Contact, 78, 96, 121, 124–126, 128–131, 171–173, 180, 181, 183, 208, 292, 316, 320 Context, 5, 7, 19–22, 27–29, 33, 34, 41, 51, 61, 79, 87, 94, 125, 173–176, 178, 179, 190, 191, 193, 200–202, 222, 229, 233, 237, 245, 283, 321 Context-dependence, 17, 22, 27, 190 Context-free, 22, 27 Context-sensitive, 20, 21, 28 Cooperation, 38, 39, 86, 201, 303 Coordination, 20, 50, 56, 87, 88, 176, 181 Correspondence principle, 28 Cultism, 89 Cultural, 9, 38–40, 47, 49–51, 53, 93, 97, 102, 103, 121, 131, 140, 141, 187, 192, 198, 212, 216, 231, 235, 239, 241, 244, 247, 290, 292, 321 Culture, 38–40, 47, 50, 51, 89, 103, 123, 141, 191, 216, 232, 233, 236, 290, 292, 294, 298 Cybernetics, 1–4, 31, 32, 37, 79, 82, 83, 89 D Database, 141, 150, 159, 167, 255, 257, 265, 289, 293, 295, 296, 303 Dative, 100, 101 Determiner Phrase (DP) hypothesis, 274, 285 Diachrony, 90, 107 Dialect chains, 157 Dialect dynamics, 147–149, 157, 158 Dialogic, 35, 50, 51, 189 Diasystem, 141–143, 154, 162, 167 Dictionary, 142, 237, 251–254, 257, 262, 267, 268 Diffusion, 17, 96, 107, 139, 140, 168 Direct Common Knowledge (DCK), 171, 173, 176, 179–181, 183 Discourse analysis, critical discourse analysis, 194 Discourses, 8, 43, 192–195, 197, 199, 207, 208, 211–214, 222, 230, 231, 247 Discrimination, 292, 299, 317, 319 Discursive process, 51, 190, 198, 200 Dissipative structures, 31 Distributed cognition, 87 Disturbance, 77, 84 Diversion, 84

Index Diversity, 5, 21, 81 Dynamic system, 77, 209 E E. Bernárdez Linguistic macrochange, 9, 93 Ecological factors in change, 102 Ecology, 88, 120, 131, 133, 141, 142, 146, 167, 171, 173, 174, 176 Ecology of Pressures Model (EPM), 173 Ecology, sociocognitive, 120 Ecosystem, 34, 37, 46, 87, 123 Elias, Norbert, 4, 120, 123 Emergence/emergent, 2, 5, 7–10, 20, 22, 26, 27, 36, 38–40, 45, 50, 54, 56, 77, 78, 81, 82, 89, 90, 107, 120, 133, 148, 149, 157, 181, 189, 195, 207, 210, 211, 214, 274, 276, 277, 290, 292 Encyclopedia, 254–256 Entropy, 24, 82, 83 Environment, 4, 17–22, 24, 27, 29, 38–40, 53, 81–84, 86–90, 98, 103, 116, 123, 167, 183, 188, 189, 209, 210, 289–295, 299, 301, 302, 316 Epistemic action, 18, 19 Epistemic framework, 61–63, 65, 69, 70 Epistemological states, 101 Epistemology, 37, 51, 133, 195, 208 Equifinality, 83 Ethnicity, 230 Ethnography, 187, 193, 194 Ethnography of communication, 192, 207, 236 Ethnoracial contexts, 10, 229, 230 Ethnoracial identities, 227, 230 Etymon, 78, 89 Event, 36, 37, 40, 80, 101, 145, 176, 179, 201, 236, 243 Evidentiality, 93, 99 Evolution, 5, 9, 19, 20, 28, 36, 40, 45, 54, 77–82, 89, 90, 95, 98, 107–110, 119, 120, 122, 124, 127–129, 131, 132, 140, 141, 167, 188, 221, 290, 315, 316, 321 Experiences, 53, 176, 178, 232, 245, 316–318, 320, 321 Explanatory patterns, 17, 18, 20–24, 27–29 Extended cognition, 19, 21, 22, 28 External stimuli, facts, 78 F Families, mixed language, 38, 126, 242, 281 Feedback, 3, 23, 35, 36, 82, 83, 176, 178

