Groups, Norms and Practices: Essays on Inferentialism and Collective Intentionality [1st ed.] 9783030495893, 9783030495909

Citation preview

Studies in the Philosophy of Sociality 13

Ladislav Koreň Hans Bernhard Schmid Preston Stovall Leo Townsend  Editors

Groups, Norms and Practices

Essays on Inferentialism and Collective Intentionality

Studies in the Philosophy of Sociality Volume 13

Series Editor Raul Hakli, Dept of Political & Economic Studies, University of Helsinki, Helsinki, Finland Managing Editors Hans Bernhard Schmid, University of Basel, Basel, Switzerland Jennifer Hudin, University of California, Berkeley, USA Advisory Editors Robert Audi, Department of Philosophy, Notre Dame University, Notre Dame, USA Michael Bratman, Department of Philosophy, Stanford University, Stanford, USA Cristiano Castelfranchi, Cognitive Science, University of Siena, Siena, Italy David Copp, University of California, Davis, Davis, USA Ann Cudd, University of Kentucky, Lexington, USA John Davis, Marquette University, Milwaukee, USA Wolfgang Detel, Department of Philosophy, University of Frankfurt, Frankfurt, Germany Andreas Herzig, Computer Science, University of Toulouse, Toulouse, France Ingvar Johansson, Philosophy, Umeå University, Umeå, Sweden Byron Kaldis, Department of Philosophy, University of Athens, Athens, Greece Martin Kusch, Department of Philosophy, University of Vienna, Wien, Austria Christopher Kutz, Law, University of California, Berkeley, USA Eerik Lagerspetz, Department of Philosophy, University of Turku, Turku, Finland Pierre Livet, Department of Philosophy, Universite de Provence, Marseille, France Tony Lawson, Faculty of Economics, University of Cambridge, Cambridge, UK Kirk Ludwig, Department of Philosophy, Indiana University, Bloomington, USA Kay Mathiessen, Information Science and Philosophy, University of Arizona, Tucson, USA Larry May, Philosophy Department, Vanderbilt University, Nashville, USA Georg Meggle, Institute of Philosophy, University of Leipzig, Leipzig, Germany Anthonie Meijers, Department of Philosophy, University of Eindhoven, Eindhoven, The Netherlands Seumas Miller, Philosophy, Australian National University and Charles Sturt University, Canberra, Australia Uskali Mäki, Academy of Finland, Helsinki, Finland Elisabeth Pacherie, Cognitive Science, Jean Nicod Institute, Paris, France Henry Richardson, Department of Philosophy, Georgetown University, Washington D.C., USA Michael Quante, Department of Philosophy, University of Münster, Münster, Germany John Searle, Berkeley, USA Michael Tomasello, Department of Developmental Psychology, Max Planck Institute, Leipzig, Germany

This book series publishes research devoted to the basic structures of the social world. The phenomena it focuses on range from small-scale everyday interactions to encompassing social institutions, from unintended consequences to institutional design. The unifying element is its focus on the basic constitution of these phenomena, and its aim is to provide philosophical understanding on the foundations of sociality. Studies in the Philosophy of Sociality covers the part of philosophy of the social sciences which deals with questions of social ontology, collective intentionality (e.g. collective knowledge, joint and collective action, joint mental states) and related philosophical topics. The series includes monographs and edited collections on philosophical and conceptual questions concerning social existence as well as conceptual and theoretical analyses of social notions and collective epistemology. In principle, all phenomena dealing with sociality are covered as long as they are approached from a philosophical point of view, broadly understood. Accordingly, the works to be published in the series are generally philosophical—without regard to philosophical schools and viewpoints—and meet the highest academic and intellectual standards. However, the series is interdisciplinary not only in an intraphilosophical sense but also in the sense of encouraging high-level work from other disciplines to be submitted to the series. Others who are active in the field are political scientists, economists, sociologists, psychologists, linguists, neuroscientists, evolutionary biologists, and researchers of artificial intelligence. The resulting interdisciplinary approach suggests new perspectives on the study of topics such as social interaction, communication, unintended consequences of action, social structures and institutions, the evolution of collective intentionality phenomena, as well as the general ontological architecture of the social world. More information about this series at http://www.springer.com/series/10961

Ladislav Koreň  •  Hans Bernhard Schmid Preston Stovall • Leo Townsend Editors

Groups, Norms and Practices Essays on Inferentialism and Collective Intentionality

Editors Ladislav Koreň Filozofická Fakulta Univerzita Hradec Králové Hradec Králové, Czech Republic

Hans Bernhard Schmid Institut für Philosophie Universität Wien Wien, Wien, Austria

Preston Stovall Filozofická Fakulta Univerzita Hradec Králové Hradec Králové, Czech Republic

Leo Townsend Institut für Philosophie Universität Wien Wien, Austria

ISSN 2542-9094     ISSN 2542-9108 (electronic) Studies in the Philosophy of Sociality ISBN 978-3-030-49589-3    ISBN 978-3-030-49590-9 (eBook) https://doi.org/10.1007/978-3-030-49590-9 © Springer Nature Switzerland AG 2021 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, expressed 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

Acknowledgments

Work on this manuscript was supported by the Austrian Research Council, FWF project I 3068, and the Czech Science Foundation (GAČR), Inferentialism and Collective Intentionality, GF17-33808L.

v

Contents

1 Introduction: Between Inferentialism and Collective Intentionality ������������������������������������������������������������������    1 Ladislav Koreň Part I Relating Inferentialism to Collective Intentionality 2 Rational Golems: Collective Agents as Players in the Reasoning Game����������������������������������������������������������������������������   21 Javier González de Prado and Jesús Zamora-Bonilla 3 Trust and Commitment in Collective Testimony����������������������������������   39 Leo Townsend 4 Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation��������������������������������������������������������������������   59 Ronald Loeffler Part II Philosophical and Empirical Cross-Pollination 5 Normative Mindshaping and the Normative Niche������������������������������   85 Jaroslav Peregrin 6 Collective Intentionality, Inferentialism and the Capacity for Claim-Making������������������������������������������������������   99 Glenda Satne 7 Wherein is Reasoning Social? ����������������������������������������������������������������  119 Ladislav Koreň 8 Making Sense of We-Awareness: Experiences, Affordances, and Practices��������������������������������������������������������������������������������������������  147 Anna Moltchanova

vii

viii

Contents

Part III Themes from Wilfrid Sellars 9 Belief Attribution as Indirect Communication��������������������������������������  173 Christopher Gauker 10 Sellars on Rational Agency as Presupposing Collective Attitudes����������������������������������������������������������������������������������  189 Jeremy Randel Koons 11 Understanding What We Ought and Shall Do: A Hyperstate Semantics for Descriptive, Prescriptive, and Intentional Sentences�����������������������������������������������������������������������  215 Preston Stovall Index������������������������������������������������������������������������������������������������������������������  239

About the Editors and Contributors

Editors Ladislav  Koreň  is an Associate Professor and Researcher in the Philosophical Faculty at the University of Hradec Králové. His current research focuses on social-­ normative aspects of human cognition, decision-making, and communication. He has published his works in the areas of epistemology, philosophy of logic, philosophy of language, and philosophy of psychology. Hans Bernhard Schmid  is a Professor of Political and Social Philosophy at the University of Vienna. His research interests include social ontology, phenomenology, and existential philosophy. Preston Stovall  is a Postdoctoral Researcher in the Philosophical Faculty at the University of Hradec Králové. He works on the philosophy of language, metaphysics, and themes in German idealism and American pragmatism. Leo  Townsend  is a Postdoctoral Researcher in the Faculty of Philosophy at the University of Vienna. He works predominantly on social epistemology and collective intentionality and has published papers on the nature of trust, group agency, and collective belief.

Contributors Jesús Zamora Bonilla  Department of Logic, History and Philosophy of Science, Universidad Nacional de Educación a Distancia, Madrid, Spain Javier González de Prado Salas  Department of Logic, History and Philosophy of Science, Universidad Nacional de Educación a Distancia, Madrid, Spain

ix

x

About the Editors and Contributors

Christopher  Gauker  Fachbereich Philosophie in der Kultur- und Gesellschaftswissenschaftlichen Fakultät, Universität Salzburg, Salzburg, Austria Jeremy Randel Koons  Philosophy Department, Georgetown University in Qatar, Doha, Qatar Ladislav  Koreň  Faculty of Philosophy, University of Hradec Králové, Hradec Králové, Czechia Ronald  Loeffler  Department of Philosophy, Grand Valley State University, Allendale, MI, USA Anna  Moltchanova  Philosophy Northfield, MN, USA

Department,

Carleton

College,

Jaroslav Peregrin  Faculty of Philosophy, University of Hradec Králové, Hradec Králové, Czechia Glenda Satne  Department of Philosophy, University of Wollongong, Wollongong, NSW, Australia Preston  Stovall  Faculty of Philosophy, University of Hradec Králové, Hradec Králové, Czechia Leo Townsend  Department of Philosophy, University of Vienna, Vienna, Austria

Chapter 1

Introduction: Between Inferentialism and Collective Intentionality Ladislav Koreň

Abstract  Inferentialism and theories of collective intentionality are two important strands in the current philosophical research. Each tradition recognizes the crucial role that collective norms and practices play in human lives; and each maintains that social attitudes or activities of sorts underlie them. But they have been barely confronted in the literature up to now. This volume brings together new essays, which tackle the issue at hand from different angles, often drawing on and comparing the core ideas developed in each approach. This introductory chapter provides a basic theoretical background and map of the terrain explored by the essays included in the volume. Keywords  Inferentialism · Collective intentionality · Groups · Norms · Social practices

1.1  Overview Inferentialism and theories of collective intentionality are two important strands in the current philosophical research. One way or another, they each hold that certain characteristically human capacities for experiencing, learning, thinking, reasoning and acting are expressions of the human kind of sociality. On the one hand, inferentialism claims that we are sapient creatures because we make claims and reason from and to claims, which is a matter of our abilities to take part in norm-governed social practices of giving and asking for reasons (cf. Brandom 1994; Kukla and Lance 2009; Peregrin 2014). It maintains that underlying this are social-normative attitudes of treating performances as appropriate or not, adopted by agents in the course of engaging in reciprocal interpretation of one another verbal and non-verbal

L. Koreň (*) Philosophical Faculty, University of Hradec Králové, Hradec Králové, Czechia e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_1

1

2

L. Koreň

behaviour. Collective intentionality theories, on the other hand, aim to provide accounts of what makes human experiences, attitudes, actions or practices shared in the first place. According to them, the key lies in our capacities (including motivations) to take collective intentional attitudes of sorts. Though different theories analyse their nature and structure differently, they all agree that they are distinct from intentional attitudes structuring individual intentional activities. Indeed, they hold in common that we are uniquely collective creatures due in great part to our capacities to share and jointly direct our minds at objects, states of affairs, goals, or values (cf. Gilbert 1989, 1996; Searle 1990, 2010; Tuomela 2002, 2007, 2013; Bratman 1999, 2014; Pettit 2001, 2002, 2003). These capacities also make possible shared goals, norms, values or institutions and participate in social practices regulated by them. Inferentialism and theories of collective intentionality are rich resources of ideas concerning the fundamental aspects of human sociality that may potentially challenge but also complement each other. Yet, regrettably, they have been barely confronted in the literature up to now. The present volume aims to fill this lacuna in current research by bringing together new essays of experts based in both traditions, which tackle the issue at hand from different angles, often drawing on and comparing the core ideas developed in each approach. This chapter provides a basic theoretical background and map of the terrain explored by the essays included in the volume.

1.2  Normative Inferentialism Normative inferentialism is at heart a theory of what makes bits of language meaningful (or concept expressing).1 It is a distinctive version of use theory of meaning that explains meanings in terms of normative-inferential roles that expressions acquire with respect to norms determining their correct use within social practices of giving and asking for reasons (cf. Brandom 1994, 2000, 2008; Kukla and Lance 2009; Peregrin 2014; Wanderer 2008; Loeffler 2018). The idea was adumbrated by Wilfrid Sellars (1953) and systematically elaborated in Robert Brandom’s Making it Explicit (1994). Brandom asks what kind of basic social-normative structure a practice must have in order to institute inferential roles conferring meanings (conceptual contents) on linguistic performances caught up in them. He submits that the basic kinds of moves within such a practice are performances with the distinctive functional role of assertion. Assertion is something that can serve as a premise providing reasons for (against) other assertions and for (against) which reasons can be provided. Its credentials are “always potentially at issue” among sapient creatures, there being a room for challenges (requests for 1  Normative inferentialism is to be distinguished from causal or dispositional versions of the inferential-role semantics that construe meanings of expressions (or contents of concepts) partly or wholly in terms of actual inferential transitions or dispositions to such transitions.

1  Introduction: Between Inferentialism and Collective Intentionality

3

reasons) to be answered by other assertions justifying them (Brandom 2000, p. 193). That is why the practice must have something like the structure of the game of giving and asking for reasons. Brandom also holds that discursive moves within such a practice must be subject to norms determining when they are appropriately made. In particular, a subset of assertional and inferential proprieties institutes a structure of normative inferential relations between performances and expressions serving as their vehicles. It is in virtue of such relations that performances and expressions acquire their semantic profiles. The idea is that a vehicle of the game of giving and asking for reasons  – i.e., language  – must comprise, at a minimum, a system of sentential expressions standing in relations of inferability (“to be a reason for”) and incompatibility (“to be a reason against”), provided these incorporate also material (non formal-logical) inferability and incompatibility, and proprieties governing the perceptual language-entry transitions and the practical language-exit transitions issuing into decisions and actions. Normative-inferential relations determine the inferential potentials of sentences. Meanings of other expressions are then accounted for as their contributions to inferential potentials of sentences in which they occur as components – i.e., their inferential roles (cf. Peregrin 2014). This holistic proposal that meaningfulness of bits of language is primarily a matter of normative inferential relations is advanced as an alternative to the representationalist accounts that maintain that meaningfulness is a matter of referential relations of linguistic to non-linguistic items (Brandom 1994, p. xvi). It is a pragmatist alternative, since the semantic level of inferential relations is reconstructed in terms of what agents are doing when producing and interpreting discursive performances of one another; in particular, what normative statuses they recognize and attribute to one another. Here Brandom devises the deontic scorekeeping model (Brandom 1994, chapter 3) purporting to lay down the normative-fine structure of a minimal linguistic practice in terms of the acknowledging, undertaking and attributing of assertional and inferential commitments to which performers might or might not be entitled. The notions of commitments and entitlements are deployed to model what it takes for one’s performance to be correct or appropriate (roughly: it is correct or appropriate for one to do what one is committed or entitled to do). The notions of their acknowledgment and attribution model practical-normative attitudes of practitioners toward performances of themselves and others: what it takes to treat someone’s performance as correct. In or by adopting such attitudes practitioners mutually keep score on the constellations of their commitments and entitlements and changes therein effected in the course of their social exchange. In view of this, the functional role of assertions can be modelled in terms of what kind of difference to the deontic score they effect. Roughly, for one’s performance to be an assertion is for it to be treated as incurring commitments and entitlements to certain other claims, undertaking responsibility to justify it when properly challenged and entitling others to reassert it on one’s authority. Their meanings can be modeled in terms of their inferential articulation: their position in the space of reasons implicitly established by such scorekeeping activities.

4

L. Koreň

If norm-governed social practices of sorts create meanings (or concepts) and with them also human sapience, what grounds norms and practices in the first place? According to Brandom and fellow normativists, only intersubjective practices institute the kind of normative statuses that constitute meaningful linguistic performances and expressions. This is partly the lesson that they extract from Wittgenstein’s considerations to the effect that rule-following  – at any rate, concept-involving activities with determinate conditions of correctness – is only possible against the backdrop of a larger social practice. Elaborating on this, Brandom proposes that the fundamental kind of interpersonal infrastructure is a mutual attribution of normative statuses by virtue of normative attitudes that practitioners take toward one another performances. This is the so-called I-Thou kind of sociality that he takes to be exemplified in the interpersonal activities of his deontic scorekeepers. Importantly, normative attitudes do not relate to normative statuses  – not primarily anyway  – as representations relate to represented contents (including intentional attitudes toward propositional contents). The explanatory strategy is to account for what it is for something to represent something as of a certain kind in terms of its functional role in norm-governed practices. If normative attitudes institute the norms of such practices, the strategy would be vitiated if it had to individuate them in terms of represented contents. Brandom’s idea is that norms constitute meaningful (conceptual) performances and expressions only if basic norms are implicit in the practice. In their most basic form, normative attitudes are primitive attitudes of practically recognizing and treating (responding to) something as appropriate or not – without yet representing it as such. The I-Thou model of sociality also provides the basis for Brandom’s view that objectivity emerges as a structural feature of social-perspectival activities of scorekeeping. Very roughly, the idea is that “objectivity” originates with scorekeepers becoming sensitive to the difference between commitments acknowledged by performers and commitments undertaken by them. They attribute to performers the latter based on what they take to be the inferential consequences of the former in light of their own commitments. Scorekeepers can thus attribute normative statuses that outrun those avowed by performers, thereby developing a basic practical sense that what is appropriate to do or say (what one is really committed or entitled to) does not coincide with what (any) one takes to be appropriate to say or do. So, from within their perspective, there is always room for the distinction between what is correct and what is taken as correct. Eventually, they recognize that their perspective is one among many with this feature (e.g., by recognizing that others assess them in this way too, or registering and resolving conflicts between their present and past commitments). Normative inferentialism submits that the social-normative structure of human discursivity is the key to understanding our sapience: uniquely human capacity to treat our experiences, attitudes and activities as placed in the logical space of reasons – i.e., as essentially something assessable in light of the norms of correct judgment and reasoning. Indeed, it construes all propositional structures, including intentional attitudes involved in deliberation and agency (beliefs, desires, intentions), as presupposing normative practices of the type described above. Thus,

1  Introduction: Between Inferentialism and Collective Intentionality

5

actions are said to be performances “caught up in our practices of giving and asking reasons as moves for which reasons can be proffered and sought” (Brandom 2001, p. 11) and “the propositional contents of the intentional states appealed to in practical reasoning presuppose assertional-inferential proprieties, and hence linguistic social practices” (Brandom 1994, p. 231). So inferentialism endorses the idea that rational agents act on and respond to reasons, but not the order of explanation of those traditional accounts of mindedness and rational agency that understand normative statuses as a superstructure presupposing a preexisting infrastructure of full-blown, propositional attitudes. Rather, to be a minded, rational agent is to be a locus of responsibility and accountability “in the space of reasons”, where reasons are understood in terms of normative statuses conferrable on agents through practices of deontic scorekeeping. This has ramifications potentially bearing on the debates about the nature and ontological status of agents and intentionality, including collectively intentional attitudes. If agency and intentionality is ultimately a matter of being treated as a genuine subject of normative statuses within the social game of giving and asking for reasons, they need not be restricted to individual agents and denied to group or collective ones (cf. Zamora Bonilla 2011; González de Prado Salas and ZamoraBonilla 2015). There is an analogy with Dennett’s intentional stance theory (Dennett 1987), according to which agency and intentionality can be attributed to a system whose functioning is complex enough to warrant explanation or prediction in terms of intentional (desire- and belief-like) states or attitudes. By analogy, it could be suggested that collective agency and intentionality can be attributed, from within the right interpretive-normative stance, wherever groups function in ways complex enough – and in some respects similar enough to individual agents – so that it is warranted to treat them as bearers of commitments and entitlements, and hence of social responsibility and accountability.

1.3  Collective Intentionality Analysis Collective intentionality is the term originally coined by John Searle (Searle 1990). One way or another, students of collective intentionality aim to botanize and illuminate group phenomena such as feeling, intending, accepting, reasoning or committing to something together; as well as norms, practices and institutions involving such attitudes or processes (cf. Tuomela 2002). An interest in it traces back to pioneering analyses of those forms of group action in which agents do something intentionally together. Such actions – variously called joint, group, shared, plural or collective – intuitively differ from individual intentional actions in that agents participating in them intend to contribute to some common goal (often the shared activity itself) and coordinate their contributions accordingly (cf. Schweikard and Schmid 2013). Particularly influential accounts of this phenomenon have been developed by the “big four”: Margaret Gilbert (1989, 1996, 2000, 2009; John Searle 1990, 1995,

6

L. Koreň

2010; Michael Bratman 1999, 2014 and Raimo Tuomela 2002, 2007, 2013). They all agree that participant in shared action act on shared intentions, not straightforwardly analysable into mere aggregations of their individual intentions. That said, they part ways when it comes to spelling out the nature of such intentions, hence what makes actions genuinely shared or collective. The controversy revolves around the issue of what exactly about intentionality it is that can be collective, or shared. Schweikard and Schmid (2013) offer a useful classification of the main positions in the dialectical space. Intentional attitudes in general and collective attitudes in particular have a tripartite structure. First, there is a subject of individual or collective attitudes or actions. Second, intentionality has a mode – namely the kind of the attitude (e.g., belief or intention). Third, intentionality has a content – that which makes the attitude to be about something. Based on this, three groups of representative accounts of collective intentionality can be distinguished. Subject-accounts argue that in order to intend and do something collectively – as a group – individuals have to form a single collective body of a kind (which might be more or less temporary). Collective intentionality is then ascribed to them collectively – as a body. The most prominent account of collective intentionality along these lines has been developed by Gilbert (cf. Gilbert 1990, 1996, 2009; another influential account is developed in Pettit 2001, 2002, 2003). It is based on the notion of a “joint commitment to act as a single body”, which can only be formed (and rescinded) jointly by the parties involved. More generally, then, joint commitment is a commitment of two or more parties to X (intend, believe, accept, act) as one. For this commitment to materialize, each of the parties must somehow express her personal readiness to be jointly committed with the other(s) in the relevant way (where this can be done in various ways). There must be a common knowledge between the participants that these expressions had taken place. Most importantly, there is a structure of mutual obligations and entitlements (though they need not be specifically moral obligations and entitlements). Two (more) parties jointly committed to X are each obligated to act in certain ways and each of them is entitled to demand the other’s conforming action – owing each other such conduct (cf. Gilbert 2009). The notion of “plural subject” of collective intentionality is then defined as follows: “A and B […] constitute a plural subject […] if and only if they are jointly committed to doing something as a body – in a broad sense of ‘do’.” (Gilbert 2006, p. 145). The other competitors in the debate usually try to avoid the assumption of such a plural subject. Content-accounts argue that the collective element is wrapped-up in the specific content of intentions. The most prominent proponent is Michael Bratman (e.g., Bratman 1999, 2014). Bratman’s main undertaking is to account for the “minimal sociality” characteristic of shared-cooperative actions that are routinely performed by normal humans. As in individual intentional actions, agents act on intentions representing their goals (outcomes to be realized) and coordinating their activity. In shared-cooperative actions, however, agents enact a shared intention to perform the action – call it ‘J’ – together. As Bratman further elaborates it, I and you share an intention to J together if (i) each of us intends that we J, (ii) each of us does so in accordance with and because of meshing subplans of each of our intentions

1  Introduction: Between Inferentialism and Collective Intentionality

7

mentioned in (i), and (iii) all this is common knowledge between us (cf. Bratman 1999, p. 143). So collectivity is built in the content of each of our individual intentions of the form “I intend that we J”. And the social glue is a matter of us acting on a shared intention to J, which is a matter of our individual intentions and other attitudes being properly interrelated. This being so, there is apparently no need to introduce new kinds of agents and/or attitudes such as irreducible group agents (minds) and/or primitive we-attitudes. We just need invoke some basic capacities of individual intentionality (which, Bratman submits, we need anyway to account for the more basic case of intra-personal and cross-temporal intentional planning and agency), plus the capacities to (meta-)represent and align our individual intentions, including meshing their subplans, within interpersonal rational planning, coordination and bargaining. In this sense, Bratman purports to offer a parsimonious reduction of collective intentionality to individual intentionality. Mode-accounts argue that collective intentionality is a special, irreducible form or kind of intentionality. Prominent authors in this group include Searle (1990, 2010) and Tuomela (2003, 2006, 2007). Searle claims that interdependent intentional I-attitudes are not enough to account for genuine joint actions that he takes to be not only mutually responsive but also cooperatively loaded. In J-ing jointly, he reasons, my and your individual intentions concern our respective contributions qua contributions to J intended by each of us as something “to be done by us”. Intuitively, to intend to J in this specific we-frame of mind is to intend to J together – as a group. What we intend to do as a group thus differs from what each of us intends as an individual member of the group, though the contents of our we-intentions and I-intentions are connected. For in we-intending to J (as our group goal) each of us also I-intends his/her particular contribution as a means of achieving that goal. In this sense, our individual intentions (their corresponding goals) derive from our over-arching we-intentions (their corresponding goals), rather than the other way round. According to Searle, we-intending is a sui generis attitude based on a biologically primitive capacity: a basic social-cognitive layer with a function and evolutionary rationale of its own  – a cooperation sustaining capacity for forming common goals and coordinating cooperative contributions. For all that, collective intentionality does not belong to some irreducible group subject (mind). All intentions, we-intentions included, are the furniture of the only minds that there are – namely individual minds. Tuomela offers a different version of the mode-account. He says that for participants to “jointly intend” to J, each must “we-intend” to J (for his account of the conditions of we-intending see Tuomela 2005, pp. 341–345). He distinguishes two different senses of this: a strong, “we-mode” sense and a weak, “I-mode” sense. Roughly, in the “we-mode sense”, the participants intend as group members because of a group reason. This contrasts with “I-mode” joint intention: when the participants intend in the I-mode they intend as private persons (cf. Tuomela 2006, p. 35). More generally, Tuomela contends that we can reason, intend and act in the `I-Mode` qua private persons or in the ‘we-mode’ qua members of the group. Putting these substantive differences in approach to one side, there is a basic agreement that collective intentionality of sorts is crucial to the understanding of

8

L. Koreň

human social cognition, and that any understanding of normative attitudes towards other agents, social norms and conventions and normative social practices presuppose collective intentionality. From this perspective, it can be argued that it is necessary to provide an understanding of the social dimension of communication, linguistic meaning and norms that avoids the problematic assumption that in order for an agent to have intentional attitudes, that agent needs to be participating in a normative social practice of mutual interpretation of verbal and non-verbal behavior. This assumption seems problematic for several reasons; but let us highlight three of them. First, the claim that intentional attitudes are or involve social-normative statuses might be considered problematic on the ground that the nature or emergence of such facts themselves is something of a mystery. In this vein, it can be charged that the inferentialist accounts of the core notions of deontic statuses socially instituted by mutual deontic scorekeeping leave something to desire. By contrast, theories of collective have always aspired to account for social ontology (in particular, Gilbert 1989, 2000; Tuomela 2002, 2013; Searle 1995, 2010; Pettit 2002, 2003) and come up with well-developed accounts of deontic status as constituted by collectively intentional attitudes such as collective acceptance or recognition. In particular, influential normativist accounts of collective intentionality such as Gilbert’s have analysed collective intentions (as well as beliefs) in part in term of joint commitments. Indeed, Gilbert (1989, 1996) has rather extensively analysed the type of normative implications, which joint commitments license. To many, it has seemed that such analyses thus promise to offer deeper elucidation of fundamental inferentialist categories in terms of collective intentional attitudes. Second, communicative exchanges of discursive moves appear to be intentional cooperative activities. If so, it may be argued that they are amenable to analyses in terms of collective intentional attitudes. Indeed, there have been attempts to develop a Grice-style account of intentional communicative interaction in this direction. They have emphasized (a) the role of intentional attitudes (in particular, communicative intentions) in communication, and (b) its cooperative- joint character (involving a joint commitment, based on a common ground of sorts, to make oneself transparent to the recipient by disclosing one’s cooperative intention to communicate something of relevance to her). Of particular interest here are hypotheses that combine elements of the collective intentionality analysis (in particular, the views of Bratman and Gilbert) and the Gricean approach to verbal and non-verbal intentional communication (cf. Tomasello 2008, 2014; Clark 1996, 2006; Levinson 2006 and Enfield and Levinson 2006). Third, theorists of collective intentionality might worry that if normative statuses implicit in shared normative practices are constitutive of intentional attitudes, it is difficult to explain both the phylogenesis and ontogenesis of such practices or requisite capacities to skilfully engage in them. Yet it has been argued in important branches of current scientific research (e.g. Tomasello et al. 2005; Tomasello 2008, 2014) that the kind of collective intentionality that involves normative attitudes and states builds on more primitive forms of joint or shared attention, goal-directed action and intentionality, and that it is in virtue of their cooperative-mindedness – or

1  Introduction: Between Inferentialism and Collective Intentionality

9

shared intentionality – that human babies come to engage in the kind of cooperative practices that ultimately result in human linguistic communication. This research program has amassed impressive evidence that human babies come to engage in the kind of cooperative practices that eventually develop into human linguistic communication (and discursive practices). According to it, joint actions require a prosocial motivation to share experience and the social-cognitive skills of shared intentionality including, in particular, joint attention and social-recursive mindreading (e.g., Tomasello et al. 2005; Tomasello 2014). This naturalistic approach is a welcome complement to “standard“ approaches of the “big four” (mentioned under the banners of “subject”, “mode”, and “content” accounts) that are quite demanding as regards their cognitive requirements (e.g., higher-order attitudes, common knowledge). In their case, it is far from clear that they can really accommodate the evidence of (at least nascent) collective intentionality in human infants or toddlers, and hence provide a developmentally plausible account thereof. At the same time, Tomasello often appeals himself to the ideas of “big four” (e.g., providing a simplified version of Bratman’s account of joint action), which makes him prima facie vulnerable to the charge that this does not sit well with his ambition to provide a plausible developmental (and phylogenetic) account of the motivations and skills of shared (joint and collective) intentionality (cf. Tollefsen 2005; Butterfill 2012; Pacherie 2013). Increasingly influential are therefore various minimalist analyses trying to account for joint action/intention without appeal to such rich, propositional structures and saving on both normative and cognitive assumptions (cf. Tollefsen 2005; Vesper et  al. 2010; Knoblich et  al. 2011; Tollefsen et al. 2011; Pacherie 2013; Butterfill 2012, 2017).

1.4  Topics Addressed in the Volume It transpires that even though inferentialism and theories of collective intentionality share in common interest in norms and shared social practices structured by them, they approach them in quite divergent ways. Take their accounts of the fundamental structure of human sociality differ markedly. Brandom conceives of it in terms of I-Thou relations structuring activities of mutual interpretation. Whereas theories of collective intentionality posit much more interdependent forms of sociality based on collective intentional attitudes. On top of that, Brandom depicts I-Thou relations as standing in a kind of tension with what he calls I-we relations: ‘[the] tendency to talk of the community as somehow having attitudes and producing performances of the sort more properly associated with individuals is neither harmless nor accidental. This facon de parler is of the essence of the communal assessment approach. It is a manifestation of the orienting mistake […] of treating I-we relations rather than I-thou relations as the fundamental social structure’ (Brandom 1994, p. 38–39).

Relatedly, each tradition recognizes the crucial role that norms and practices play in human lives; and each maintains that social attitudes or activities of sorts underlie

10

L. Koreň

them. But they theorize the nature of such attitudes or activities differently. Focusing attention on the normative structure of discursive practices, inferentialists claim that their underlying social attitudes are ultimately normative attitudes of practically treating performances as appropriate or not, adopted by human agents in the course of engaging in reciprocal interpretation of one another verbal and non-verbal behaviour. Collective intentionality theories, by contrast, provide analyses of what makes experiences, attitudes, actions or practices shared in the first place. According to them, the key lies in the capacities to take collective intentional attitudes of sorts. This raises a host of intriguing issues that deserve closer exploration. Surprisingly, though, these issues are rarely so much as raised, still less thoroughly explored in the literature. With few exception (e.g., Zamora Bonilla 2011 and De Prado Salas and Zamora-Bonilla 2015), the two research programs have not been confronted with each other in the received literature, and no attempt has been made so far to explore and critically assess their relationship (as regards premises, desiderata, possible complementarities, etc.). This volume fills this lacuna in current research. It consists of ten essays divided into three parts: three essays in Part I show that some of the theoretical resources developed by inferentialists may be put to good use by theorists of collective intentionality, and vice versa; four essays in Part II examine dimensions along which philosophical and empirical research on human reasoning and collective intentionality can benefit from more cross-pollination; finally, three essays in Part III examine themes from inferentialism and collective intentionality that arise in the work of Wilfrid Sellars.

1.4.1  P  art I “Relating Inferentialism to Collective Intentionality” It can seem, at first blush, that collective intentionality and inferentialism are too different in their focus, methods and theoretical commitments to be brought into fruitful dialogue. Yet, the papers in this part of the volume demonstrate that inferentialism and collective intentionality are neither wholly orthogonal, and hence irrelevant, to one another, and nor do they stand in any kind of irreconcilable tension. On the contrary, the guiding thought in these papers is that bringing together inferentialism and collective intentionality may prove theoretically fruitful; that some of the theoretical resources developed by inferentialists may be put to good and proper use by theorists of collective intentionality, and vice versa. The first two papers show this by invoking inferentialist concepts to address perennial problems for theorists of collective intentionality, while the third invokes a collective intentional concept in order to address a particular problem for one leading version of inferentialism. In their contribution, González de Prado Salas and Zamora Bonilla show how a broadly inferentialist approach to collective belief can help to overcome a serious challenge faced by traditional accounts of collective belief. The challenge, put bluntly, is that what we call collective belief is not really belief properly so-called.

1  Introduction: Between Inferentialism and Collective Intentionality

11

This is because, although we readily ascribe beliefs to groups of various shapes and sizes, these ‘collective beliefs’ seem to lack some of the essential features of genuine beliefs: they can be formed ‘at will’ and on the basis of non-epistemic, practical reasons. This has led some philosophers to suggest that collective belief is not belief at all, but is instead some form of ‘acceptance’, where this is understood as an attitude that can be formed at will, and for non-epistemic reasons. González de Prado Salas and Zamora Bonilla respond to this challenge by adopting an inferentialist approach to belief, which associates belief with a particular normative or deontic status. With this conception of belief in hand, they argue that the belief-like attitudes of collectives are actually subject to the norms that govern belief, even if these norms can be broken more blatantly in the case of groups than in the case of individuals. Whereas González de Prado Salas and Zamora Bonilla focus on the internal epistemic lives of groups – on the ways they form and are normatively responsible for their beliefs – Townsend zooms in on the external aspect, i.e., the epistemic role groups play in the wider discursive community. Like individuals, groups are able to actively participate in the dissemination of knowledge by means of testimony, and Townsend’s particular focus is on one leading account of collective testimony, due to Miranda Fricker (2012). According to Fricker, in testifying, a group makes a commitment of trustworthiness to its audience, thus undertaking a certain sort of responsibility and making itself liable to blame or sanction should it fail to fulfil this commitment. Townsend argues that the way Fricker construes this commitment means it fails to underwrite a genuinely epistemic form of responsibility. He suggests that a better way to model the commitment involved in testimony would be along broadly Brandomian lines, as an undertaking of specific epistemic ‘task-­ responsibility’, to vindicate one’s entitlement to the claim being made, should it be challenged. In contrast to both Townsend’s paper, and that of González de Prado Salas and Zamora Bonilla, Loeffler’s paper employs a concept developed within collective intentionality – the notion of a ‘we-attitude’ – in order to answer a challenge faced by Brandom’s inferentialism. Loeffler’s particular focus is on Brandom’s notion of implicit scorekeeping, which is central to his overall explanatory project. Scorekeeping, according to Loeffler, is a form of mindreading, but contemporary mainstream approaches to mindreading tend to see it as always involving explicit higher-order thought about the other’s mental activity. These mainstream approaches to mindreading thus seem to undermine this key Brandomian notion of implicit scorekeeping, and so threaten Brandom’s broader project. Loeffler responds to this perceived threat by arguing that the idea of implicit scorekeeping can be vindicated if we conceive of belief as a ‘propositional we-attitude’. What this means is that the very capacity to have beliefs oneself already involves the recognition and evaluation of others qua believers, where this interpersonal doxastic recognition and evaluation is not mediated by explicit higher order thoughts about others’ mental activity.

12

L. Koreň

1.4.2  Part II “Philosophical and Empirical Cross-Pollination” Inferentialists and collective intentionalists theorize social practices in terms of their normative and intentional structures respectively. To this end, they may draw on classical philosophical methods, including (re-)interpreting historical ideas considered inspirational, engaging in conceptual analyses, constructing illuminating models, providing phenomenological analyses of the phenomena, and the like. At the same time, coordination, cooperation, collectivity, communication or norms are all social phenomena of interest to social, cognitive and behavioural sciences. Indeed, much cutting-edge, empirically-oriented research has recently focused on their evolutionary, cultural, ontogenetic and psychological aspects. The question therefore suggests itself whether there are any salient points of contact and prospects for cross-fertilization. The four papers in this part – by Peregrin, Satne, Koreň and Moltchanova – show in concreto that and how philosophical approaches could profit from paying close attention, including constructively critical, to empirical research, and vice versa. This, in fact, has been a progressive trend in the burgeoning literature on collective intentionality, where prominent psychologists like Michael Tomasello already are among the key players. (Incidentally, all four papers take account of Tomasello’s research). In the inferentialist literature, it is still a rare sport to engage seriously with what science has to say about man as a natural and social creature. But the contributions in this part of the book show that the sport is worth trying out. Peregrin addresses the question of why and how humans evolved into unique hyper-social, that is, cultural creatures. He elaborates the idea that this was due in great part to having evolved abilities to shape minds of one another via social processes of imitation, learning, teaching, norm enforcement and, of course, language. However, he proposes that the core mechanism of mindshaping consists in establishing, learning and teaching norms or rules (Sellars’s “ought-to-bes” implying “ought-do-dos”). This idea, he submits, is consistent with and further enriches recent naturalistic models of cumulative cultural niche construction. Plus, it can be read as offering an alternative to the mainstream view that it was the evolution of mindreading that grounded also the evolution of human ultrasociality, including culture and language. And rather than being grounded in collective intentionality, as Tomasello claims, Peregrin concludes that also collective intentionality might be the product of the processes of establishing and socially maintaining and reproducing rule-governed social activities or practices. According to Satne, inferentialists and collective intentionalists both share interest in shared practices and both are partly right and partly wrong about them. Inferentialists are right that contentful intentional attitudes cannot be presupposed as unexplained explainers, especially if we look at the matter from an evolutionary or developmental perspective. They also score a point when maintaining that determinate contents require conditions of correctness that can emerge only within

1  Introduction: Between Inferentialism and Collective Intentionality

13

shared normative practices. But Brandom’s own reconstruction of the social game of giving and asking for reasons in terms of I-Thou relations underlying deontic score-keeping is found inadequate, because it does not show how such shared norms and practices can be constituted in the first place. Here collective intentionalists score a point when they point out that shared practices do require some kind of we-­ intentionality. However, shared activities and we-intentionality develop from the basic to increasingly complex. Contentful attitudes are co-constituted with normative practices at later stages, building on shared activities that are regulated by more basic kinds of collective intentionality that is not contentful. This, Satne submits, is consistent with both the evolutionary and developmental perspective. Koreň addresses yet another topical issue of common interest to inferentialists, collective intentionalists and empirical scientists alike – namely, the social nature of human reasoning. Inferentialism offers a philosophical elaboration of the idea that reasoning is primarily a competence to play the social game of giving and asking for reasons. A rather similar idea has been elaborated from the naturalistic perspective by Hugo Mercier and Dan Sperber, who argue that reasoning evolved primarily for social justification and argumentation. And Michael Tomasello has developed an alternative version based on his own influential account of shared (collective) intentionality: i.e., reasoning evolved primarily for truly cooperative argumentation and joint or collective decision-making. Koreň argues that although Mercier-Sperber’s naturalistic hypothesis is promising, their account of how reasoning works faces some philosophical objections that inferentialism is in a better position to avoid. On the other hand, “naturalized” inferentialists could find congenial much of what Mercier and Sperber have to say about the social origins and functions of reasoning. Finally, he argues that Tomasello’s account may go too far in the collectivist tradition. He concludes that a more plausible naturalistic account of the sociality of human reasoning would likely recognize a mix of different social contexts and functions of reasoning. In her essay, Moltchanova aims to justify the proposal that there is a genuine form of “plural self-awareness” that differs from extant conceptions framed in terms of content, mode or subject. On her account, we-awareness is not reducible to two (or more) “I”s both (all) having a thought to the effect that “we J together”. Nor do those “I”s entertain their thoughts in a special we-mode, as opposed to an I-mode. Nor is it a matter of there being a plural subject with an undifferentiated experience. Rather, there is a multitude of “we-parts” that are the loci of the awareness of one “we”. We-awareness of we-parts can be temporary (up to short-lived, e.g., lasting only as long as the shared activity). Also, it does not exclude that the loci of we-­ awareness can also be he loci of I-awareness. Along the way, Moltchanova brings to bear empirical research to support this kind of view as well-motivated. For instance, the mere fact that other people might be co-present and attend to the same object, or that they might join us in a joint activity, can subliminally transform the way we process stimuli. As if our egocentric frame was replaced with an allocentric frame, indicating, she submits, we-awareness of sorts.