327 Figurational complexity, 120 Figurational sociology, 120 First-Order Cybernetics, 3, 32 Flexibility, 17, 23, 26–29, 265 Fractal structure, 201 Fragmented vocabulary, 22 Friction, 86 Frontiers, 149, 167 Functionalism, 195 G Genetic epistemology, 46, 50, 60, 61 Genètics, 78 Genotype, 88 Germanic languages, 97, 99 Giant component, 25, 26, 285 Goal, 3, 10, 20, 39, 52, 54, 55, 63, 66, 67, 79, 80, 86, 87, 107, 172, 183, 195, 229, 251, 254, 260, 261, 265–267, 269, 282 Grammaticalization, 99, 101 Greenlandic, 97, 98, 274 H Habitus, 125, 194 Halliday, 190, 191 History, 9, 36, 45, 49, 54, 78, 80, 81, 88, 97, 103, 129, 167, 171, 172, 174, 177, 178, 183, 193, 194, 232, 244, 246 Human beings, 4, 7, 9, 48, 53, 93, 94, 96, 119, 122–124, 133, 173, 193 I Icelandic, 93, 99–103 Ideologies, 9, 129, 171, 174, 175 Immigrant groups, 127 Impersonal passive, 101 Index, 62, 146, 196, 238, 251, 254–256, 274 Indigenous languages, 183 Indirect Common Knowledge (ICK), 171, 173, 176, 178–183 Information, 3–5, 19–21, 24–27, 32, 34, 36, 37, 53, 59, 82–85, 87, 89, 99, 102, 103, 115, 152, 158, 162, 164, 173, 176, 178, 209, 210, 255, 260–264, 267, 269, 276, 292, 303 Intelligence, 85, 86, 88, 107, 245 Intentionality, 37, 211, 214, 221, 222 Interaction, 5–9, 33, 38–40, 56, 60, 84, 86, 87, 89, 94, 107–109, 116, 125, 126, 128, 133, 140, 171, 172, 175, 176, 180, 181,

328 190, 192, 193, 208–210, 212, 216, 220–222, 230, 239, 243, 275, 299, 316, 317 Interdisciplinary, 2, 4, 5, 8, 33, 45, 50, 52, 56, 57, 70, 107, 124, 133, 139, 230, 303 Interest, 28, 47, 48, 109, 124, 127, 132, 174–176, 194, 197, 200, 214, 217, 239 Internal stimuli, facts, 78 Intonation, 10, 207, 208, 211–214, 223, 241 K Knowledge, 1, 5, 6, 9, 21, 22, 27, 32, 44–53, 56, 58–61, 63, 65, 70, 71, 82, 84–88, 99, 102, 108, 121, 123, 125, 126, 128, 141, 146, 171, 172, 176–183, 189, 190, 192, 193, 197, 198, 200, 207, 208, 213, 216, 218–221, 235, 244, 251, 252, 257, 260–263, 269, 295, 313–315, 321–323 L Language, 3, 7–10, 17, 18, 20–22, 25–29, 33, 38, 43, 47, 48, 50, 51, 77–84, 86–90, 93–100, 102, 103, 107–110, 115, 116, 120–132, 139–142, 146, 150, 158, 165, 168, 171, 172, 174–181, 189–193, 195–198, 207–212, 214, 227–237, 239–241, 244–247, 252, 253, 255, 258, 262, 274–277, 282, 285, 286, 290–292, 304–310, 314, 315, 319, 320 Language-based sharing, 20 Language contact, 9, 96, 97, 119, 120, 124, 128, 132, 134, 146 Language production, 7 Language shift, 119, 126, 130, 131, 177 Languaging, 119, 120, 123, 124, 132, 193 Lemma, 251 Lexeme, 201 Lexical concept, 200 Lexical database, 257 Lexicon, 5, 93, 102, 149, 158, 233, 234, 257, 263, 264, 269, 292, 301 Lingüístic, 1–5, 7–9, 24–27, 29, 36, 47, 50, 77, 78, 80–84, 86–88, 90, 94–97, 99, 108, 110, 115, 116, 122, 140, 141, 145, 146, 149–152, 174–177, 181, 184, 190, 193–196, 202, 207, 211–213, 221, 227, 229–231, 233–235, 239, 243–247, 267, 274–277, 281, 286, 291, 299, 301, 313–317, 319, 320, 322, 323 Linguistic change, 77–80, 83, 84, 90, 94, 95, 97, 168 Linguistic coherence, 24, 25, 29