14

L. Koreň

1.4.3  Part III “Themes from Wilfrid Sellars” Inferentialism is the legacy of Wilfrid Sellars` analysis of conceptual-propositional content in terms of normative-functional roles of linguistic (and derivatively mental) items “in the logical space of reasons”. Many have explored in detail this ancestry (Brandom 1994; O’Shea 2007; deVries 2005). Less attention, however, has been devoted to the fact that it was Sellars who introduced at the scene the very concept of “we-­intention”, which was subsequently recognized by the proponents of the analysis of collective intentionality as their source of inspiration (cf. Tuomela and Miller 1988; O’Shea 2007; Schweikard and Schmid 2013). Even though it has been registered in the received literature that, in Sellars’ practical philosophy, we-intention plays the role of making attitude-dependence of value judgments compatible with their claim to intersubjective validity, no systematic and detailed attempt has been made to locate Sellars’ conception of we-intention in the broader context of his inferentialism. One way or another, the papers included in this last part draw inspiration from Wilfrid Sellars’ wide-ranging research on the inferentialist theory of meaning, philosophy of psychology and his conception of practical philosophy as based on collective intentions. In his famous “Myth of Jones”, Sellars (1956) proposed that occurrent thoughts can be understood on the model of overt episodes of speech – i.e., as covert speech-­ like episodes that might occur in absence of or antecedently to overt episodes of speech and other goal-directed behavior. Subsequently, Sellars has been interpreted as a pioneer of the theory-theory of mind – the idea that our common-sense understanding of each other and ourselves as minded agents consists in a theory of unobservable mental episodes and states  – paradigmatically, beliefs and desires  – intervening in a law-like manner between perception and rational-­ intentional behavior. Indeed we draw on the theory when predicting and explaining one another behavior (paradigmatically based on belief-desire laws of sorts). And some have assumed that decision theory provides a precise formal explication of this common-sense theory. In his essay, Gauker (himself one of Sellars’ last students) argues that the view of belief-desire attribution as an explanation or prediction of behavior is misplaced. On his account, though such attributions do sometimes explain behavior, they don’t predict it. In place of this conception, Gauker argues that belief-desire attribution has a primarily communicative function: roughly, to attribute a belief is to make an assertion on someone else’s behalf, and to attribute a desire is to issue a command on someone else’s behalf. Gauker argues that despite having a grounding in social practices of communication, this view can nevertheless be seen as realist about beliefs and desires, provided we take into account the fact that they are norm-governed realities. Koons examines Sellarsian theories of meaning, on which inferential role is foregrounded, and focuses on the idea of the conceptual as the rule-governed. Having explained the crucial role of pattern-governed behavior – rule-related but not yet rule-following  – he argues that in the conceptual case it requires that rules of

1  Introduction: Between Inferentialism and Collective Intentionality

15

material inference be couched in a meta-language. And, specifically, they must be understood as expressions of collective intentions of sorts. After detailing Sellars’ theory of linguistic meaning, and the distinction between what ought-to-be with regard to one’s linguistic dispositions and what one (or another) ought-to-do to see to it that these dispositions are as they ought to be, the author argues that these dispositions within language-learners are shaped by collective intentions shared among language-­trainers. In the course of spelling out the view the author relates Sellars’ theory of meaning to debates between Gilbert and Tuomela in the literature on collective intentionality, arguing that while we need to appeal to Tuomela’s strong wemode attitudes, Gilbert’s view better captures the community-wide implicit collective intentions (commitments) to shape one another’s linguistic dispositions that lie at the back of the explicit collective intentions that Tuomela appeals to. Stovall’s essay provides a semantic framework for understanding descriptive sentences as representations of possible worlds and both prescriptive and intentional sentences as expressions of plans of action. Though he criticizes and revises some of Sellars’ views, this provides a clear formulation of Sellars’ claim that moral judgments are in some sense a kind of collective intention. For Sellars, to say that an action is obliged or forbidden is to express commitment to a collective intention concerning what anyone in the relevant community would or would not do. A view of this sort was championed by Allan Gibbard in the 1990s and early 2000s, and Stovall shows how Gibbard’s planning-based semantics can be extended to account for the contents of sentences that give voice to individual and collective intentions, expressed with the modals ‘I shall’ and ‘we shall’. In doing so he argues that this allows us to understand both moral reasoning and individual and collective intentionality as species of the genus practical rationality, and he shows that it delivers a number of results that any such semantics should be expected to model. Acknowledgments  Work on this chapter was supported by the joint Lead-Agency research grant between the Austrian Science Foundation (FWF) and the Czech Science Foundation (GAČR), Inferentialism and Collective Intentionality, GF17-33808 L.

Bibliography Brandom, R. (1994). Making it explicit: Reasoning, representing, and discursive commitment. Cambridge, MA: Harvard University Press. Brandom, R. (2000). Articulating reasons: An introduction to Inferentialism. Cambridge, MA: Harvard University Press. Brandom, R. (2001). What do expressions of preference express?. In C.W. Morris & A. Ripstein (Eds.), Practical Rationality and Preference: Essays for David Gauthier (pp. 11–36). Cambridge: Cambridge University Press. Brandom, R. (2008). Between saying and doing: Towards an analytic pragmatism. Oxford: Oxford University Press. Bratman, M. (1999). Faces of intention: Selected essays on intention and agency. Cambridge: Cambridge University Press. Bratman, M. (2006). Dynamics of sociality. Midwest Studies in Philosophy, XXX, 1–15.

16

L. Koreň

Bratman, M. (2014). Shared agency. Oxford: Oxford University Press. Butterfill, S. (2012). Joint action and development. The Philosophical Quarterly, 62(246), 23–47. Butterfill, S. (2017). Coordinating joint action. In M. Jankovic & K. Ludwig (Eds.), The Routledge handbook of collective intentionality (pp. 68–82). London/New York: Routledge. Clark, H. (1996). Using language. Cambridge: Cambridge University Press. Clark, H. (2006). Social actions, social commitments. In Roots of human sociality: Culture, cognition and interaction (pp. 126–150). Oxford: Berg Press. Dennett, D. (1987). The intentional stance. Cambridge, MA: MIT Press. deVries, W. (2005). Wilfrid Sellars. New York: McGill-Queen’s University Press. Enfield, N. J., & Levinson, S. C. (Eds.). (2006). Roots of human sociality. New York: Berg. Fricker, M. (2012). Group testimony? The making of a collective good informant. Philosophy and Phenomenological Research, 84(2), 249–276. Gilbert, M. (1989). On social facts. New York: Routledge. Gilbert M (1990). Walking together: a paradigmatic social phenomenon. In Gilbert 1996, pp. 177–194. Gilbert, M. (1996). Living together: Rationality, sociality, and obligation. Lanham/Boulder/New York/London: Rowman and Littlefield. Gilbert, M. (2000). Sociality and responsibility: New essays in plural subject theory. Lanham/ Boulder/New York/London: Rowman and Littlefield. Gilbert, M. (2006). A Theory of Political Obligation. Oxford: Oxford University Press. Gilbert, M. (2009). Shared intention and personal intentions. Philosophical Studies, 144, 167–187. González de Prado Salas, J. G., & Zamora-Bonilla, J. (2015). Collective actors without collective minds an inferentialist approach. Philosophy of the Social Sciences, 45(1), 3–25. Knoblich, G., Butterfill, S., & Sebanz, N. (2011). Psychological research on joint action: Theory and data. In Psychology of learning and motivation (Vol. 54, pp.  59–101). Burlington: Academic. Kukla, R., & Lance, M. (2009). “Yo!” and “Lo!”: The pragmatic topography of the space of reasons. Cambridge, MA: Harvard University Press. Levinson, S. C. (2006). On the human interaction engine. In N. J. Enfield & S. C. Levinson (Eds.), Roots of human sociality (pp. 39–69). Oxford: Berg Press. Loeffler, R. (2018). Brandom. Cambridge: Polity Press. Mercier, H., & Sperber, D. (2017). The enigma of reason. Cambridge, MA: Harvard University Press. O’Shea, J. R. (2007). Wilfrid Sellars. Cambridge/Malden: Polity Press. Pacherie, E. (2013). Intentional joint agency: Shared intention lite. Synthese, 190(10), 1817–1839. Peregrin, J. (2014). Inferentialism: Why rules matter. Basingstoke: Palgrave Macmillan. Pettit, P. (2001). Collective intentions. In N. Naffine, R. J. Owens, & J. Williams (Eds.), Intention in law and philosophy (pp. 241–254). Farnham: Ashgate. Pettit, P. (2002). Collective persons and powers. Legal Theory, 8, 443–470. Pettit, P. (2003). Groups with minds of their own. In F. Schmitt (Ed.), Socializing metaphysics (Vol. 2003, pp. 167–193). New York: Rowman and Littlefield. Schmid, H. B. (2009). Plural action: Essays in philosophy and social science. Dordrecht: Springer. Schweikard, D. P., & Schmid, H. B. (2013). Collective intentionality. In Stanford Encyclopedia of philosophy. http://plato.stanford.edu/entries/collective-intentionality/ Searle, J. (1990). Collective intentions and action. In P. Cohen, J. Morgan, & M. E. Pollack (Eds.), Intentions in communication. Cambridge, MA: Bradford Books/MIT Press. Searle, J. (1995). The construction of social reality. New York: The Free Press. Searle, J. (2010). Making the social world: The structure of human civilization. New York: Oxford University Press. Sellars, W. (1953). Inference and meaning. Mind, 62(247), 313–338. Sellars, W. (1954). Some reflections of language games. Philosophy of Science, 21(3), 204–228. Sellars, W. (1956). Empiricism and the philosophy of mind. In H.  Feigl & M.  Scriven (Eds.), Minnesota studies in the philosophy of science (Vol. I, pp. 253–329). Minneapolis: University of Minnesota Press.

1  Introduction: Between Inferentialism and Collective Intentionality

17

Sellars, W. (1963). Philosophy and the scientific image of man. In R. Colodny (Ed.), Science, perception, and reality (pp. 35–77). Atascadero: Ridgeview Publishing. Tollefsen, D. (2005). Let’s pretend. Children and joint action. Philosophy of the Social Sciences, 35(1), 75–97. Tollefsen, D., Perron, D., & Dale, R. (2011). Naturalizing joint action: A process-based approach. Philosophical Psychology, 25(3), 385–407. Tomasello, M. (2008). Origins of human communication. Cambrdige, MA: Harvard University Press. Tomasello, M. (2014). A natural history of human thinking. Cambrdige, MA: Harvard University Press. Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences, 28(05), 675–691. Tuomela, R. (2002). The philosophy of social practices: A collective acceptance view. Cambridge: Cambridge University Press. Tuomela, R. (2003). The We-mode and the I-mode. In F. Schmitt (Ed.), Socializing metaphysics: The nature of social reality (pp. 65–91). Lanham: Rowman and Littlefield. Tuomela, R. (2005). We-intentions revisited. Philosophical Studies, 125, 327–369. Tuomela, R. (2006). Joint intention, We-mode and I-mode. Midwest Studies in Philosophy, 30(1), 35–58. Tuomela, R. (2007). The philosophy of sociality: The shared point of view. New  York: Oxford University Press. Tuomela, R. (2013). Social ontology: Collective intentionality and group agents. New  York: Oxford University Press. Tuomela, R., & Miller, K. (1988). We-intentions. Philosophical Studies, 53(3), 367–390. Vesper, C., Butterfill, S., Knoblich, G., & Sebanz, N. (2010). A minimal architecture for joint action. Neural Networks, 23(8–9), 998–1003. Wanderer, J. (2008). Robert Brandom. Stockfield: Acumen. Zamora Bonilla, J. (2011). Rationality in the social sciences: Bridging the gap. In I. C. Jarvie & J. Zamora-Bonilla (Eds.), The Sage handbook of the philosophy of social sciences (pp. 721–738). London: SAGE Publications.

Part I

Relating Inferentialism to Collective Intentionality

Chapter 2

Rational Golems: Collective Agents as Players in the Reasoning Game Javier González de Prado and Jesús Zamora-Bonilla

Abstract  We ordinarily attribute beliefs and other intentional states to collective entities. These attributions can be vindicated from a theoretical perspective that holds that: (i) collective entities can behave as rational agents in our argumentative practices of giving and asking for reasons; and (ii) attributions of beliefs are interpretative tools aiming to make sense of the behaviors and perspectives of agents, and to keep track of their commitments. However, it is not immediately clear that groups have attitudes that play the same role in our argumentative and interpretative practices as the beliefs of individuals. It seems that the belief-like attitudes attributed to collective agents lack some of the distinctive features of the beliefs of individuals. More specifically, collective (but not individual) agents seem to be able to form beliefs against the available evidence at will, moved by rewards or other practical considerations. This has made some authors (Wray, Synthese 129(3): 319–333, 2001; Wray PRO 18: 363–376, 2003; Meijers, ProtoSociology 16: 70–85, 2002; Meijers, PRO 18: 377–388, 2003; Hakli, Cogn Syst Res, 7(2–3): 286–297, 2006) argue that the belief-like attitudes of collectives are actually some form of acceptance (which does not involve the commitment to truth characteristic of belief). On our view, this proposal is misguided. We shall argue that the belief-like attitudes of collectives are actually subject to the norms that govern belief, even if these norms can be broken more blatantly in the case of groups than in the case of individuals. In order to explain why this is so, we appeal to the fact that doxastic deliberation in groups is mediated by the deliberative actions of the relevant members, which may be rationally motivated by practical considerations. Keywords  Collective belief · Deliberation · Judgment aggregation · Acceptance · Transparency

J. González de Prado (*) · J. Zamora-Bonilla Department of Logic, History and Philosophy of Science, Universidad Nacional de Educación a Distancia, Madrid, Spain © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_2

21

22

J. González de Prado and J. Zamora-Bonilla

2.1  Collective Believers It is part of our ordinary practice to attribute beliefs and other cognitive states to collective entities, such as firms, sport teams, governments or institutions.1 Speakers often say things like ‘Microsoft believes we are on the threshold of quantum computation’ or ‘The government is certain that the economic situation will improve during the next year.’ On the face of it, these attributions may seem puzzling. After all, groups lack the sort of phenomenology and neural structure characteristic of individual human believers. One could expect such phenomenology and biological constitution to be essential aspects of what it is to be a believer. Can we then vindicate our attributions of collective beliefs, or should we rather regard such attributions as simply metaphorical or loose talk? We think that an independently attractive functionalist-pragmatist account of belief attribution makes it possible to justify our ascriptions of beliefs (and other cognitive states) to collectives. According to this approach, having a belief is a matter of being in a dispositional state that plays a given role in a suitable practice – in particular, in relation to action, reasoning and language use (see Brandom 1994; Sellars 1956). Having a belief that p will amount to being disposed to engage in certain patterns of action, speech and reasoning. For instance, an agent believing that it is raining will be typically disposed to open her umbrella (if she is outdoors and has the goal to remain dry), to assert ‘It is raining’ (if it is conversationally relevant) and to infer that the streets will be wet. When we attribute to an agent the belief that it is raining, we would be ascribing to her these types of dispositions. Note that we have not mentioned the phenomenology of agents or their brain states. As long as the agent has suitable behavioral and linguistic dispositions, it will make sense to attribute beliefs to her, regardless of her underlying psychological constitution. In this way, it may be in principle appropriate to attribute beliefs to collective entities, provided that they manifest the relevant dispositions. On our preferred view, defended at length by Brandom (1994, 2000), the dispositions characterizing what it is to be a believer are norm-governed. More specifically, attributions of beliefs take place within normative practices of giving and asking for reasons. Beliefs are characterized by their role as premise-states of pieces of reasoning. Believing that p involves a commitment to such a proposition and to its inferential implications (so that one cannot correctly believe propositions logically or materially incompatible with p while correctly believing p). In turn, correctly believing that p puts one in a position to correctly believing its inferential implication q, if one transitions to that belief by properly reasoning from one’s correct belief that p. Likewise, when one correctly believes the premises of good pieces of practical reasoning, one is in a position to correctly undertake the course of action recommended by the conclusion of such reasoning. In public discourse, we 1  The notion of collective belief has been defended, among others, by Gilbert (1987, 1989, 1994, 1996, 2002) and Tollefsen (2002, 2015). For further discussion, see List (2005), List and Pettit (2011), Lackey (2016), and Kallestrup (2016).

2  Rational Golems: Collective Agents as Players in the Reasoning Game

23

typically express our beliefs by asserting their contents. So, in public argumentation an agent may justify her belief that q by asserting a further proposition p that she correctly believes and that entails q. Similarly, an agent may justify her actions by citing considerations that she correctly believes and that are the premises of good pieces of practical reasoning recommending those actions. This is the standard way of providing reasons for one’s attitudes and actions. As we will conceive of them here, normative reasons are considerations that count in favor of some attitude or response, and to which agents may appeal in order to justify such attitude or response (i.e. to vindicate its correctness).2 Paradigmatic believers are competent players of justificatory games of demanding and offering reasons. By attributing beliefs to agents, we make sense of their moves in such normative games (i.e. we rationalize the agent’s behavior). Thus, having beliefs requires being capable of participating in certain normative practices, and not so much being in any specific phenomenological or neural state. Our proposal is that it makes sense to attribute beliefs to groups that are sufficiently responsive to reasons, i.e. groups that are proficient players in a normative game of giving and asking for reasons (see González de Prado Salas and Zamora-­ Bonilla 2015). Note that what is crucial is not that there exist rational deliberations among the members of the group, but that the group itself as a unity, as a collective agent, can take part in relevant argumentative practices and engage in exchanges of reasons with other agents, so that its behavior and attitudes become evaluable as reason-sensitive. An important intuition behind our proposal is that collective agents are a type of artificial entities, something that humans create in order to achieve goals that would be otherwise difficult to attain, but that, as all other artefacts, may end having consequences that were not expected by their creators at the beginning, and in some cases may end having something like a ‘life of their own’, in the sense that they can be taken as emergent beings that, up to a point, behave in a way not totally controllable by the individuals. To say it with a classical image, the collective agent would be a kind of golem. As all artificial beings, they can function better in some cases and worse in others; they can be the result of a deliberate planning, or they can emerge and evolve more or less spontaneously; they may have very different kinds of forms, and they are irreducible to a simple set of principles that can be summarised in one clean formula (that is why we renounce to offer a precise ‘definition’ of collective agent, collective belief, or the like).3 The essential feature of a rational collective agent is that it can function as a subject in a public deliberative practice, i.e. that it can be regarded as responsive to reasons, so that it makes sense for other  See Parfit (2011), Scanlon (1998), and Raz (1975).  We also want to remain neutral about whether the constitution of a collective agent requires that its members adopt attitudes in the we-mode (Tuomela 2013; also Schmitz 2017). In general, we wish to be as non-committal as possible about how the members of a group must interact in order for the group to behave as a collective agent (i.e. as a proficient player in normative practices of giving and asking for reasons). It may well be that there is not a unique answer applying to all types of collective agent. 2 3

24

J. González de Prado and J. Zamora-Bonilla

agents to attribute to the group commitments to certain propositions and to their inferential implications. So, the idea we have in mind is that intentional attitudes may be attributed to groups that are able to play the role of competent participants in argumentative conversational practices. If we can engage in a conversation with a collective entity, then it makes sense to attribute beliefs and other attitudes to it. A modification of the classic Turing test might be helpful here. Imagine that you are performing what superficially looks like an instance of the Turing test (i.e., you are passing questions to some kind of interface, and receiving some answers from it), but in which your task is not to discriminate whether the answers are being produced by a human being or by a machine, but to tell whether these answers are produced by an individual or by a group (acting as a genuine collective agent, and not merely expressing the views of an Arrovian dictator, say). If there are groups capable of passing this Collective-Turing test (i.e. groups that cannot be distinguished by their answers from an individual interlocutor in the context of a Turing test-like process), then it is possible for groups to behave as collective agents with beliefs and other intentional attitudes – although, of course, passing the test is only a sufficient condition for being a collective agent, not a necessary one. It is important to stress that we are not assuming that there is always a rule-like algorithm (for example, an ‘aggregation mechanism’) that takes as inputs the beliefs of the members of the group and produces as an output the groups’ collective attitudes.4 There may be some groups whose belief-formation procedures are structured according to explicit algorithmic rules, for instance some voting rule. Actually, we will sometimes appeal to such groups as simple toy examples. But surely collective agents will not always include such rule-like decision mechanisms. In particular, not all groups function by making their members vote on each proposition the group has to take a stand on (in the same way that the curves of a road are not designed by making all the engineers, workers and politicians engaged in the project vote). More frequently, the collective agent will be made to hold a certain view as the result of some process of deliberation by its members, a process that can be more or less explicitly regimented, but that is certainly not always reducible to a voting rule, and in the course of which the members themselves will often change their own individual opinion, even if it does not coincide (not even in the end) with that of the group. Furthermore, many collective agents (which, remember, are a kind of tool) exist with the aim of advancing epistemic goals that are difficult to attain by means of uncoordinated individual actions, both thanks to the division of cognitive labour (e.g. Kitcher 1990; Weisberg and Muldoon 2009; Wagenknecht 2016) and to the overcoming of individual biases (e.g. Mercier and Sperber 2017). In this way, what we will often observe is that epistemic collective agents tend to be more rational than their individual members (as far as those specific epistemic goals are taken into  There are numerous mathematical proofs, in the judgment aggregation literature, showing that, given some reasonable assumptions, it is impossible to aggregate profiles of rational individual beliefs in a completely coherent way (see List and Pettit 2011). 4

2  Rational Golems: Collective Agents as Players in the Reasoning Game

25

account). In many cases, the point of creating collective agents is not to ‘preserve’ the rationality presupposed in each individual’s system of beliefs (as assumed by much of the theoretical work on judgment aggregation), but rather to overcome the shortcomings of the cognitive capacities of the individual members. One last clarification is that the constitution of the group as a collective agent does not eliminate the cognitive agency and autonomy of its members as individuals. Although groups can attain some epistemic goals that are beyond the reach of individuals (or that these can attain less efficiently), the collective agent’s attitudes are in the end derivative of the epistemic agency of its members and of the processes of deliberation that take place among them. Hence, the constitution of the collective agent’s belief does not imply that disagreements between its members cease to exist, nor that these are forced to renounce their own beliefs as autonomous individuals. The collective agent will simply be an additional subject in the public argumentative practices in which it is entitled to take part, and it will have its own normative position in such public practices. In principle, there may be disagreements among the beliefs of the collective agents and the beliefs of some (even most) of its members. And, in many cases, the dissenting members of a collective agent will have the freedom to express their dissenting voices and to act on the basis of them, at least when they do not act as public speakers or representatives of the group. These considerations will be crucial in our discussion below. In the last section of the paper, we shall examine some examples that will further illuminate the picture of collective agency that we endorse. However, our aim here is not to offer a full argument for this sort of picture (we have done so elsewhere, see González de Prado Salas and Zamora-Bonilla 2015). Rather, we want to discuss a specific objection to the idea that collective entities can play the role of believers, that is an objection to the idea that collective entities can actually take part competently in the types of practices that license belief ascription. It has been argued that the collective attitudes of groups can be at most acceptances, but not beliefs, because such attitudes are not subject to some of the rationality constraints that define what it is to believe. In particular, it seems that the doxastic attitudes of groups can be adopted voluntarily, do not aim solely at truth and can be overtly influenced by pragmatic factors. One can object that an attitude with these features does not behave as belief does, so it should be classed as a different type of attitude, perhaps an acceptance (Wray 2001, 2003; Meijers 2002, 2003; Hakli 2006; for discussion, Gilbert 2002; Tollefsen 2002; Mathiesen 2006). Our aim here is to resist this line of argument. We will argue that groups may be subject to the norms and rationality requirements distinctive of belief. In the next section we discuss the rational constraints characteristic of belief and the norms that may underlie them. After that, we show that the attitudes of collectives can actually be governed by the norms of belief, even if it is true that in the case of collectives these norms may be broken in a more blatant way than in the case of individuals. In the rest of the paper we explore the sources of these differences between individual and collective belief.

26

J. González de Prado and J. Zamora-Bonilla

2.2  Rational Belief and Practical Considerations Belief is distinctively connected with truth and evidence. Williams (1973: 148) tried to capture these distinctive connections by claiming that belief aims at truth.5 In general, it is difficult to form beliefs that you take to be false, or that go against the evidence available. If you see that it is raining, it is hard to get yourself to believe that it is not raining. Relatedly, we do not seem to be able to form beliefs at will, in the sense that we cannot decide to form a belief just because it would be desirable or useful to do so.6 As has been noted in the literature, normal agents are not usually in a position to form beliefs motivated merely by practical incentives to do so (Williams 1973; Shah 2003; Shah and Velleman 2005; McHugh 2012, 2014a). Imagine that a powerful genie offers you a great reward if you form a belief that is not supported by your evidence (say, the belief that London is the capital of Italy). No matter how great the reward is, for normal agents it would seem impossible to do as instructed by the genie and voluntarily form the requested belief. In general, there is not a straightforward route from desiring to form a belief to getting to form it. The apparent impossibility of forming beliefs as a response to practical incentives suggests that in doxastic deliberation only epistemic considerations can be recognized as reasons to believe (where epistemic considerations are considerations that concern the truth of the relevant proposition). Practical considerations having to do with the desirability of forming the belief that p cannot weigh in the deliberative process as reasons to believe that p (Shah and Velleman 2005; Hieronymi 2008). Of course, practical factors may have a causal impact in the formation of beliefs, for instance in the form of wishful thinking or biases (e.g. confirmation bias). Likewise, practical considerations may move the agent to try to put herself in a position where she is likely to form the required belief (say, she can undergo hypnosis, experiment with different forms of self-suggestion or expose herself to being brainwashed). Yet in general it is not easy to make you form a belief that you take to be unsupported by evidential considerations – you can only do so in indirect, roundabout ways. It seems that, when we engage in explicit deliberation about what to believe, we cannot effectively treat practical considerations as reasons to believe some proposition (Williams 1973). Why is the deliberative formation of beliefs not responsive to practical reasons, at least in an overt, self-aware way? A possible line of response is that doxastic deliberation is transparent, in the sense that deliberating about whether to believe p automatically gives way to deliberation about whether p (that is, about whether p is true).7 According to Hieronymi (2008), beliefs about p embody the agent’s answer  For discussion of this idea and different ways of developing it, see among others Wedgwood (2002), Shah and Velleman (2005), Owens (2003), Steglich-Petersen (2006), Boghossian (2008), McHugh (2014b), and Whiting (2010). 6  This does not mean that we lack control over our beliefs. It can be argued that beliefs are under rational control, insofar as we form them in response to the reasons available to us (McHugh 2012, 2014a; Hieronymi 2008). 7  For proposals along these lines, see Shah and Velleman (2005) and Hieronymi (2008). 5

2  Rational Golems: Collective Agents as Players in the Reasoning Game

27

to the question whether p. Thus, deliberating about what to believe in relation to p amounts to considering what is the correct answer to the question ‘p?’.8 Arguably, the correctness of an answer to the question ‘p?’ hinges on the truth of p. Therefore, only evidential considerations bearing on the truth of p are relevant for settling this type of question. In this way, to the extent that the belief that p is an answer to the deliberative question ‘p?’, it is only correct if p is true (see Whiting 2010; McHugh 2014b). This would be why only epistemic considerations can be openly recognized as reasons to believe when engaging in doxastic deliberation. Moreover, the existence of a truth-involving standard of correctness governing beliefs would explain why beliefs that happen to be formed in a way that disregards the evidence (say, because of some recalcitrant prejudice) are assessed as incorrect  – as somehow epistemically defective. To be sure, the difficulty to form beliefs voluntarily or against the evidence available might be taken as a mere psychological fact, or at any rate a fact that is not to be accounted for in normative terms (see Bykvist and Hattiangadi 2007; Glüer and Wikforss 2013). However, as the last paragraph shows, a natural explanation of why we can only believe for epistemic reasons (and not for practical ones) appeals to some truth-involving normative standard governing belief – for instance, the norm that beliefs are correct, appropriate or permissible only if true. According to the explanation given above, when deliberating about what to believe regarding p, agents are trying to answer the question whether p. Insofar as answers to this question are governed by a truth-involving standard of correctness, only evidential considerations can properly settle it. In what follows, we assume that this sort of truth-involving normative standard is characteristic of the attitude of believing.9

2.3  The Atypical Features of Collective Belief Many of the distinctive features of individual belief discussed in the previous section are not found in the belief-like attitudes attributed to collective entities (for ease of exposition, we will keep calling such attitudes collective beliefs, suspending judgment for the moment about whether they constitute genuine beliefs or not). It seems that practical considerations can affect the formation of collective beliefs in a far more direct, widespread and flagrant way than what is observed in the case of individual believers. In particular, there is no obvious barrier to the adoption of collective beliefs as a response to practical incentives – in this sense, collective beliefs

 Take ‘p?’ as shorhand for a grammatically correct formulation of the question whether p. Indeed, as a matter of psychological fact, individuals hardly ever deliberate asking explicitly to themselves ‘Should I believe that p?’, rather than simply ‘p?’. The sort of transparency we are discussing would explain why this is the case. 9  Most of the main points we want to make could also be made in terms of weaker norms that do not require beliefs to be true, for instance the evidentialist norm that one ought to adjust one’s beliefs to the evidence available (see Conee and Feldman 2004). We will focus on truth-involving norms, for the sake of simplicity, and because of their plausibility. 8

28

J. González de Prado and J. Zamora-Bonilla

seem to be voluntary in a way that individual beliefs are not (Wray 2001, 2003; Meijers 2002, 2003, Hakli 2006). Imagine, as a simple toy example, a group where majority rule is the mechanism to decide what collective belief is to be adopted: the members of the group vote in favour or against adopting the collective belief that p, and the option supported by more votes is chosen. Now, in principle all kinds of considerations can motivate the members of the group to vote one way or another. Perhaps they are only moved by epistemic considerations, but they can also have in mind practical considerations, such as the potential benefits of getting the group to adopt the belief that p. In this way, the members of the group can (rationally) decide to vote in favour of the collective belief that p, even if as individuals they believe that ¬p (and take the evidence to support such a belief). If the members of this type of group are offered a reward for making the group adopt the collective belief that p, they can easily comply: they just need to vote accordingly. The lesson of the example does not depend on the specific details about the group’s decision mechanisms. Something similar could happen in a group where decisions about what to (collectively) believe are reached via internal deliberations among the members (the members argue with each other until they agree on adopting a certain collective doxastic attitude). Again, the members may be moved by practical considerations when deciding what position they will defend in this intersubjective deliberative process – they may decide to advocate a view that they know is actually false and goes against the available evidence. It may be argued that these examples only show that the collective beliefs of a group can be voluntarily controlled by its members – so that the practical concerns of the members can play a direct role in the adoption of the collective belief. Yet the examples would not show that the group, as a collective agent, can form beliefs at its own will, or that the group can recognize practical considerations as reasons to believe (see Gilbert 2002; Gilbert and Pilchman 2014). One first thing to say is that, even if this were so, practical factors (more specifically, the practical interests of the members) would still play a more direct and pervasive role than in the case of individual beliefs. Moreover, there is no reason to think that there cannot be groups where beliefs are routinely formed directly in response to the group’s desires and intentions to form them. We can imagine a group where there are two different decision mechanisms, the first one for forming collective desires to adopt collective beliefs, and the second one for adopting collective beliefs. It could happen that the outcomes of the first decision process serve as inputs for the second one, so that the (collective) desires to believe formed via the first decision process directly determine the collective beliefs produced as outcomes in the second decision process (the group just adopts the beliefs that it desires to adopt). Furthermore, it could be that, in the context of public justificatory practices, the group cites the practical considerations that backed its collective desires to believe as its reasons for forming the corresponding collective beliefs. In a case like this, the group could be said to be forming beliefs motivated by practical considerations concerning the desirability of forming such beliefs (which, of course, does not mean that these practical considerations are actually good reasons to adopt the relevant beliefs). Perhaps these types of examples are not best described as cases where the group recognizes practical

2  Rational Golems: Collective Agents as Players in the Reasoning Game

29

considerations as reasons to believe (i.e. as considerations that favour the belief and make it fitting), but rather as cases where the group has practical reasons to desire to adopt a belief (i.e., a cognitive commitment), and manages to directly form such a belief in response to this desire. However, this would still constitute a significant difference with individual belief-formation, where it is not typically possible to go directly from a desire to form a belief to actually forming it. The dissimilarities between individual beliefs and the belief-like attitudes of collectives may make one conclude that such collective attitudes are not actually beliefs, but some other type of attitude  – for instance, some form of acceptance (Wray 2001, 2003; Hakli 2006; Meijers 2002, 2003). There are different ways of characterizing the attitude of acceptance, but for our purposes here it is enough to think of it as an attitude that is similar to belief in some of its inferential and practical implications, but that only involves endorsing its content in certain restricted contexts and for certain specific purposes.10 In this way, a lawyer may accept the innocence of her client only in the context of defending him, but not when she is not acting as a lawyer. Likewise, a scientist may accept some claim in the context of a certain research project, perhaps due to its predictive success, even if she remains neutral about its truth (or even if she knows the claim to be strictly false, although perhaps a good approximation). As these examples show, accepting a claim is compatible with not believing it. Moreover, in general an agent can decide at will whether to accept some claim (for certain purposes). After all, we can accept a claim merely for the sake of argument in order to see what follows from it. And scientists may accept some view as a research assumption motivated by the funding prospects associated with undertaking a research project presupposing such a view. In the contexts where the agent’s acceptance of p is in play, her behaviour may be similar in many respects to that of an agent that believes p. So, the agent may be disposed to assert that p, to offer support for p in the face of appropriate challenges and to act as if p were true. If this is so, it could seem that we can properly characterize the behaviour of a group with a belief-like attitude towards p by attributing to it an acceptance of p, rather than a belief that p. In this way, there would be no problem in granting that the attitude of the group towards p may be adopted at will, on the basis of practical considerations. In the next section, however, we argue that this is not a satisfactory approach. Unlike acceptances, the belief-like attitudes of groups are subject to the normative standards governing belief.

2.4  Acceptance and the Norm of Belief As we have suggested above, it is plausible to think that belief is governed by a truth-involving normative standard, so that believing involves a context-­independent commitment to the truth of the proposition believed. If you believe p but at the same time take it to be false, you are bound to have an incorrect attitude.

10

 For discussion on the notion of acceptance, see Cohen (1995) and van Fraassen (1980).

30

J. González de Prado and J. Zamora-Bonilla

Correctly believing that p is incompatible with correctly believing that p is false (or with correctly believing ¬p). This unrestricted commitment to truth is not found in the attitude of acceptance characterized in the previous section. An agent can always correctly accept p (in a certain context, for certain purposes) while leaving open the possibility that p is false, indeed even while believing that p is false – for instance, we may accept p as a useful approximation. In this way, an agent may correctly accept, in different contexts, each of two incompatible propositions (i.e. propositions that cannot be both true). For example, it may be that, when tackling a problem that does not require precise results, a scientist accepts some claim that works well as a simplifying approximation, whereas she accepts instead some more cumbersome non-­simplified claim in the context of dealing with other problems where precision is paramount. Now, if the belief-like attitudes of groups were to be modeled always as acceptances rather than genuine beliefs, we would expect groups to be able to adopt such attitudes towards incompatible propositions, in different contexts, without having to reject one of the attitudes as incorrect. Moreover, it would not be possible to say of groups that they accepts some proposition as true, although they actually believe it is false, like we can say of individuals – for groups would never really believe anything, they would only accept propositions in certain (but not necessarily all) contexts. Individuals can make a distinction between what they believe and what they merely accept, and the pragmatic-inferential norms they are publicly subjected to when forming and revising their beliefs are noticeably different from the norms they must follow when managing mere acceptances. What we want to argue is that this pragmatic difference between the norms governing beliefs and the norms governing mere acceptances also applies to group agents. To be clear, we are not arguing that groups cannot accept propositions as true without believing them. In principle, groups will be able to do this, in the same way that individuals can. Our claim is that groups also adopt belief-like attitudes that are subject to normative constraints different from the norms operative for mere acceptances. More specifically, the belief-like attitudes attributed to groups seem to be sensitive to the sort of truth-involving normative standards governing belief. We have argued that groups may adopt belief-like attitudes that go blatantly against the available evidence (perhaps moved by practical incentives). However, these attitudes will be assessed as incorrect and may be criticized by other participants in the relevant practices of giving and asking for reasons. Imagine, for instance, that the government expresses a belief-like attitude towards the claim that the unemployment rate has been reduced, when the evidence clearly indicates that it has gone up. Surely, the government’s attitude will be criticizable, given its disregard for the available evidence. Moreover, the correctness of collective belief-like attitudes that ignore the evidence available will not be vindicated by citing practical incentives to form that attitude. Such practical considerations will not be taken to bear on the correctness of the group’s belief-like attitude. Thus, the belief-like attitudes of groups that depart from the normative standards governing belief will be typically seen as violations of such standards.

2  Rational Golems: Collective Agents as Players in the Reasoning Game

31

In light of these observations, we can conclude that the belief-like attitudes of groups are indeed subject to the norms of belief. What is characteristic of beliefs is not that they always satisfy such norms, but rather that they are always subject to them. After all, these norms are often transgressed by the beliefs of individual agents, which nonetheless keep counting as beliefs – think for instance of cases of bias or wishful thinking. So, we submit that the belief-like attitudes of groups do not behave as non-doxastic acceptances, but rather play the same normative role as the beliefs of individual agents (see Mathiesen 2006). Therefore, we should not be wary of regarding such attitudes as collective beliefs. To be sure, there remains the question of why collective agents can flout the norms of belief in more overt, glaring ways than individuals. We discuss this issue in the next section.

2.5  Non-transparent Collective Doxastic Deliberation In Sect. 2.2 we argued that the fact that we cannot easily form beliefs at will, moved merely by practical considerations, is related to the transparency of doxastic deliberation. The belief that p constitutes an answer to the question ‘p?’, so deliberating about what doxastic attitude towards p is correct amounts to deliberating about whether p. We want to suggest that things work differently in the case of collective beliefs. Typically, collective doxastic deliberation is not transparent, because it is mediated by the deliberative choices of the members of the group. This may make groups more vulnerable than individuals to the influence of practical considerations when forming their (collective) beliefs, constituting a possible source of epistemic failure against which individuals tend to be more protected by their natural psychological constitution. In general, collective beliefs are formed as a result of the actions performed by some of the members in some decision-making process. A simple example, discussed above, is a group where members vote in order to decide what collective belief to adopt. But the actions of members will also be involved in more sophisticated methods of collective belief formation. Think of a group where collective beliefs are adopted via a deliberative process in which the members argue for different positions until an agreement is reached. The resulting collective beliefs will depend on what positions the members decide to defend in the negotiation. Call the actions performed by members of a group in the context of collective decision-­ making processes deliberative actions. The crucial point is that these deliberative actions performed by the members may be rationally guided not only by evidential considerations about whether p, but also by further practical considerations concerning whether it is desirable that the group comes to believe that p. The question considered by the members of the group is not directly ‘p?’, but ‘Is it desirable that the group believes that p?’ So, in collective deliberations, the members will consider the question about what doxastic attitude is desirable for the group to adopt, and will act in ways that promote the group’s adoption of such an attitude (e.g. they will vote for it). It may well be that the individual members have good (practical) reasons to

32

J. González de Prado and J. Zamora-Bonilla

foster a collective belief that they know is false or goes against the evidence, and that will be irrational at the collective level (for instance, in the sense that it will be incoherent with other things believed by the group). For example, some members of the group may have a practical interest in getting the group to act on certain beliefs they know to be false – say, because the political agenda of these members will be favored if the group acts on such beliefs. Thus, it may be perfectly rational for the members of the group (as individuals) to try to manipulate the collective deliberative process so that the group adopts an irrational attitude – in particular, a belief that goes against the available evidence. Perhaps there is some limit to the amount of internal manipulation that a group can suffer before losing its coherence as unified, rational collective agent. Yet the point remains that deliberations within groups tend to be more exposed to the direct influence of practical considerations than in the case of individuals, given that they typically proceed via the deliberative actions of the group members. The situation is analogous to cases where an individual agent has practical reasons to make other agents believe something false or against the evidence. An agent may have good practical reasons to lie, with the intention of inducing false beliefs in others (say, by lying the agent may be trying to mislead a killer searching for an innocent victim). Likewise, if you are a skilled hypnotist, you may have (practical) reasons to cause some agent to form a certain irrational belief, perhaps to avoid her emotional distress. As noted in Sect. 2.2, an agent can also have practical reasons to try to influence her own beliefs in these sorts of ways. However, when the agent is directly deliberating about the correct answer to the question ‘q?’, practical factors can at most play an implicit causal role. And, in general, it is difficult for an individual agent to manipulate successfully her own beliefs against the evidence she acknowledges. By contrast, as we have seen, in principle there may be groups where the members directly manipulate the group’s doxastic commitments, guided by their practical interests and going against the evidence that they (as individuals) acknowledge. To be sure, in many cases the members of a group will have exclusively epistemic goals when engaging in collective doxastic deliberation, so that their deliberative actions will be guided by epistemic considerations (Mathiesen 2006). And when this is so, collective deliberation can afford epistemic goods that would not be available to isolated individual agents (as we will show in next section). What we want to argue here is that there may also be other cases where the deliberative actions of the members of the group respond to non-epistemic, practical considerations. When this happens, rational deliberative actions by the members of the group may lead to irrational attitudes at the collective level (say collective beliefs against the available evidence, or incoherent collective attitudes). It should also be noted that in groups with defective decision-making procedures, the group may have blatantly irrational beliefs even if the agents perform rational deliberative actions guided only by epistemic considerations. For instance, it is well known that decisions by voting may lead to different voting paradoxes. In a group where beliefs are formed by voting, incoherent collective beliefs may result from

2  Rational Golems: Collective Agents as Players in the Reasoning Game

33

votes that reflect the members’ coherent individual doxastic attitudes (see List 2005; List and Pettit 2011). Again, in these cases the flagrant irrationality of the collective belief has its roots in the non-transparency of the deliberative process through which such a belief was adopted. Regardless of whether the members have purely epistemic aims, the beliefs of the group are formed as a function of the views of other agents (the members), rather than as a result of a transparent deliberation in which the same agent that is to adopt the belief considers whether p. This may open the door to forms of irrationality that are less frequent in individual doxastic deliberation. Of course, it is to be expected that some collective agents will implement mechanisms aimed at protecting themselves from such risks of irrationality, to the extent that the rationality of these agents is an important factor for their success or their very survival. In general, successful collective agents will resort to strategies for limiting the impact of major sources of irrationality. To sum up, group deliberation is non-transparent in a way that individual doxastic deliberation is not, given that it goes through the individual deliberation of the members about what deliberative action they should perform in order to influence the group’s belief-formation in desirable ways. This explains why practical factors can play a more explicit and overt role in the formation of group beliefs.