Index M Machine learning, 293, 303 Macrochange, 93–95, 97–99, 103 Macroevolution, 9, 93, 95, 103 Matrix, 39, 151–153, 156, 159, 162, 163, 166, 167, 222 Mazatec, 9, 139, 141–154, 157–159, 165, 167, 168, 307 Meaning, 1, 20, 33, 34, 37, 59, 85, 89, 94, 101, 102, 109, 150, 151, 180, 190, 191, 193, 195, 196, 200, 201, 209, 211, 212, 214, 221, 233, 235, 237, 240, 257, 262, 263, 268, 291, 292, 316–320, 322 Means of expression, 207, 212, 214, 221 Mediator, 88 Memory aid, 19, 21 Mental lexicon, 10, 252, 253, 262, 264, 269, 275 Mental model, 43, 46–48, 59 Mental reality, 180 Metacognition, 43 Metacognitive strategy, 43, 44 Metaknowledge, 269 Meta-point of view, 43, 44, 52 Metastability, 24, 77 Metastable state, 93 Method of thought, 50, 51 Methodology, 6, 33, 56, 57, 119, 133, 187, 193, 194, 198, 274 Methods-quantitative/qualitative, 62, 70, 212 Mexico, 9, 127, 139, 141, 172, 174, 177, 181, 184, 304–309 Minimum spanning tree, 162, 164 Model, 4, 8–10, 18, 22–24, 27–29, 31, 33, 34, 36–41, 43–48, 54–60, 62–71, 82, 83, 85, 93, 107, 108, 110–112, 115, 120–123, 127, 133, 140, 148, 149, 154, 157, 158, 162, 164, 167, 168, 173, 189, 192, 197, 200, 209, 210, 212, 222, 273, 290–292, 302, 316, 320 Model(s)/modeling, 2, 6, 10, 18, 20, 21, 24, 26, 27, 33, 39, 43, 44, 47, 48, 52, 55–57, 62–64, 69–71, 82, 83, 122, 127, 133, 139, 140, 158, 290, 292, 301, 302 Morin, Edgar, 4, 31, 32, 34, 41, 52, 57, 63, 120, 187 Morphogenesis, 36, 37 Morphogenetic processes, 10, 207 Morphology, 88, 93, 102, 165, 274 Multi-agent systems, 31, 41 Multicultural socialization, 232 Multi-dimensional scaling, 165, 166, 168

Index Multimodality, 195, 196 N Narrative (or narrativity), 88, 192, 194, 221, 222, 241, 243 Natural language, 1, 5, 8, 9, 32, 47, 48, 59 Navigation, 253, 260–262, 265, 266, 269 Needs, 5, 7, 19, 26, 48, 54, 55, 61, 94, 96, 122, 123, 174, 179, 181, 194, 195, 200, 234, 262, 263 Negotiation, 114, 174, 222 Neighbor network, 24, 40, 78, 156, 165, 251, 254, 257, 261, 263 Network, 2, 4, 6, 23–26, 35, 61, 77, 79, 81, 82, 84–88, 98, 141–145, 152, 154–158, 162, 164, 168, 171, 175, 176, 179–183, 199, 200, 208, 210, 211, 254, 257, 258, 260–262, 274–277, 281–286 Neural network, 23, 77, 81, 85, 88 Nominative, 100, 101 Non-nominative subject, 101 Nordic languages, 99 Normalized threshold, 161 Nouns, 100, 150, 151, 158, 159, 161, 163–166, 262, 277, 278 O Objectification, 44, 49–53, 57, 70 Observed system, 32, 46, 53, 57, 70 Observing systems, 3, 32, 34, 46, 53, 57, 70 Old icelandic, 99, 100 Old Norse, 101 Ontological beliefs, 58, 61, 65 Organism, 98, 189, 267 Organization, 21–23, 28, 32, 36, 44–46, 52, 54, 55, 58, 60–62, 79, 83, 87, 88, 94, 158, 179, 188, 200, 251–253, 262, 269, 275 Organization of thought, 46 P Paradigm, 1, 2, 4–6, 8, 9, 31–34, 38, 41, 43–45, 49, 50, 52, 57, 60–63, 70, 71, 77, 79, 100, 139, 187, 188, 190, 195, 278, 280 Patrimonial word, 89 Pattern, 17, 18, 20, 24, 27, 60, 61, 63, 154, 162, 188, 190, 211, 213, 216, 219, 235, 245, 265, 286, 301, 316, 317, 321 Performance, 18–20, 80, 88, 113, 114, 195, 229, 230 Phase transition, 24, 87, 94 Phenotype, 88 Phonology, 165, 167, 274, 275 Pluri-contextual, 17, 22, 27, 28