2.6  The Epistemic Virtues of Collective Agency We have argued that it is possible that the collective beliefs of a group are blatantly irrational and overtly break the norms governing belief (in a way not to be found in normal individuals), even if the attitudes of the members of the groups remain perfectly rational. However, this point should not obscure the fact that collective entities are often formed precisely to allow their members to improve their epistemic position, and to avoid epistemic shortcomings associated with individual perspectives. An obvious way in which collective agents may go beyond the epistemic limitations of individuals is by pooling the evidence and epistemic skills of their members. However, plausibly there are other, less evident epistemic advantages associated with collective agency. We conclude the paper by offering a few examples of collective believers, with the aim of illustrating some of the potential epistemic benefits of collective agency. The collective constituted by the co-authors of a scientific paper (in particular, of a two-authored paper) offers a clear example of an ‘epistemic collective agent’ providing access to epistemic rewards beyond the reach of individual agents (see Zamora-Bonilla 2014). Co-authorship may be justified in many different ways, but the most common and obvious one is by appeal to the division of cognitive labour. In general, the collaborating authors will have different epistemic capacities and resources – each one may know more than the other about some of the topics that need to be tackled in the paper. Because of this, co-authorship often entails some degree of mutual trust between the authors: regarding some claims the paper makes,

34

J. González de Prado and J. Zamora-Bonilla

one author will be more capable of offering a competent defence than the other, even if the defence that is actually written is the result of a deliberation between both. This is more evident if we consider the paper not as a finished and perfect work, but as a single piece in a longer public conversation: if the paper is challenged in the future by other academics, the co-authors’ responses need not be themselves co-authored (though they of course may), and obviously each co-author may end up offering different arguments or interpretations about what they wrote together. However, as one of us recently argued, co-authorship can be shown to have epistemic advantages even in cases when the division of cognitive labour is not the main reason behind the collaboration (Zamora-Bonilla 2014). This is particularly clear if we picture scientific inquiry as taking place within normative practices of giving and asking for reasons. Think of the scientific paper not as an isolated item of knowledge, but as an argument which is a portion of a longer social process of deliberation, justification, criticism and contestation. In this type of social argumentative practice, past commitments (by some members of the scientific community) are appealed to in order to justify further commitments (not necessarily by the same individuals), and the credit (or symbolic reward) a certain scientist gets depends on the use that further researchers make of the items of knowledge that such a scientist has advanced or ‘authored’. Assuming this sort of picture of scientific practice, it is plausible to think that many instances of co-authorship will be explained because the authors prefer to be recognised and rewarded by something that is a consequence of their individual contributions taken together, but that does not follow from each author’s ‘part’ in isolation. Imagine, as a simple example, that author A has proved that p entails q, and author B has proved that p, but both prefer to be recognised as the joint discoverers of q (say, an important mathematical theorem), rather than as the individual discoverers of the truth of the less newsworthy claims ‘p’ and ‘p entails q’. Hence, the epistemic motivations for co-authorship do not only derive from the trivial fact that different individuals may have different, complementary capacities and resources, but from an important result about the cognitive division of labour: the fact that in general the ‘solution’ to some research question does not arise as a mere juxtaposition of single items of knowledge, but as an inferential consequence of the premises and arguments provided by the co-authors. The proposition q constituting the solution to the problem the co-authors wanted to solve is, obviously, believed by both A and B, but the ‘social status’ of that proposition (as an item of knowledge within the ‘long conversation’ in which the activity of their scientific discipline consists of) may be legitimately taken as something that ‘belongs’ to both co-authors as a single subject – i.e. it is the collective agent formed by A and B (A&B) who is entitled to get the credit for providing a proof for q. And, of course, since the acceptance of q by the rest of the scientists will mainly depend on the epistemic quality of the arguments offered in A&B’s paper, A and B have a strong incentive to let the paper contain the soundest arguments they can elaborate, as far as the practice of evaluating scientific claims within their scientific discipline is governed by strict epistemic norms.

2  Rational Golems: Collective Agents as Players in the Reasoning Game

35

The explicit appeal to collective beliefs in relation to co-authorship is evident when we consider a case in which the scientific community challenges and perhaps ends up rejecting A&B’s conclusion – due to further evidence and arguments available after the publication of the paper by A&B. In cases like this, it would make sense to say: ‘Even if A&B believed that q, we now know that q is false’. As we argued in Sect. 2.1, belief ascriptions in this type of context should not be seen as making reference to any phenomenological or neural state, but rather as a way in which an interpreter keeps track of the commitments of (individual or collective) agents, without the interpreter having to undertake such commitments. Co-authorship teams can be seen as the simplest collective unit of scientific agency. At the other end of the spectrum we find the entity constituted by a whole scientific discipline. Disciplines, we think, offer a further example of how collective epistemic practices afford potential epistemic rewards not available to individuals working in isolation. One may be initially doubtful that scientific disciplines behave as unified collective agents – at least, this is not as clear as in the case of co-authorship. In particular, disciplines as such will not always be able to undertake commitments and defend them in cohesive ways in argumentative social practices. However, there is a sense in which scientific disciplines come closer to constitute a collective agent. Disciplines sometimes function as custodians of the consensus views emerging in the field and of the arguments backing those views. Indeed, it may be argued that disciplines may be legitimately taken as the real depositaries of the scientific knowledge in an area of science, in the sense that they have the highest authority when some item of knowledge belonging to that area is appealed to or discussed in the rest of society (and, to begin with, in other scientific disciplines). After all, it may well be that no individual scientist has the means to acquire herself all the knowledge of the discipline and to master the different epistemic skills and resources underpinning such knowledge. As opposed to what happened with teams of co-authors, in the case of disciplines the boundaries of the group are not even well defined: who is a legitimate member of the group, and how much voice a given member has in the determination of the group’s positions, are often questions without definite answers. But that is only a problem for judgment aggregation views of collective attitudes, which reduce the formation of a collective belief to a set of voting rules or other algorithmic formal procedures. Our approach to collective agents does not presuppose any of that formal machinery (except in those cases where the group explicitly adopts it), but allows us to think of a collective agent as an entity defined by the normative position it occupies in a practice of giving and asking for reasons. In this sense, the scientific discipline (organised through more or less fluid structures like associations, congresses, textbooks and syllabi) defines, both for its members and for the outsiders, what claims and arguments must be taken as valid, which ones as merely conjectural, which others as admissible-but-not-compulsory etc. It also determines when an individual is allowed to act as a spokesperson for the scientific field (i.e. when an individual counts as a qualified expert). And, though some positions that are endorsed by the scientific discipline will be endorsed by all or by most of its

36

J. González de Prado and J. Zamora-Bonilla

members, it will often be the case that, due to the division of labour within it, many things that ‘are known’, are really unknown by most of the individual scientists belonging to the field (because they lie outside of their more specific area of specialization) – not to mention those cases in which, though the community accepts some claim, it also allows some degree of dissent about it or about its correct interpretation. Scientific disciplines illustrate the potential epistemic advantages of engaging in collective scientific practices. Individual researchers clearly benefit epistemically from working within the context of consolidated scientific disciplines. In addition to providing scientists with access to evidence, arguments and epistemic resources beyond the reach of a single isolated researcher, the intersubjective networks constituting disciplines may also function as a tool for correcting and limiting some forms of bias and irrationality that are difficult to avoid for individuals. Think for instance of the phenomenon known as confirmation bias. Researchers tend to give more weight and pay more attention to evidence that confirms their views. However, within a certain scientific discipline there will typically be researchers defending alternative positions. So, even if a researcher favouring a view p may tend to disregard evidence against that view, there will be other researchers that defend ¬p and will therefore be inclined to highlight evidence against p. Thus, it can be expected that in the sort of public intersubjective disputes that underlie the formation of disciplines and the emergence of consensus views, the potential confirmation biases of different individual researchers will be checked and corrected. Indeed, confirmation bias may even become epistemically advantageous in certain collective contexts, in that it may lead to a beneficial form of division of cognitive labour. Agents would specialize in finding evidence supporting the claims they defend, which will be then scrutinized and criticized by other agents defending alternative claims. Thus, the resulting views collectively regarded as the established position in the discipline will be free of some of the biases and sources of irrationality that tend to affect individual views. This is not to say that some biases present at the individual level may not be reflected at the collective level as well, or that collective scientific practices may not introduce their own form of bias. Yet it seems that collective scientific interactions will help scientists avoid some of the potential epistemic limitations of individual research.

Bibliography Boghossian, P. (2008). Content and justification. Oxford: Oxford University Press. Brandom, R. (1994). Making it explicit. Cambridge, MA: Harvard University Press. Brandom, R. (2000). Articulating reasons. Cambridge, MA: Harvard University Press. Bykvist, K., & Hattiangadi, A. (2007). Does thought imply ought? Analysis, 67, 277–285. Cohen, L. J. (1995). Essay on belief and acceptance. Oxford: Oxford University Press. Collins, S. (2013). Collectives’ duties and collectivization duties. Australasian Journal of Philosophy, 91(2), 231–248.

2  Rational Golems: Collective Agents as Players in the Reasoning Game

37

Collins, S. (2017). Duties of group agents and group members. Journal of Social Philosophy, 48(1), 38–57. Conee, E., & Feldman, R. (2004). Evidentialism: Essays in epistemology. Oxford: Clarendon Press. Gilbert, M. (1987). Modelling collective belief. Synthese, 73(1), 185–204. Gilbert, M. (1989). On social facts. Princeton: Princeton University Press. Gilbert, M. (1994). Remarks on collective belief. In F. Schmitt (Ed.), Socializing epistemology: The social dimensions of knowledge (pp. 111–134). Lanham: Rowman and Littlefield. Gilbert, M. (1996). Living together: Rationality, sociality, and obligation. Lanham: Rowman and Littlefield. Gilbert, M. (2002). Belief and acceptance as features of groups. PRO, 16, 35–69. Gilbert, M., & Pilchman, D. (2014). Belief, acceptance, and what happens in groups. In J. Lackey (Ed.), Essays in collective epistemology. Oxford: Oxford University Press. Glüer, K., & Wikforss, Å. (2013). Against belief normativity. In T. Chan (Ed.), The aim of belief (pp. 80–99). Oxford: Oxford University Press. González de Prado Salas, J., & Zamora-Bonilla, J. (2015). Collective actors without collective minds: An inferentialist approach. Philosophy of the Social Sciences, 45, 3–25. Hakli, R. (2006). Group beliefs and the distinction between belief and acceptance. Cognitive Systems Research, 7(2–3), 286–297. Hieronymi, P. (2008). Responsibility for believing. Synthese, 161(3), 357–373. Kallestrup, J. (2016). Group virtue epistemology. Synthese, 1–19. https://doi.org/10.1007/ s11229-016-1225-7. Kitcher, P. (1990). The division of cognitive labor. The Journal of Philosophy, 87(1), 5–22. Lackey, J. (2016). What is justified group belief? Philosophical Review, 125(3), 341–396. List, C. (2005). Group knowledge and group rationality: A judgment aggregation perspective. Episteme, 2(1), 25–38. List, C., & Pettit, P. (2011). Group agency: The possibility, design, and status of corporate agents. Oxford: Oxford University Press. Mathiesen, K. (2006). The epistemic features of group belief. Episteme, 2(3), 161–175. McHugh, C. (2012). Epistemic deontology and voluntariness. Erkenntnis, 77(1), 65–94. McHugh, C. (2014a). Exercising doxastic freedom. Philosophy and Phenomenological Research, 88(1), 1–37. McHugh, C. (2014b). Fitting belief. Proceedings of the Aristotelian Society, 114, 167–187. Meijers, A. (2002). Collective agents and cognitive attitudes. ProtoSociology, 16, 70–85. Meijers, A. (2003). Why accept collective beliefs?: Reply to Gilbert. PRO, 18, 377–388. Mercier, H., & Sperber, D. (2017). The enigma of reason. Cambridge, MA: Harvard University Press. Owens, D. (2003). Does belief have an aim? Philosophical Studies, 115, 283–305. Parfit, D. (2011). On what matters. Oxford: Oxford University Press. Raz, J. (1975). Practical reason and norms. Oxford: Oxford University Press. Scanlon, T. (1998). What we owe to each other. Cambridge, MA: Harvard University Press. Schmitz, M. (2017). What is a mode account of collective intentionality? In G. Preyer & G. Peter (Eds.), Social ontology and collective intentionality (pp. 37–70). Cham: Springer. Sellars, W. (1956). Empiricism and the philosophy of mind. Minnesota Studies in the Philosophy of Science, 1(19), 253–329. Shah, N. (2003). How Truth Governs Belief. Philosophical Review, 112(4), 447–482. Shah, N., & Velleman, J.  D. (2005). Doxastic deliberation. The Philosophical Review, 114(4), 497–534. Steglich-Petersen, A. (2006). No norm needed: On the aim of belief. The Philosophical Quarterly, 56, 499–516. Tollefsen, D. P. (2002). Challenging epistemic individualism. PRO, 16, 86–117. Tollefsen, D. P. (2015). Groups as agents. New York: Wiley. Tuomela, R. (2013). Social ontology: Collective intentionality and group agents. Oxford: Oxford University Press. Van Fraassen, B. (1980). The scientific image. Oxford: Clarendon.

38

J. González de Prado and J. Zamora-Bonilla

Wagenknecht, S. (2016). Social epistemology of research groups. London: Palgrave Macmillan. Wedgwood, R. (2002). The aim of belief. Noûs, 36, 267–297. Weisberg, M., & Muldoon, R. (2009). Epistemic landscapes and the division of cognitive labor. Philosophy of Science, 76(2), 225–252. Whiting, D. (2010). Should I believe the truth? Dialectica, 64(2), 213–224. Williams, B. (1973). Problems of the self. Cambridge: Cambridge University Press. Wray, K. B. (2001). Collective belief and acceptance. Synthese, 129(3), 319–333. Wray, K. B. (2003). What really divides Gilbert and the rejectionists? PRO, 18, 363–376. Zamora-Bonilla, J. (2014). The nature of co-authorship: A note on recognition sharing and scientific argumentation. Synthese, 191, 97–108.

Chapter 3

Trust and Commitment in Collective Testimony Leo Townsend

Abstract  In this paper I critically discuss Miranda Fricker’s ‘trust-based’ view of collective testimony—that is, testimony that comes from a group speaker. At the heart of Fricker’s account is the idea that testimony involves an ‘interpersonal deal of trust’, to which the speaker contributes a commitment to ‘second-personal epistemic trustworthiness’. Appropriating Margaret Gilbert’s concept of joint commitment, Fricker suggests that groups too can make such commitments, and hence that they, like individuals, can ‘enter into the second-personal relations of trust that characterise testimony’ (Fricker, Philos Phenomenol Res 84:272, 2012). I argue that this choice to appropriate Gilbert’s concept of joint commitment betrays a deep problem in Fricker’s account—a misconstrual of both the object and the subject(s) of the commitment a speaker makes in testifying. After developing this criticism, I outline an alternative way of construing the speaker’s commitment, which can be applied to both collective and individual testimony. Keywords  Collective testimony · Epistemic trust · Joint commitment · Assurance view

3.1  Introduction In recent years a number of philosophers have advanced the idea that trust is central to testimony—that coming to believe something on someone’s say-so is a matter of trusting the speaker for the truth.1 There are various different ways of spelling this out, but one leading proposal is that the speaker, in testifying, extends a kind of offer or invitation to her audience to trust her, which the audience, if she believes the speaker, takes up. Seen like this, testimony aims at establishing trusting relations   See, esp. Holton (1994), Hinchman (2005), Hieronymi (2008), Faulkner (2011), and McMyler (2011). 1

L. Townsend (*) Department of Philosophy, University of Vienna, Vienna, Austria e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_3

39

40

L. Townsend

between speaker and hearer, relations that give rise to a distinctively ‘I-thou’ or ‘second-personal’ form of normativity. The act of telling someone something is then one side of a transaction, which, if completed, effects a normative shift in the way speaker and hearer stand towards one another. If the invitation to trust is accepted, the speaker then becomes answerable to her audience for the truth of what she says, and perhaps also a suitable target of reactive attitudes and blame if what she tells him turns out to be false. Correlatively, the audience, in accepting the speaker’s invitation to trust, acquires a right to challenge the speaker, and to defer challenges to his testimonially-acquired beliefs back to her. Miranda Fricker (2012) has developed a view of testimony along these lines and has attempted to extend it to collective testimony—that is, to cases in which testimony is offered by, or in the name of, a group.2 At the heart of Fricker’s account is the idea that testimony has its home in what she calls an ‘interpersonal deal of trust’ (Fricker 2012: 258), to which the speaker, in telling her audience that p, contributes a commitment to ‘second-personal trustworthiness concerning whether p’ (Fricker 2012: 271). Appropriating Margaret Gilbert’s concept of joint commitment, Fricker then suggests that groups too can make such commitments, and hence that they, like individuals, can ‘enter into the second-personal relations of trust that characterise testimony’ (Fricker 2012: 272). In this paper I critically discuss Fricker’s approach to the nature of collective testimony. In short, I think that Fricker is right to focus on the ‘second-personal’ or ‘I-thou’ features of testimony, and to attempt to ground those features in some sort of commitment on the part of the speaker. But she goes wrong, in my view, in how she construes the speaker’s commitment—as a commitment to proving worthy of the audience’s ‘second personal epistemic trust’, rather than as a commitment to the truth of what she says, made to a second person.3 I begin by detailing Fricker’s ‘trust-based’ view of the speech act of testimony (Sect. 3.2.1), and explaining how she extends this view, with the help of some familiar Gilbertian machinery, to cases of collective testimony (Sect. 3.2.2). I then argue that this choice to appropriate Gilbert’s concept of joint commitment betrays a problem in Fricker’s account—a misconstrual of both the object (Sect. 3.3.1) and the subject(s) (Sect. 3.3.2) of the commitment a speaker makes in testifying. I end off by outlining an alternative way

 Note that Fricker’s paper is about both the nature of the speech act of testimony and the epistemology of testimonially-acquired beliefs. In this paper I set the epistemological issues aside and focus only on her view of the speech act. 3  I do not mean to overstate the contrast between Fricker’s account and the account I try to develop in its stead. In particular, I do not mean to suggest that an interpersonal commitment to the truth of p stands in some kind of opposition to a commitment to interpersonal epistemic trustworthiness. On the contrary, if you are committed to proving worthy of your interlocutor’s epistemic trust then you will naturally also be committed to the truth of what you tell her. So I am not proposing a wholesale rejection of Fricker’s account, only a change in emphasis. More specifically, much of my critical argument should be seen as taking issue with Fricker’s construal of the speaker’s commitment as first and foremost a commitment to trustworthiness, and only derivatively a commitment to the truth of what is said. Thanks to Preston Stovall for prompting me to clarify this. 2

3  Trust and Commitment in Collective Testimony

41

of construing the speaker’s commitment, which I think does a better job, and which can be applied to both collective and individual testimony (Sect. 3.4).

3.2  Fricker’s Account of Collective Testimony 3.2.1  The Deal of Second Personal Epistemic Trust Fricker sees her account as a version of the so-called ‘assurance view’ of testimony.4 Like other proponents of this approach, she ties the distinctiveness of testimony as a source of knowledge to certain features of the relationship between the testifier and audience. A central idea here is that the act of testimony somehow renders the speaker responsible to the audience, and hence that the audience acquires a correlative right against the speaker. Richard Moran, for instance, puts this in terms of the speaker ‘making himself accountable, conferring a right of complaint on his audience should his claim be false’ (Moran 2005: 21). And in a somewhat similar vein, Benjamin McMyler (2011) claims that testimony accords the audience certain rights and entitlements relating to the justification of her testimonially-acquired beliefs: In acquiring knowledge and justified belief on the basis of testimony, an audience is entitled to defer epistemic challenges to her testimonial beliefs back to the original testifier. (McMyler 2011: 7)

It is important to note that these features—the responsibility of the speaker, and the correlative rights and entitlements of the audience—are interpersonally directed normative statuses. The speaker owes the truth of what she says to her audience, and the audience holds a right of complaint or entitlement to defer challenges against the speaker. And this is thought, by some proponents of this general view, to make both speaker and audience susceptible to certain ‘reactive’ attitudes, especially when things do not go well in the testimonial exchange. If, for instance, the speaker speaks falsely, this may license a ‘sense of personal betrayal’ (Fricker 2012: 259) on the part of the audience, and, similarly, if the speaker’s testimony is refused by the audience, this may amount to a certain kind of ‘insult’ (Anscombe 1979) or ‘slight’ (Hinchman 2005). It is these ‘second personal’ or ‘I-thou’ features of testimony— the directed rights and obligations of speaker and hearer, and the attendant reactive attitudes—that proponents of the assurance view are typically concerned to highlight and explain. For her part, Fricker attempts to account for these features of testimony by developing what she calls a ‘trust based conception of the speech act of testimony’ (Fricker 2012: 259). The conception is ‘trust based’ because she thinks that it is in the nature of the speech act of testimony that it aims at establishing relations of epistemic trust between speaker and audience, so that, all going well, the audience   Other assurance theorists include Ross (1986), Moran (2005), Hinchman (2005), and McMyler (2011). 4

42

L. Townsend

will end up trusting the speaker for the truth of what she says. When that happens we have what Fricker calls a deal of second personal epistemic trust. I will explain just what Fricker thinks is involved in such a deal of trust momentarily. But I first want to draw attention to the fact that trust is implicated in Fricker’s account of testimony in two subtly different ways, one ‘illocutionary’ and the other ‘perlocutionary’. Roughly put, the illocutionary-perlocutionary distinction marks a difference between what may be done in uttering something and what may be done by that utterance (Austin 1962: 107–8). So if I perform a locution with the illocutionary force of, for example, an invitation, and you (my audience) see that this is what I am doing with my words, then something already gets done, you get invited, just in virtue of my performance, and your recognition of it as being a performance of that kind. Of course, if I do invite you then there will typically be some further consequences—that is, something beyond your recognition of what illocutionary thing I am up to with my utterance—that my words strive towards. I may have invited you with the intention that you would accept my invitation, and come to my dinner party. That is the perlocutionary dimension of the utterance. So whereas it is up to the speaker herself what illocutionary status her speech has, if any, it is not up to her what perlocutionary effects it accomplishes. This is why it is the hallmark of illocutionary speech acts that the speaker can intelligibly lay claim to a certain authority in the performance of them, by prefacing what she says with ‘I hereby’, as in ‘I hereby invite you’, or ‘I hereby warn you’, and so on. With this distinction in mind, consider this claim Fricker makes about the speech act of testimony and its relation to trust: the speech act [of testimony] centrally involves the speaker making a commitment of trustworthiness to the addressee (worthiness, that is, of his second-personal trust) with respect to p. (Fricker 2012: 270)

I think this must be read as a claim about the illocutionary significance of the speech act of testimony. In other words, a commitment of trustworthiness being made is what is done simply in a speaker performing the speech act of testimony, by telling her audience that p. This seems to fit the picture of illocution alluded to above, because as speakers we have the authority to commit ourselves in various ways through our speech.5 For example, someone might say ‘I hereby commit to reducing my carbon footprint’ and therein become actually so committed. But this claim about the illocutionary significance of testimony—that it involves a speaker’s commitment to second personal epistemic trustworthiness—does not provide the whole sense in which the testimony is ‘trust based’, according to Fricker. This is because the speaker’s commitment to trustworthiness aims at establishing something further, namely, putting in place the ‘deal of trust’ between speaker and audience:

 A number of theorists have suggested that some notion of commitment is essential to both illocutionary acts in general (see, e.g., Moran 2013; Geurts m.s), and to assertion in particular (see, e.g., Peirce 1934; Brandom 1983; MacFarlane 2011). 5

3  Trust and Commitment in Collective Testimony

43

In offering her word in response to [her audience’s] epistemic need, the [testifier] supplies her side to an interpersonal deal of trust. (Fricker 2012: 258)

Here I think we must understand Fricker as referring to the perlocutionary dimension of the speech act of testimony. This is because a deal of trust, as I will explain in a moment, requires a substantial contribution from both the parties, and this means that forming such a deal is not something that is up to the speaker in the way the illocutionary significance of her speech is up to her. While inviting trust may fall within in the realm of the illocutionary, establishing trust does not (one could hardly say, ‘I hereby get you to trust me’). So although Fricker glosses testimony as ‘an illocutionary speech act’ (Fricker 2012: 253) and characterises her account as a ‘trust based conception of the speech act of testimony (Fricker 2012: 259), it seems to me that trust enters the picture in both an illocutionary way and a perlocutionary way. And, in fact, in view of the explanatory project on offer—the primary ambition of which is to explain the directed normative statuses and susceptibility to reactive attitudes of both speaker and hearer—it seems as though priority should be accorded to the perlocutionary dimension. Something connected with trust is done in the speech act itself (a commitment to epistemic trustworthiness is forged) but we only begin to understand why someone would do that thing when we look beyond this, to the paradigmatic perlocutionary point of testimony, which, for Fricker, is the establishment of a deal of trust between the speaker and her audience. This is because the commitment a speaker makes in testifying is put forward as a kind of enticement for the audience, so that she will decide to invest her trust, and hence seal the deal of trust. This gives an idea of how Fricker relates trust to the speech act of testimony; now we can look more closely at what she thinks the relevant sort of trust involves. The primary sort of trust she thinks is relevant to testimony is second personal epistemic trust. Fricker follows a number of recent theorists of trust in construing interpersonal trust as a three-place relation.6 More specifically, it is a special kind of reliance of one party (the truster) on another party (the trustee) for something (the object of trust). What distinguishes ‘second personal’ trust from other forms of trust is that it emerges out of a particular kind of communicative transaction between the parties, through which a deal of trust is struck, and this transaction produces certain interpersonally directed obligations and entitlements. According to Fricker, a deal of trust has a sort of dynamic, reflexive structure. The truster makes it known to the trustee that he is prepared to trust her in some specific matter, and by doing this actually provides the trustee with a reason to prove trustworthy—the reason being that he is trusting her. For her part, the trustee makes it known to the truster that she will honour his trust by proving trustworthy in the matter in which she is trusted, and that she will do so, at least in part, for the very reason provided by the truster (that he is trusting her). This commitment to trustworthiness made by the trustee then serves to support the truster’s original investment  See, e.g., Baier (1986), Holton (1994), and Pettit (1995).

6

44

L. Townsend

of trust, because it is reasonable to trust those who are committed to trustworthiness. This need not always be the precise sequence of events when a deal of trust is struck; the point is just that there must be, on one side, an investment of trust, and on the other side, a commitment to trustworthiness, and the parties need to communicate their respective contributions to one another, in order to reinforce the trust relation. A typical occasion for trust of this sort would be someone’s promising to do something to help out her friend, and having her friend accept the promise. So Sally might say to Anne, ‘Go ahead, drink all you like, I will give you a ride home’, and Anne might reply, ‘Great, thanks!’ and in this way they strike a deal of trust between them. Deals of trust like this qualify as ‘second personal’, for Fricker, in part because of the way the parties provide one another with reasons to participate in the deal, and in part because of the distinctively directed normative relations they effect. When Sally makes her promise of a ride, and the promise is understood and accepted by Anne, Sally thereby incurs a directed obligation to Anne to give her a ride home. In other words, she has not only placed herself under an obligation to give someone, Anne, a ride home, but now owes it to Anne to give her a ride home, in the sense that Anne is entitled to demand of Sally that she fulfils the obligation. This is why, should Sally fail to deliver Anne the promised lift she will not simply have gone wrong, but will instead have wronged someone—Anne herself. Put differently, if Sally fails to fulfil her promise made to Anne, then Anne becomes the victim, not merely part of the occasion, of Sally’s failure.7 And this is because of the deal of trust Sally and Anne had struck, a deal through which Sally willingly incurred certain directed obligations to Anne. Fricker thinks that testimony, like promising, has its home in deals of second personal trust, with the primary difference between promising and testimony residing in the object of trust—what the truster trusts the trustee for. Promising typically involves some sort of practical undertaking: the promiser commits to the promisee to doing something, such as giving her a ride home. Testimony, by contrast, involves an epistemic object: in testifying that p, the testifier ‘accepts responsibility for her inquirer’s belief as to whether p’ (Fricker 2012: 258). And this means that testimony, like promising, makes its recipient susceptible to a ‘personal betrayal’ (Fricker 2012: 259), and exposes the speaker to ‘the implicit threat of a powerful kind of blame’ (Fricker 2012: 269). So here, in summary, is Fricker’s trust based view of the speech act of testimony. The basic strategy is to use trust to explain those second personal normative features of testimony that, as I noted above, proponents of the assurance view are especially keen to account for. It is important to note that trust features both in terms of the illocutionary dimension of the act and in terms of the perlocutionary effects at which it typically aims. The illocutionary significance of testimony is that it commits the speaker to second personal epistemic trustworthiness, and the

 Cf. Thompson (2004: 340).

7

3  Trust and Commitment in Collective Testimony

45

perlocutionary point of such commitment is the establishment of a deal of second personal epistemic trust between speaker and hearer.

3.2.2  H  ow Fricker Extends the Account to Collective Testimony Having clarified her trust based account of testimony, Fricker explores whether it can be extended to cases where the testimonial source or ‘speaker’ is some kind of collective body, such as when a team of researchers co-author a report, or when a statement is put forward in the name of a group by an authorised spokesperson. She asks: Can a group be so constituted that it is capable of standing in the distinctively second-­ personal relations of trust that are proper to testimony? (Fricker, 2012, p. 269)

Fricker thinks that groups are capable of this, and to support her view she appropriates Margaret Gilbert’s concept of ‘joint commitment’.8 A joint commitment, as it is defined by Gilbert, is a commitment of and by several people to doing something as a single body, where possible substitutions for ‘doing something’ include both actions, such as walking together, and attitudes, such as believing that p, or intending to phi.9 So Fricker’s idea is simply that the characteristic object of a testimonial commitment, namely, proving worthy of an audience’s second personal epistemic trust, is the possible object of a Gilbertian joint commitment. Just as an individual testifier makes a personal commitment of epistemic trustworthiness in her testimony, so too can a collective speaker—a ‘plural subject’, as Gilbert would call it—make or express a joint commitment of epistemic trustworthiness in its speech.10 So a statement put forward in the name of a group will count as collective testimony just if it somehow involves the group members’ joint commitment to second personal epistemic trustworthiness in the matter it concerns. And when such a commitment is met with an investment of epistemic trust on the part of the group’s audience, there will be a deal of second-personal epistemic trust on one side of which is a group: there will be a We-thou deal of epistemic trust. Fricker thinks that Gilbert’s concept of joint commitment is especially well-­ suited to the task of underpinning her account of collective testimony because of

 See, esp. Gilbert (2006, 2014).  See Gilbert (2002: 41): ‘Joint commitments are always commitments to “act as a body” in a specified way, where “acting” is taken in a broad sense. Thus people may jointly commit to deciding as a body, to accepting a certain goal as a body, to intending as a body, to a believing as a body a certain proposition, and so on’. 10  I say here ‘make or express’ –as I discuss later, Fricker is unclear on whether the joint commitment involved in group testimony is a new joint commitment formed in the act of speaking or whether, in speaking, the group expresses a previously formed joint commitment that is in some sense constitutive of the group. 8 9

46

L. Townsend

what she sees as a sort of normative affinity between the Gilbertian framework and her assurance view of testimony. Specifically, just as the normative features of testimony have a particular kind of interpersonal or ‘second personal’ character, so too do the normative features of joint commitment.11 According to Gilbert, when a set of people are jointly committed in some way then each one of them incurs a distinctive participatory obligation: an obligation to play her part in their effort to fulfil the joint commitment. But these member-level obligations are actually directed obligations: each party owes her conformity to the joint commitment to all the other parties to the commitment.12 And each one of them also has, in virtue of their joint commitment, a directed entitlement: the entitlement to demand that the others comply, and the standing to rebuke any non-compliers for their failure to comply. This is why Gilbert claims (and Fricker approvingly quotes this claim) that ‘joint commitment is the clearest possible context for interpersonal betrayal’ (Gilbert 2006: 150). To illustrate just how it is that a joint commitment to epistemic trustworthiness is meant to underpin collective testimony, Fricker provides an instructive example. She asks us to imagine a government-appointed committee that is given the task of producing a report on the health risks of certain food additives—a report which will then be published so that the public can learn of the findings. Such a group would have to be constituted, Fricker thinks, by way of a joint commitment to epistemic trustworthiness with respect to what they will say about those health risks in their report (such a commitment ‘comes with the job’, Fricker (2012: 272) claims), and this joint commitment somehow forms the basis for the (implicit) invitation to trust that the report communicates to the public. This means that when a member of the public reads and believes the report, a deal of trust is sealed between the committee and this citizen—the citizen entrusts his views vis-a-vis those health risks to the committee, and the committee undertakes responsibility for his, and other citizens’, beliefs. And so, when it later transpires that certain members of the committee had been swayed by bribes from certain food companies, this amounts to a ‘personal betrayal’ of the citizen by the committee: Our citizen is betrayed. He trusted the committee’s word, but they abused his trust, and doubtless that of many other members of the public. (Fricker, 2012, p. 273)

 Fricker is rather vague about the suitability of the Gilbertian framework for her purposes: she says her approach to testimony and Gilbert’s plural subject theory ‘involve the same materials’ and hence that (in appropriating Gilbert’s notion of joint commitment ‘we are clearly in the right register’ (Fricker 2012: 271). 12  According to Gilbert, ‘obligations of joint commitment are, in the parlance of deontic logic, “directed obligations”. A logical correlate of my obligation to you to do such-and-such is your right against me that I do such-and-such’ (Gilbert 2002: 50). 11

3  Trust and Commitment in Collective Testimony

47

3.3  Two Concerns with Fricker’s Account I now want to raise two concerns for Fricker’s trust based account of the speech act of testimony, and in particular for the way she construes the commitment a speaker makes in testifying. These concerns relate to her general account of testimony, not specifically to her proposal for extending this account to groups. But the way she attempts to extend the proposal to groups, by appropriating Gilbert’s concept of joint commitment, is especially revealing of (what I take to be) the problems in her account.

3.3.1  The Object-Related Concern The first problem concerns the object of the speaker’s commitment, that is, what it is a commitment to. According to Fricker, in testifying a speaker makes a commitment with the following object: second personal epistemic trustworthiness. Notice that although this object makes reference to something epistemic—epistemic trust—it is itself a practical commitment, in the sense that fulfilling it requires the committed party to do something, namely, live up to, or honour her audience’s trust, should it be invested. And, as it turns out, Fricker must construe the object of the commitment in this broadly practical manner if her attempt to appropriate Gilbert’s notion of joint commitment is to be feasible. This is because, for Gilbert, all joint commitments are practical in precisely that sense: they are commitments of and by a set of people to ‘doing something’ or ‘acting’ as a single body (Gilbert 2002: 41). Accordingly, the responsibility that is incurred by a testifier, when her commitment to trustworthiness is met by an investment of trust on the part of her audience (when their deal of trust is sealed), is also at base a practical form of responsibility. Fricker characterises this as a matter of ‘responsibility taken for the other person’s epistemic status’ (Fricker 2012: 258) and what is meant by ‘epistemic status’ here is the audience having a particular belief on the matter at hand, and that belief being true and warranted. So the deal of trust involves the audience entrusting her beliefs vis-a-vis p to the speaker, who thereby takes responsibility for having furnishing the audience with certain beliefs. This is why, when things go wrong, and the audience is left with a false or unwarranted belief, notions like betrayal and blame are directly brought into play. If you commit to being trustworthy in this sense—if I entrust my beliefs to you, and you commit to honouring that trust—then the way I will hold you responsible for your failure to fulfil that commitment is by blaming you when things go wrong. But this, it seems to me, is not what we mean when we talk of epistemic or doxastic responsibility. To be responsible for belief—whether one’s own belief, or, in the case of testimony, someone else’s—is not a matter of answering for the influence one might have exerted in the production of certain psychological states. If I am responsible for the belief that p (whether mine or yours), this is not, except

48

L. Townsend

derivatively, a responsibility in respect of someone’s having the attitude of believing p, but is instead a responsibility I have in respect of p’s truth. To see this, consider our practices of challenging and answering challenges—of holding, and being held, responsible—in respect of our own beliefs. If you avow your belief that p to me, you make yourself liable to a certain kind of challenge: I may now ask you, ‘Why do you believe that p?’ It is a familiar point that this sort of challenge does not ask for a causal history of your state of mind, but rather for your reasons for believing—for considerations that will rationalise your belief and so vindicate your entitlement to it. But it is worth noting that you will not count as accepting my challenge unless you answer with the right kind of reasons for believing. That is, it would not do for you to respond with purely ‘attitude-related’ reasons for belief—reasons that attempt to vindicate your believing that p with reference to the attractiveness or worthwhile-ness of holding the attitude of believing that p. Instead, what you need to respond with, if you are to really take responsibility for your belief, is ‘content-related’ reasons: considerations that bear on the truth of the content of the belief, that is, p itself. Something similar goes for the responsibility that testifiers undertake for the beliefs of the recipients of their testimony. If you tell me that p, and I believe you, then you are in some sense responsible for my belief. You might also bear some responsibility, as Goldberg (2006) and McMyler (2011) have suggested, for what I go on to tell others, and hence what the recipients of my testimony believe. This is supported by so-called epistemic buck-passing practices—which are governed by what McMyler (2011: 62) calls the ‘epistemic right of deferral’, meaning that the recipient of testimony is entitled to defer challenges to his testimonially-acquired beliefs back to the original speaker. But this responsibility is strictly ‘content-­ related’—it is a responsibility in respect of the content of what you told me, and hence what I believed and told others. That this is so is evident from our discursive practice. If what you told me is faulty, or is placed in doubt by one of my interlocutors, then my first recourse will not be (as Fricker appears to have it) to charge you with a ‘personal betrayal’ or to target you with ‘a powerful form of blame’. Instead, I will direct an epistemic challenge to you: ‘Why did you tell me that p?’ Here again I am not asking why you saw fit to induce a possibly faulty belief in me; instead, what I am demanding is reasons that bear on p’s truth. To be responsible for a belief that p, whether yours or someone else’s, is first and foremost to be liable to answer these kinds of challenges, challenges to p’s truth.13  Note that in her discussion of the epistemology of testimony, Fricker does sometimes speak of ‘epistemic responsibility’ as necessary background condition on testimonial warrant. For example, she claims that ‘provided there is sufficient background reason for the addressee to assume the testifier has exercised epistemic responsibility and is speaking sincerely, then the testifier’s (ethical-epistemic) reason to be trustworthy regarding p generates a correlative (purely epistemic) reason for the addressee to take her at her word’ (Fricker 2012: 262, italics in original). Epistemic responsibility in this sense is something inquirers and speakers (should) exercise: it is matter of being diligent, careful, rigorous (etc.) in the conclusions one draws and the pronouncements one makes. By contrast, the sort of epistemic responsibility I am suggesting is missing from Fricker’s account is responsibility in the sense of ‘answerability’ or ‘accountability’ (cf. Watson 2004)— 13

3  Trust and Commitment in Collective Testimony

49

So my first objection to Fricker’s account is that by making the object of the speaker’s commitment practical—a commitment to proving worthy of the audience’s second personal trust—she obscures the fundamental form of responsibility relevant to testimony. Testifiers are answerable to challenges to the truth of what they say, and not, except derivatively, for the role they played in inducing a state of mind in their audience.