329 Positive transfer, 21 Post-cognitivism, 198 Potential landscape(s), 25 Power, 7, 35, 111, 127, 130, 131, 175, 176, 180, 216, 221, 227, 229, 230, 243 Pragmadialectics, 195 Pragmatic, pragmatics, 9, 36, 96, 98, 102, 158, 190, 193–196, 201, 207, 213, 214, 218, 219, 221 Pressures, 7, 81, 122, 131, 171, 173–180, 183 Pressures in conflict, 173, 175 Problematization, 43, 48–50, 53, 55, 70 Proposition, 59, 61, 62 Prosody, 9, 84, 207, 208, 211, 214, 221, 222 Psycholinguistics, 207, 273, 313 R Race, 227–232, 235, 241, 242, 247 Raciolinguistics, 10, 227, 229–231, 244 Reductionism-emergentism, 26 Relativity, 4, 88, 89 Renormalization procedure(s), 28 Representation, 1, 2, 32, 43, 54, 64, 129, 130, 154, 165, 166, 171, 177, 197, 282, 286, 291, 292, 317–320 Responsibility of the agent, 100 Rhetoric constructivist rhetoric, 195 S Science communication, 21, 29 Science-natural environment system, 18, 19 Scientific beliefs, 8, 43, 57–66, 69, 71 Scientific evolution, 28 Scientific method, 89 Scientific practice, 17, 18, 20–22, 26, 27, 29, 43, 47, 48, 58, 60, 134 Scientific thought, 4, 44, 48 Search, 22, 24, 39, 49, 81, 130, 188, 232, 251, 252, 254, 258, 260–266, 269 Second-order cybernetics, 3, 32, 37 Selection, 6, 34, 37, 60, 79–81, 83, 88, 125, 130, 191, 198, 201, 254, 303, 319 Selective pressure, 28 Self-organized, 32, 36, 173 Self-organizing/self-organization, 5, 7, 8, 17–20, 23, 24, 27, 29, 32, 36, 38, 56, 79, 82, 87, 88, 94, 120, 122, 132, 173, 188, 210, 211 Self-reflexivity, 35–37 Semantic memory, 58, 59 Semiotics, 187, 192, 194, 196, 202, 291 Sense, 2–4, 19, 20, 22, 28, 34, 35, 37, 41, 47, 48, 50–52, 58, 59, 71, 77, 80–82, 94,

330 133, 173, 177–180, 182, 196, 209, 211, 212, 214, 217, 227, 233, 234, 236, 238, 240, 245, 255, 263, 265, 303, 314–317 Social, 3, 4, 6–9, 17–21, 26, 28, 31–39, 41, 43–51, 53–55, 57–61, 63, 67, 79, 80, 83, 84, 88, 96, 98, 107–109, 111, 115, 116, 121–125, 127, 128, 130–133, 139–141, 167, 171–174, 176–180, 183, 190, 192, 194, 196–201, 208, 211, 216, 219, 222, 227, 229–231, 233, 235, 236, 317 Social brain, 84 Social cognition, 8, 20, 21, 43, 57, 84 Social communication, 20, 43, 47 Social intelligence, 50 Social network, 176, 179 Social practices, 59 Social representation, 47, 57 Social simulation, 44, 57, 63, 67, 68, 71 Society, 3, 7, 8, 21, 34, 35, 37, 38, 40, 41, 47, 51, 53, 58, 78, 83, 84, 87, 107–112, 122–124, 127, 132, 144, 168, 178, 179, 193, 196, 199, 219, 246 Sociocomplexity, 9, 120 Socio-cultural organization, 94 Sociogenesis, 45, 59, 60 Sociolingüístics, 4, 7–9, 88, 120, 121, 208, 236, 237 Sociology, 21, 23–25, 31, 34, 41, 82, 83, 120, 121, 192 Socio-verbal, 47, 60 Source of knowledge, 99 Spanish, 90, 127–131, 144, 145, 147, 171, 172, 181–183, 202, 208, 212–215, 223, 228, 233, 236, 275, 276 Spatial relations, 100 Speaker, 97, 109, 110, 129, 175–177, 181, 190, 191, 195, 217, 219–221, 232, 243, 244, 274 Speaking, 7, 27, 35, 77, 85, 94, 125, 128, 129, 132, 144, 146, 175, 180, 227, 231, 235, 236, 241, 242, 244–246, 252 Speech act, 174, 191, 195, 212, 213, 219, 220, 236 Sprachbund, 97 Stability, 17, 24, 26, 29, 36, 87 Stable state, 77 Standard Average European (SAE), 97 State, 5, 9, 22, 24, 25, 44, 45, 49, 80, 81, 85, 87, 94, 96, 101, 102, 119, 125, 127, 129–132, 144, 146, 171, 173–175, 179, 183, 189, 191, 201, 219, 251, 258, 261, 263, 266, 318, 322

Index State of the world, 9, 80, 171, 174, 175, 179, 183 Stigmergy, 88 Structure, 5, 7, 10, 17, 24–26, 28, 35–38, 41, 43–45, 47, 50, 54–57, 59, 62–64, 83, 85, 88, 96, 123, 124, 144, 152, 156, 158, 162, 166, 180, 182, 188–191, 207, 214, 233, 237, 253, 265, 274–276, 278, 279, 281, 284, 286, 290–292, 315–317, 319, 321 Subject, 20, 32, 34–36, 41, 44, 47, 48, 50, 51, 58, 61, 65, 67–70, 82, 89, 100, 101, 123, 182, 183, 187, 188, 191–193, 208, 209, 219, 221, 222, 227, 231, 244, 277, 278, 281, 301, 303, 322 Subsystem, 95, 96, 98, 211, 221, 317 Success, 38, 108, 109, 130, 176, 177, 181–183, 222, 245, 256, 269 Supersystem, 86 Synchrony, 90, 221 Synergi/synergetic, 27, 94 Syntactic theory, 277, 286 System, 3, 6, 17–21, 27–29, 33–39, 41, 43–45, 53–58, 61–64, 70, 71, 77–88, 90, 93–98, 102, 103, 108, 109, 112–115, 124–127, 130–132, 141, 150, 172, 181, 188, 189, 191, 196, 207, 209–214, 219–221, 223, 230, 234, 235, 245, 254, 261, 263–266, 269, 290, 313, 314, 317, 318 System of scientific beliefs, 43 System of thought, 44, 62, 87 Systems theory, 2, 4, 17, 31, 79, 82, 86, 188, 189 T Technological embodiment, 19–21 Temporal relations, 99 Thesaurus, 258, 265 Thought system, 8, 43, 46, 48–50, 52, 57, 61, 70 Time, 4, 7, 20, 22, 23, 27, 28, 32, 37, 40, 45–47, 49, 56, 70, 78, 80–82, 88–90, 93–98, 100, 102, 107–110, 112–115, 119–121, 123, 124, 127, 130–133, 139, 140, 162, 172–176, 178–181, 187–189, 191, 194, 196, 198, 199, 201, 207, 209, 210, 212, 216, 218, 221, 222, 227, 230, 232, 238, 240–243, 246, 252, 255, 257, 260, 263, 266, 276, 290, 296, 303, 317, 318, 322, 323 Tip-of-the-Tongue problem (ToT), 10, 251 Transdisciplinarity, 5, 9, 82, 187, 189, 194, 202, 208

Index U Unifying/general themata, 22, 24–26, 28, 29 Utmost Common Routine (UCR), 173–175 V Valencia, 127, 129, 131 Values, 9, 39, 53, 54, 58, 61, 64–66, 68, 70, 110, 112, 113, 141, 150–152, 159, 171, 174, 175, 212, 213, 222, 290, 298–301, 303, 322 Variation/variational, 7, 9, 69, 78–81, 90, 144, 230, 283, 313, 320, 321

331 Variety, 36, 40, 51, 82–85, 119, 132, 144, 146, 147, 149, 162, 165, 181, 234, 236, 245, 274, 303, 317 Variety of action, 84, 87 W Word, 24, 51, 78, 108–110, 115, 158, 182, 194, 198, 233, 242, 251–258, 260–269, 274, 275, 280, 283 Word access, 251, 254, 258, 260, 261, 268 Word order, 98 Writing, 150, 244, 252