3.3.2  The Subject-Related Concern The second problem I wish to raise for Fricker’s account relates to the subject(s)— or, perhaps better, the intersubjectivity—of the testifier’s commitment, by which I mean who is committed to whom. The problem is made apparent, I think, by Fricker’s choice to appropriate Gilbert’s notion of joint commitment in service of her account. According to Fricker, a group that is constituted (at least in part) through joint commitment to second personal epistemic trustworthiness would therein be well-suited to giving testimony, because it is precisely this sort of commitment that makes the group responsible to its audience in the way characteristic of testimony. Now I have already complained that the nature of this responsibility is mischaracterised by Fricker, who sees it as a practical sort of responsibility that makes the speaker liable to betrayal and blame, instead of (what would be more appropriate) a sort of epistemic answerability that makes the speaker liable to epistemic challenges and rational criticism. In this section, however, I want to set that worry aside and ask whether the nature of the commitment, as Fricker construes it, gives the hearer the standing to hold the group responsible in any way at all. At this point it is worth explicitly noting something that is apt to be quite confusing in Fricker’s account of group testimony, concerning the precise role of the joint commitment to second personal epistemic trustworthiness. This is that, on Fricker’s account of group testimony, this joint commitment does not play the parallel role as that played, in her account of individual testimony, by the speaker’s commitment to second personal epistemic trustworthiness. Rather, the joint commitment to second personal epistemic trustworthiness is conceived by Fricker as constitutive of the group itself (it is the ‘making of the collective good informant’), and so is established prior to the speech act of group testimony. By contrast, as we saw earlier, the speaker’s commitment to second personal epistemic trustworthiness in individual testimony is something that is only established in the illocutionary act of telling itself. So the joint commitment to epistemic trustworthiness invoked by Fricker as constitutive of a ‘collective good informant’ is by no means a collective-level counterpart of the commitment an individual speaker makes in testifying.

that is, of being liable to certain challenges or criticisms. Thanks to Preston Stovall for encouraging me to address this.

50

L. Townsend

Nonetheless, Fricker is clear that group testimony, like individual testimony, also aims at establishing trust, and hence that, when groups give testimony, they, like individual testifiers, therein commit to second personal trustworthiness. But now we might wonder what the relation is between the original (constitutive) joint commitment—the one that, for example, comes with the job of being a member of the food-­ additives committee – and this second, illocutionary commitment—the kind that is made in the act of publishing the report. Are these two separate commitments (with the very same object), or is the communicative commitment some kind of expression or signalling of the prior, group-constituting commitment? Unfortunately Fricker does not clarify this important issue. Still, I think there are good reasons to favour the second option, that the invitation to trust extended by the group speech act (e.g., the publication of the report) serves to signal or express the already-present joint commitment to second personal epistemic trustworthiness. For one thing it is not clear what explanatory purpose is served by the postulation of a group-constituting joint commitment to epistemic trustworthiness when such a commitment could be formed in the act of group testimony itself. Moreover, Fricker suggests that in cases of group testimony the distinctive normative implications of testimony—such as the audience’s susceptibility to betrayal—are grounded in the (original) joint commitment: where a testifier fails to be trustworthy, she reneges on a commitment—her side of the deal of trust—so that the addressee is thereby betrayed. Now the kind of commitment at stake in the formation of a plural subject is just the right kind of commitment to generate the possibility of such a betrayal. (Fricker 2012: 271)

This is where I think the second problem with Fricker’s account lies. I think that a joint commitment, even one to second-personal epistemic trustworthiness, is not at all the right kind of commitment to generate the possibility of testimonial betrayal. In the case of the food-additives committee, Fricker appears to hold that the betrayal of the citizen by the committee is somehow grounded in, or otherwise connected with, the failure of the committee to honour their constitutive joint commitment to second personal epistemic trustworthiness in respect of their report. However, given the normative structure of joint commitment, it is hard to see how this failure of the committee—their ‘reneging’ on their joint commitment—could in fact amount to a betrayal of the citizen. To see this, note that there are, in general, two different kinds of ‘failure’ made possible by the existence of a joint commitment. The first kind of failure arises from its jointness. This is the way in which the various parties to a joint commitment will be failing one another if they do not fulfil their obligations of conformity to the joint commitment. This is the sort of failing that Gilbert has in mind when she claims that joint commitment can give rise to interpersonal betrayal. But it should be clear that, on its own, this is not going to be of any help in explaining how someone outside of the joint commitment could get betrayed. If the members of the food-additives committee are jointly committed to proving worthy of the public’s trust, then each one of them owes it to all of the others—the other members of the committee, not the public—to play her part in their actually

3  Trust and Commitment in Collective Testimony

51

proving to be trustworthy. Hence if certain members are corrupted and so do not fulfil their directed participatory obligations, and as a result the group ends up failing to prove trustworthy, then there is, to be sure, an interpersonal betrayal that gets perpetrated. But it is not one that involves the audience of the committee’s report, because it all takes place within the committee. So while the joint commitment to epistemic trustworthiness can explain the possibility of a certain sort of internal betrayal within the less-than-trustworthy testifying group—a betrayal of some members by others—it doesn’t help at all to explain the betrayal of the citizen by the group.14 The second kind of failure made possible by the existence of a joint commitment arises not from its jointness but from the fact that it is a commitment at all. This is a more basic sort of rational failure: the failure of the subject of a commitment (in the case of a joint commitment, the plural subject) to fulfil the object of their commitment. This possibility arises, according to Gilbert, because of what commitments are, namely ‘facts about one’s normative situation—about what one has reason to do’ (Gilbert 2013: 899, italics in original). So, for example, if I commit myself to reducing my carbon footprint but fail to do so (without having changed my mind) then I fail to act as reason demands. Here too, however, it is not obvious how this sort of failure could give rise to the citizen’s betrayal in Fricker’s example. The members of the committee were jointly committed to epistemic trustworthiness with respect to their report about food additives, and in light of the corruption within the committee they certainly failed to fulfil that commitment. But this sort of failure is not inherently interpersonal: it is not a matter of the committee failing the citizen, only their failure to fulfil their commitment. It doesn’t help that the object of the commitment makes reference to the interpersonal matter of proving worthy of the public’s trust, because that alone cannot entitle the (members of) the public to the fulfilment of the obligation. By analogy, if I make a personal commitment to tip every busking musician I come across, these busking musicians do not thereby acquire a right to my tips, or the standing to demand that I tip them, or the entitlement to feelings of betrayal if I do not. Such a commitment, though it involves other people in its content, is not inherently interpersonal because it is not a commitment I made to those people. It is a personal commitment, and, as such, its fulfilment is strictly my business. The same goes, I think, for the commitment to epistemic trustworthiness that helps to constitute the food-additives committee. This is a commitment of and by the committee, and though it makes reference to the citizenry who will read the report, it is not a commitment made to them. This means that the fact that the committee fails to prove

 To be clear, I am not attributing this claim to Fricker. In fact Fricker seems well aware that joint commitment could not ground the citizen’s susceptibility to betrayal in this manner, because the ‘intersubjective commitments that make up joint commitment […] are not […] commitments on the part of the group to another party’ (Fricker 2012: 271). But what work does joint commitment do, in that case? My point here is simply that joint commitment cannot help to ground the normative dimension of testimony in this or any other manner. Thanks to Niels de Haan for suggesting that I make this clearer. 14

52

L. Townsend

trustworthy does not ground the citizen’s betrayal, nor does it give the members of the public any right of complaint or license any reactive attitudes on their part. Specifically, it doesn’t give them any such right or license any such attitudes just because they (members of the public) themselves feature in the object of the committee’s commitment, which is a commitment to prove worthy of their (the public’s) trust. It seems, then, that a Gilbertian joint commitment to second personal epistemic trustworthiness does not, after all, help to explain the distinctive intersubjective normativity of group testimony  – the way in which testimony confers certain directed rights and entitlements on the audience, and makes the audience susceptible to a ‘personal betrayal’. Insofar as Fricker thinks this kind of commitment can ground the intersubjectivity of group testimony, I think this betrays a broader problem with her account of testimony in general. To put it bluntly, it looks like, for Fricker, testimony is a way of forming and signalling a commitment to proving worthy of someone’s second personal trust, should it be invested. (In the case of individual testimony the forming and the signalling both happen in the act of telling, while in group testimony the forming seems to pre-date the signalling.) But to do that is not to commit to another person, only to commit oneself ‘at’ them. The relevant difference is, roughly speaking, the same as that between announcing one’s intention to help someone and promising them that one will. In the former case one might well produce expectations in the audience, but one does not license or entitle their reliance on your help, nor would they have been betrayed should you fail to live up to your intention. In the same way, I think, someone to whom a commitment to second personal epistemic trustworthiness is signalled does not therein acquire the right and entitlements properly associated with testimony.

3.4  An Alternative Form of Second Personal Commitment I have suggested that Fricker’s account of collective testimony fails because of the way she construes the commitment a speaker makes in testifying, and that this is made especially clear by her choice to appropriate Gilbert’s notion of joint commitment in service of her account. Signalling one’s commitment to second personal epistemic trustworthiness does not, I argued, make the speaker responsible for the truth of what she says (the object-related worry), nor does it make the speaker responsible to her audience (the subject-related worry) in the right sort of way. Nevertheless, I think that Fricker’s idea that the speaker makes a certain kind of ‘second personal commitment’ in testifying can be preserved, if we make suitable adjustments to the way that commitment is construed. This is what I attempt to do in this section. Let us begin with the object of the testifier’s commitment—what she, in testifying, commits herself to. Instead of committing herself to proving worthy of her audience’s trust, as Fricker has it, I propose that the testifier commits herself to the truth of what she says.

3  Trust and Commitment in Collective Testimony

53

There is, of course, nothing remotely original about this proposal. It is very standard to classify the speech act of telling as a member of the broader illocutionary category of ‘assertives’, and one prominent way of understanding this broader category is with reference to the idea of commitment to the truth of what is said.15 According to Searle’s influential taxonomy, for example, the [illocutionary] point or purpose of the members of the assertive class is to commit the speaker (in varying degrees) to something’s being the case, to the truth of the expressed proposition. (Searle 1979: 12)

But what is it, exactly, to commit oneself to the truth of a proposition? To be so committed is to have bound oneself to something’s being actually the case, in some kind of normatively thick sense of ‘bound’. The normativity of the commitment characteristic of assertion is frequently cashed out in deontological terms: the speaker willingly undertakes responsibility for the truth of what she asserts; she incurs certain obligations relating to the propositional content of her utterance. It is not, of course, that by committing herself to the truth of p that she imposes a duty on herself to make it the case that p. Instead, to commit oneself to the truth is to make oneself liable to—i.e., responsible for answering—certain distinctively epistemic challenges: In producing assertions, [speakers] become responsible in the sense of answerable for their claims. That is, they [undertake] a specific task responsibility, namely the responsibility to show that they are entitled to the commitment expressed by their assertions, should that entitlement be brought into question. (Brandom 1994: 173)

This helps to remedy the first problem I raised for Fricker’s construal of the commitment a speaker makes in testifying. Construing the speaker’s commitment as a commitment to proving trustworthy forces us, I argued, to assimilate the speaker’s responsibility into a practical or ethical mould, and this has the unwelcome effect of obscuring the distinctively epistemic form of responsibility that is proper to testimony. For Fricker, when a piece of testimony is accepted, there is straightaway a prospect of ethical let-down in the offing: the addressee becomes vulnerable to a ‘personal betrayal’ (Fricker 2012: 259), and the speaker is exposed to ‘the implicit threat of a powerful kind of blame’ (Fricker 2012: 269). In the context of ordinary testimony, of one person telling another something, such as what the time is, or what they had for lunch, such talk may seem somewhat overblown. But, more significantly, it actually obscures the fundamental normative import of testimony, which is epistemic in character. When you tell me that p I acquire the standing to challenge you with respect to p, and to defer certain challenges I myself may be subjected to with respect to my believing p, back to you. Correlatively, you, in telling me that p, make yourself answerable to my challenges and the challenges of others that I defer back to you; you must, in responding to these challenges, vindicate your own entitlement to the commitment expressed, or retract your assertion. It is this kind of

15

 See MacFarlane (2011) for a discussion of this tradition of understanding assertion.

54

L. Townsend

epistemic answerability, rather than the practical responsibility for influencing another’s beliefs, that is most directly relevant to testimony. So adjusting the object of the speaker’s commitment—construing it as her commitment to p’s being true, rather than to her proving worthy of her audience’s trust—allows us to better capture the distinctive epistemic form of answerability that arises from testimony. But it might seem that this attractive consequence comes at too high a cost, because it does away with the central interpersonal element in Fricker’s account. For Fricker wants to do justice to the ‘profoundly intersubjective’ (Fricker 2012: 258) character of testimony, and this is why she makes the object of the speaker’s commitment refer to something second personal (namely, second personal epistemic trustworthiness). My proposal thus seems to strip her account of the resources it needs to capture the interpersonal character of testimony. This brings us to the second part of my proposal, which relates to the intersubjectivity of the speaker’s commitment, i.e., who is committed to whom. I suggested that, despite the fact that the object of the speaker’s commitment on Fricker’s account makes reference to something second personal, it nevertheless fails to alter the normative situation of the addressee. Specifically, the addressee does not acquire a right to, or a standing to demand, the fulfilment of the commitment by the speaker. So there is a sense in which, although the commitment is other-involving and other-­ addressed, it is nonetheless a kind of monadic or first-personal commitment. It is a commitment of and by the speaker (who may, in the case of collective testimony, be a plural subject), and hence the forging of such a commitment does not itself give outside parties (‘others’) any rights to the fulfilment of the commitment. In this connection I propose a further adjustment to Fricker’s construal of the speaker’s commitment. We need to construe the nature of the commitment the speaker makes in testifying not in Fricker’s first-personal terms but in second personal terms. That is, we need the commitment to be of a sort made by the speaker towards her addressee, where the sense of ‘towards’ is not simply a matter of the addressee’s being implicated in the content of the commitment. Here too I should be clear that my proposal is not a novel one. Indeed, one of the authors whose work Fricker explicitly aligns herself with, Richard Moran, has recently argued that the speech act of testimony should be understood in something like the way I am recommending. But, in contrast to Fricker, Moran sees the intersubjective character of the speaker’s commitment arising not out of the object of the commitment but out of the very fact that testimony is an illocutionary speech act. According to him, to tell someone that p is to commit to p’s being true by way of a ‘relational act’ (Moran 2013: 123), one that is ‘second personal in its very nature’ (Moran 2013: 121). Hence although the responsibility associated with this commitment is epistemic—because it is a commitment to p’s being true—it is also a responsibility the speaker has specifically in respect of the addressee of her testimony. She is answerable to her audience for challenges to the truth of p. Such relational responsibility and commitment is the defining characteristic, according to Moran, of illocutionary acts:

3  Trust and Commitment in Collective Testimony

55

in illocutions generally the speaker makes himself responsible in particular ways to another person. It is in the nature of an illocutionary act to be undertaken toward another person, to be an act performed with regard to, or to, another person, as in ‘I told him the news’, ‘I asked him to leave’, ‘I warned her about the car’. (Moran 2013: 127)

One might still wonder what exactly it means to say that acts of telling, or illocutionary speech acts in general, are essentially relational or second personal acts. Moran gestures at how this might be spelled out by alluding to the idea that illocutionary speech acts aim at producing a peculiar, self-referential sort of ‘effect’ in the audience: they aim to be recognised as the kind of acts they (thereby) are. That is, it is the hallmark of illocutionary acts that they are successful as soon as they are understood, by the audience, as they are meant to be understood, by the speaker. My effort to invite or warn or tell you is achieved—you actually are invited, warned or told—as soon as you recognise what I am up to with my words, namely performing an illocutionary act of that sort. As Jennifer Hornsby (in a paper cited by Moran) notes, the fact that recognition is, in this way, central to the notion of illocution shows how the communicative use of language depends on people being ‘sufficiently in harmony’ (Hornsby 1994: 193) or ‘attuned’ to one another to be able to identify, and so help to accomplish, what they are doing with their words. Hornsby calls this condition of harmony, upon which communication depends, ‘reciprocity’: A background of reciprocity is what provides for someone’s making of some noises constituting her performance of an illocutionary act. Only where there is reciprocity are the conditions in place for someone to say something to someone, or to warn someone of something, or whatever. Reciprocity is a feature of the human situation which allows for a meeting of minds. (Hornsby 2003: 301)

A fuller development of this way of characterising the ‘profoundly interpersonal’ nature of testimony is clearly in order. But even with only this bare sketch in hand we can see a contrast between this way of thinking and Fricker’s. For Fricker, certain acts of communication, such as promises or testimony, have the perlocutionary aim of establishing I-thou relations of trust between speaker and hearer. That is, they are contributions towards a prospective ‘deal of trust’ which would, if sealed, generate various directed obligations and entitlements for the parties to the deal. But the Moran-Hornsby point is that acts of this kind do not so much establish trust as depend on it already being in place. To speak, in the sense of illocution, is already to trust the one to whom one speaks to recognise what one is up to in speaking. So there is a sense in which the interpersonal trust relevant to testimony must be more basic than Fricker allows. Whereas she identifies trust as something that is established through communication, the Moran-Hornsby approach sees it as the condition upon which communication depends. My proposal for how the speaker’s commitment should be construed is thus as follows. In telling her audience that p, a speaker makes a commitment to p’s being true to her audience. The first ‘to’ in this formulation specifies the object of her commitment: what she makes herself answerable for. The second ‘to’ specifies the subject or subjective target of her commitment: whom she makes herself

56

L. Townsend

answerable to. Maintaining a distinction between the object and the subject of the speaker’s commitment allows us to understand how testimony manages to be both assertion-­like and promise-like. What makes testimony similar to assertion is that it has an epistemic object, the truth of the proposition expressed. But whereas assertions can be ‘offered up impersonally into public space’ (Kukla and Lance 2009: 166), testimony is always more personally directed—a matter of ‘I told him’, as Moran (2013: 127) puts it. But that is just to say that it is in the nature of testimony, but not of assertion, that the subject of the speaker’s commitment is a specific person. What one is responsible for is held constant in both assertion and testimony, but whom one is responsible to varies. With respect to whom one is responsible to, testimony is more akin to promising, because, in promising, one incurs a form of liability or answerability specifically and only in the respect of the promisee for doing what one has promised to do. Can groups commit themselves in this way, when they, for instance, jointly author a report, or have a spokesperson deliver a statement in the name of the group at a press conference? On my proposal this is not only a question about the capacities of groups—whether or not they ‘have what it takes’ to come out with testimony—but also a question about the bounds of reciprocity and recognition. No matter how well their internal modes of organisation allow them to determine and articulate matters of fact, groups will not succeed in telling anybody anything unless there is receptiveness on the part of their would-be audiences to what they are up to with their words. But it is clear, I think, from our practices of holding groups responsible for the truth of what they say, that groups are often recognised as performing the speech act of testimony. By way of illustration, consider a statement made by a spokesperson for the Obama Administration, that senior officials in the Russian government were behind the cyber attacks that influenced the 2016 US presidential election. This statement counts as collective testimony, in my view, because it commits the Administration itself—as opposed to simply committing the spokesperson (or the president) personally—to the truth of its content, to its audience. It is not that the Administration commits to proving worthy of the public’s trust in the very specific matter of whether senior officials in the Russian government were behind the cyber attacks. Rather, through the statement, the Administration makes a commitment to the public to the accuracy of the statement (to its actually being so that senior officials in the Russian government were behind the cyber attacks). This is evident from the fact that the primary sort of challenge that the statement makes the Administration vulnerable to is a challenge raised against the truth of what has been claimed, rather than a complaint that they have failed to prove trustworthy.

3  Trust and Commitment in Collective Testimony

57

3.5  Conclusion It is in this spirit, I suggest, that claims put forward by or in the name of a group are typically to be understood. That is, we understand these claims as having the same illocutionary status that they would have were they performed by or in the name of single person. This means that groups are, at least sometimes, credited with the same ‘illocutionary authority’ as competent individual speakers of a language are typically credited with—and it is this, rather than their claims to trustworthiness, which should be seen as underpinning collective testimony. For a group to constitute its utterance as an act of testimony is for it to be able to make, and to be recognised as making, an epistemic commitment to some specific audience. Insofar as trust is implicated in this way of understanding collective testimony it is not via the idea that we might entrust our beliefs to the word of group, but rather that we might recognise what a group is up to with its words, and so allow the group to have its say. Acknowledgements  Thanks to audiences in Lund and Vienna for helpful discussion, as well as an anonymous reviewer and, especially, Preston Stovall for generous written comments. My research is financially supported by the Austrian Research Council, FWF project ‘Inferentialism and Collective Intentionality’ (I 3068).

Bibliography Anscombe, G. E. M. (1979). What is it to believe someone? In C. F. Delaney (Ed.), Rationality and religious belief. Notre Dame: University of Notre Dame Press. Austin, J. L. (1962). How to do things with words. Oxford: Clarendon Press. Baier, A. (1986). Trust and antitrust. Ethics, 96(2), 231–260. Brandom, R. (1983). Asserting. Noûs, 17(4), 637–650. Brandom, R. (1994). Making it explicit. Cambridge, MA: Harvard University Press. Faulkner, P. (2011). Knowledge on trust. Oxford: Oxford University Press. Fricker, M. (2012). Group testimony? The making of a collective good informant. Philosophy and Phenomenological Research, 84, 249–276. Geurts, B. (2019). Communication as commitment sharing: speech acts, implicatures, common ground. Theoretical linguistics, 45(1–2), 1–30. Gilbert, M. (2002) Considerations on joint commitment. Chapter 2 in Gilbert, M. (2014). Joint commitment. New York: Oxford University Press. Gilbert, M. (2006). A theory of political obligation: Membership, commitment, and the bonds of society. Oxford: Oxford University Press. Gilbert, M. (2013). Commitment. In H. LaFollette (Ed.), The international encyclopedia of ethics (pp. 899–905). London: Blackwell. Goldberg, S. (2006). Testimony as evidence. Philosophica, 78, 29–52. Hieronymi, P. (2008). The reasons of trust. Australasian Journal of Philosophy, 86(2), 213–236. Hinchman, E. (2005). Telling as inviting to trust. Philosophy and Phenomenological Research, 70(3), 562–587. Holton, R. (1994). Deciding to trust, coming to believe. Australasian Journal of Philosophy, 72(1), 63–76.

58

L. Townsend

Hornsby, J. (1994). Illocution and its significance. In S. Tsohatzidis (Ed.), Foundations of speech-­ act theory: Philosophical and linguistic perspectives. London: Routledge. Hornsby, J. (2003). Free speech and hate speech: Language and rights. In R. Egidi, M. dell’Utri, & M. de Caro (Eds.), Normativit Fatti Valori. Macerata: Quodlibet. Kukla, R., & Lance, M. (2009). Yo!’ And ‘Lo! Cambridge, MA: Harvard University Press. MacFarlane, J. (2011). What is assertion? In J. Brown & H. Cappelen (Eds.), Assertion. Oxford: Oxford University Press. McMyler, B. (2011). Trust, testimony and authority. New York: Oxford University Press. Moran, R. (2005). Getting told and being believed. Philosophers Imprint, 5, 1–29. Moran, R. (2013). Testimony, illocution and the second person. Proceedings of the Aristotelian Society supplementary volume, 87, 115–135. Peirce, C. S. (1934). Judgment and assertion. In Collected papers of Charles Sanders Peirce (Vol. V, pp. 385–387). Boston: Harvard University Press. Pettit, P. (1995). The cunning of trust. Philosophy and Public Affairs, 24(3), 202–225. Ross, A. (1986). Why do we believe what we are told? Ratio, 1, 69–88. Searle, J. (1979). Expression and meaning. Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Thompson, M. (2004). What is it to wrong someone? A puzzle about justice. In R. J. Wallace, S. Scheffler, P. Pettit, & M. Smith (Eds.), Reason and value: Themes from the moral philosophy of Joseph Raz. Oxford: Clarendon Press. Watson, G. (2004). Agency and answerability: Selected essays. Oxford: Oxford University Press.

Chapter 4

Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation Ronald Loeffler

Abstract  The concept of implicit scorekeeping plays a central explanatory role in Robert Brandom’s theory of discursive rationality. This is the concept of a speaker’s attributing, unmediated by explicit higher-order thoughts, discursive commitments and entitlements of various types to his or her interlocutors in conversation. According to Brandom, the totality of such implicit scorekeeping efforts, distributed over historically situated discoursing individuals, literally institutes the norms of discursive reasoning governing their conversations. Yet Brandom offers no positive theory of implicit scorekeeping. I argue that in light of the broader literature on mindreading, this omission is problematic. My goal in this chapter is to develop a positive theory of implicit scorekeeping and to indicate how implicit scorekeeping, as modelled by this theory, promises to carry the explanatory burdens Brandom assigns to his merely negative conception. Key to my proposal is the idea that beliefs are propositional we-attitudes, that is, that a belief that p intrinsically involves, qua doxastic attitude (as opposed to qua contentful state), a certain interpersonal dimension. The central parts of this chapter motivate and develop this conception of belief. As it turns out, believers so conceived attribute beliefs, both as held and as to be held, to others unmediated by higher-order thoughts, that is, implicitly. Inserting an adjusted and expanded version of this conception of belief into Brandom’s theory then yields the sought positive conception of implicit scorekeeping, which, as I shall indicate, holds promise to carry the explanatory burdens Brandom assigns to his merely negative conception. Keywords  Normative attitudes · Belief · Cognitive we-attitudes · Inferential rationality · Mindreading

R. Loeffler (*) Department of Philosophy, Grand Valley State University, Allendale, MI, USA e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_4

59

60

R. Loeffler

4.1  Introduction A central theme in what might be called the interpretive tradition in post-WWII Anglo-American philosophy is the idea that the ability to reason ourselves constitutively depends upon our abilities to interpret others as reasoning creatures in the context of interpreting speech as we communicate with them. Interpreting (recognizing, treating) our discursive interlocutors as reasoning creatures – as creatures engaging in conceptually contentful, propositional thought, as drawing inferences and detecting incompatibilities between their thoughts, as seeking reasons and evidence for, or against, what is to be believed or intended, as having or lacking propositional knowledge, etc.  – in light of interpreting their speech and, conversely, interpreting their speech in light of interpreting them as reasoning creatures, not only is essential for communicating linguistically with them but also, more interestingly, constitutes us as rational beings ourselves. We could not reason ourselves, if we could not interpret others as reasoning in the context of interpreting their, and our, speech as we discourse with them. Donald Davidson’s work is, of course, the locus classicus of this interpretive tradition. More recently, Robert Brandom has advanced a distinctly Kantian and pragmatist approach to reason in this interpretive tradition. Following Kant, Brandom thinks of reasoning as wholly normative. Reason is nothing but a vast system of norms, and reasoning nothing but a subject’s effort to follow these norms. Propositions and concepts are aspects of this system of norms, in accordance with Kant’s dictum that concepts are rules (Kant 1998: A 106). Moreover, Brandom’s approach to reason is pragmatist in the sense that reason so conceived is explained in terms of a rational being’s ability to use language in specifically discursive social practices. Reasoning is fundamentally discursive, in the sense that Brandom identifies the original form of reasoning, in which all other forms are based, with an aspect of a speakers’ skillful use of language in certain simple forms of linguistic communication: autonomous discursive practices, or “language games one could play though one played no other” (Brandom 2008: 27). The ability to participate in discourse ultimately consists entirely in normative pragmatic know-how, that is, in practical skills to use language more or less appropriately under given discursive circumstances and to treat one’s interlocutors as contributing more or less appropriately to the exchange. In autonomous discourse, all such normative pragmatic know-how is implicit, in the sense that it does not presuppose or rely on any unexplained, prior abilities to reason conceptually. Specifically, the participants’ sensitivity to (what they take to be) the norms governing their discourse, which is an essential aspect of their normative pragmatic know-­ how, emphatically involves no conceptualizations of the purported norms  – no explicit, propositional knowledge or beliefs that such and such are the norms. Of course, we can have explicit beliefs and knowledge of norms. Yet Brandom, appealing to regress arguments about rule following from Wittgenstein and Sellars, argues that having such explicit knowledge is inessential for being able to follow norms (Brandom 1994: 21, 27–29) and that, to the contrary, having any explicit,

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

61

propositional knowledge of norms depends upon prior normative pragmatic knowhow to follow norms merely implicitly in practice. Reason and reasoning themselves emerge as features of the participants’ employment of such pragmatic know-how in, ultimately, autonomous discourse and are, in this sense, explained by it. An essential part of the normative pragmatic know-how employed in discourse is a participant’s interpersonal skill to treat her interlocutors as more or less competent participants, by tracking and evaluating each interlocutor’s previous conversational moves and by attributing to him, depending on the course of the conversation, constellations of commitments and entitlement to further discursive moves. Brandom calls such interpersonal tracking and evaluating scorekeeping. In autonomous discourse, the ability to keep score does yet again not involve any unexplained, prior abilities to use concepts, but is yet again merely a complex practical skill to track and evaluate the moves, commitments, and entitlements of one’s interlocutors implicitly in practice. Specifically, it does not involve any explicit, higher-order thoughts, about any interlocutor, that he has made such and such moves or that she is such and so committed and entitled. Of course, we can form such higher-order thoughts, but Brandom maintains that this ability, like any ability to use concepts, is once again ultimately grounded in prior, implicit pragmatic abilities in autonomous discourse  – including abilities to keep score unmediated by such higher-order thoughts. A participant’s implicit treatment of the performances exchanged in autonomous discourse as governed by certain norms, and her implicit treatment of her interlocutors as more or less faithfully following these norms – her scorekeeping – amounts to her interpreting these performances as significant or meaningful in certain ways and her interpreting her interlocutors as contributing in certain more or less appropriate ways to the exchange. A key thesis of Brandom, which places his Kantian pragmatism in the interpretationist tradition, is that such interpretive activity, distributed over discoursing participants, explains why their discourse is norm-­ governed in the first place, and in particular why it is governed by norms that deserve to be identified with norms of conceptual reasoning. The norms of discourse themselves are somehow instituted or created through the discoursing participants’ implicitly treating each other, and the performances exchanged, as governed by certain norms. Thus, since Brandom identifies the norms constituting reason, propositions, and concepts in broadly Kantian fashion with certain norms of discourse, and reasoning with an effort to follow these norms, the participants’ mutual interpretive activity somehow creates or institutes the norms of reason themselves and makes their efforts to follow these norms instances of genuinely rational activity. Brandom calls this explanatory doctrine phenomenalism about norms, alluding to the fact that, according to it, what the norms appear to be to the participants as they reason discursively creates or institutes the very norms governing their discursive reasoning themselves (Brandom 1994: 25, 627; for discussion cf. Loeffler 2005, 2018: 212–222). This is why Brandom insists that all of an autonomous discursive practitioner’s pragmatic know-how is, in the sense explained, merely implicit. The normative pragmatic know-how speakers bring to bear when they use language in autonomous discourse is supposed to explain the linguistic performances exchanged

62

R. Loeffler

as propositionally meaningful and some of the participant’s states and activities as rational and  conceptually contentful. If some of this normative pragmatic-know involved unexplained prior abilities to reason conceptually, the phenomenalist explanation would be circular and hence would fail. This chapter thematizes Brandom’s concept of implicit scorekeeping and its alleged norm-instituting role. Scorekeeping is a form of mindreading, that is, a form of tracking and evaluating another’s cognitive activity – to wit, his discursive moves, commitments, and entitlements. Yet as I shall elaborate, mainstream approaches to mindreading today widely agree that mindreading essentially comprises explicit higher-order thought about the other’s cognitive activity. Thus, from the viewpoint of mainstream approaches to mindreading, Brandom’s notion of implicit scorekeeping is illegitimate. Since this notion is pivotal for his pragmatist explanatory project to explain conceptual reasoning in terms of normative pragmatic know-how employed in autonomous discourse, including know-how to keep score implicitly, its illegitimacy would thus threaten to thwart this project. My goal is to outline a strategy for those sympathetic to this project to reply to this challenge. The next section introduces more relevant details of Brandom’s theory of autonomous discourse and scorekeeping. Section 4.3 elaborates the challenge from contemporary theories of mindreading. Section 4.4 motivates and introduces a conception of belief qua propositional we-attitude, according to which having any beliefs oneself comprises recognizing and evaluating others as believers too, unmediated by explicit, higher-order thoughts, that is, implicitly. Section 4.5 inserts an adjustment of this conception of belief into Brandom’s theory, thus yielding, apropos the challenge from mainstream approaches to mindreading, a positive and potentially attractive conception of implicit scorekeeping in autonomous discourse, which, moreover, seems suited to carry at least some of the explanatory burdens that Brandom places on his conception of implicit scorekeeping. Section 4.6 looks more closely at one of these burdens and indicates a sense in which implicit scorekeepers in the developed sense regard themselves, and each other, as bound by objective norms of reasoning. The theory outlined in Sects. 4.4, 4.5 and 4.6 will raise many issues regarding the natures of belief and propositional knowledge, the psychology of mindreading, the pragmatics of linguistic communication, and the natures of normativity, rationality, and our collectively intentional abilities. I cannot address most of these issues in this limited space. Still, I hope that this theory will appear in a sufficiently attractive light to be seen as meriting further exploration.

4.2  Acknowledging and Attributing Commitments Assertion is the central speech act in autonomous discourse, Brandom argues, and basic rational activity is inferential activity in the medium of assertions and declarative sentences. Such reasoning consists in forming intrapersonal normative attitudes of two basic types, and in altering these attitudes systematically as the conversation

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

63

progresses, in accordance with what one takes to be the norms governing the conversation. First, each participant acknowledges certain assertional commitments concerning certain declarative sentences of the fragment of language in use. Roughly, to acknowledge an assertional commitment to a sentence S is to be disposed to assert S flat out (paradigmatically, when asked whether S) and to regard oneself as obliged or required to being so disposed, given the norms governing discourse.1 Second, each participant acknowledges, for each sentence S of the (finite) fragment of the language in use, certain inferential commitments, relating S in the given context as premise or conclusion to certain other sentences S′, S″, etc. (the sentences the speaker takes to follow from S, or from which she takes S to follow from), as well as what might be called incompatibility commitments, relating S in the given context to yet further sentences S*, S**, etc. (the sentences the speaker takes to be incompatible with S). Acknowledged inferential and incompatibility commitments regarding S thus concern pairs of sentences, each pair including S itself, and they amount to treating S in the given context as incompatible with, or as inferentially related to, the other member of the pair. (Unless indicated otherwise, I shall for brevity’s sake use “inference”, “inferential commitment”, etc. in the following inclusively, to cover both inference and incompatibility and the corresponding commitments and activities.) Acknowledged assertional and inferential commitments are normative attitudes in the sense that the subject, in forming these attitudes, is sensitive to being bound by, and to follow what she takes to be, norms governing discourse, including norms of inference. However, as said, in autonomous discourse, all such sensitivity is implicit in that it never involves any conceptualizations of what one takes to be the norms – any knowledge or belief that such and such are the norms. Moreover, in autonomous discourse all inferences are material, in the sense that their validity (if so) does not depend on premises containing logical vocabulary, such as the conditional or negation. Correspondingly, in acknowledging inferential commitments, the speakers do not use unexplained logical concepts (Brandom 1994: Ch. 2).2 Given Brandom’s commitment to inferential role semantics, the propositional meaning of any declarative sentence S of the fragment of language in use consists in S’s inferential role, that is, in the norms of inference relating S to certain further sentences of the language. Correlatively, a speaker’s interpretation (linguistic understanding) of S consists in the inferential commitments she acknowledges regarding S, that is, in the inferential role she implicitly takes S to have. Moreover, to believe that p, according to Brandom, is to acknowledge an assertional commitment to a declarative sentence S, that is, to be disposed to assert S flat out, in accordance with what one implicitly takes to be the norms, while interpreting S as meaning that p,

 Each participant also acknowledges entitlements to certain assertions. For brevity’s sake, I shall focus only on commitments here, though I envision my story to extend straightforwardly to entitlements as well. 2  Many of these inferences are non-monotonic and their validity is, accordingly, often relative to collateral commitments. I shall abstract away from this complication. 1

64

R. Loeffler

that is, while acknowledging a set of inferential commitments regarding S to this effect. Original discursive reasoning thus consists in acknowledging successive constellations of assertional and inferential commitments in autonomous discourse, given what one takes to be the norms governing the conversation. This intrapersonal cognitive activity occurs in the context of the interpersonal scorekeeping activity. Each participant tracks, evaluates, and forms normative expectations about each interlocutor’s rational activity, more or less accurately and comprehensively, by attributing to him successive constellations of assertional and inferential commitments. Attributing commitments to one’s interlocutors is thus a second, interpersonal type of normative attitude characterizing autonomous discourse, alongside the intrapersonal normative attitude of acknowledging commitments oneself. Specifically, at each stage, the speaker attributes to each interlocutor two constellations of commitments: first, the constellation of assertional and inferential commitments the speaker takes the interlocutor to acknowledge in fact at that stage (the interlocutor’s de facto beliefs and interpretive activity, according to the speaker) and, second, the constellation of assertional and inferential commitments the speaker takes the interlocutor to have, that is, the commitments he is bound to acknowledge at that stage, according to the speaker (the beliefs and interpretations the interlocutor should hold, and make respectively, according to the speaker) – whether or not he does so in fact. Of course, these two constellations always overlap. Much of what the speaker takes to be the interlocutor’s de facto doxastic and inferential activity is also activity the speaker regards as correct. As said, a key aim of Brandom’s pragmatism is to show that the combination of these interlocking intrapersonal and interpersonal activities, distributed over the discoursing participants, somehow institutes or creates the normativity of discourse itself, hence of reasoning – in two senses, actually. First, due to the fact that each participant’s intrapersonal doxastic cum inferential activity takes place in the context of mutual scorekeeping, this intrapersonal activity appears to her as governed by shared, objective norms. Treating one’s own doxastic cum inferential activity as governed by objective norms involves treating it as governed by independent standard of correctness governing everybody’s doxastic cum inferential activity, hence it involves sensitivity that one’s own doxastic cum inferential activity may be mistaken in light of these standards. The mutual interpersonal scorekeeping activity is the reason why each participant has this sensitivity. Second, Brandom is a realist about norms, and argues that the combination of intrapersonal doxastic cum inferential activity and interpersonal scorekeeping activity, distributed over the participants, somehow institutes or creates the norms governing autonomous discourse, and in particular the norms of reasoning, themselves. In the reminder, I shall say nothing about whether the norms of reasoning themselves, if any, can be explained in this way, and shall focus entirely on the first claim, that a participant’s sense of being governed by objective norms is due to the fact that her intrapersonal doxastic cum inferential activity takes place in the context of mutual scorekeeping.

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

65

4.3  Mindreading By “higher-order thoughts” I shall mean in the following any cognitive states with meta-representational conceptual contents, that is, any states that in virtue of their conceptual content represents someone as being in certain psychological states or as being psychologically active in certain ways, or as being correct, mistaken, required, allowed, or committed etc. to be in these states or activities. Clearly, in autonomous discourse, the interplay between the participants’ intrapersonal and interpersonal activity of acknowledging and attributing commitments can help to non-circularly explain each participant’s intrapersonal activity as rational, concept-using activity only if the interpersonal scorekeeping is implicit in the sense that it is unmediated by higher-order thoughts. If scorekeeping in autonomous discourse involved unexplained abilities to conceptually meta-represent an interlocutor’s assertional and inferential commitments – as acknowledged or as had – the envisioned explanation would be viciously circular. Thus, in autonomous discourse, a scorekeeper must attribute to each interlocutor H a system of beliefs – that is, of acknowledged assertional and inferential commitments  – and evaluate this system, without forming higher-order thoughts with contents of forms such as that H (correctly/incorrectly) believes that p that H is assertionally committed to S that H should not believe that q that H (correctly/incorrectly) infers S′ from S that H (correctly/incorrectly) treats S and S* as incompatible etc.

containing psychological concepts such as belief, inferring, treating as incompatible, etc. and normative concepts such as correct, incorrect, commitment, should, etc. But how should such implicit scorekeeping be positively understood? Brandom does not tell us, but is satisfied with characterizing it merely negatively, as the ability to track and evaluate another participant’s doxastic cum inferential activity without forming higher-order thoughts. Nor does he offer, as far as I can see, any philosophical arguments that we must have the ability to keep score implicitly, no matter how this is to be positively understood - by contrast to the regress of rule following arguments he offers in support of the idea that we must be able to follow norms implicitly. Yet while stipulating a merely negatively characterized ability to keep score implicitly is perhaps not inherently problematic, it is problematic in light of contemporary debates on the psychology of mindreading, that is, of our ability to interpret, explain, and evaluate the behavior of others by attributing mental states and activities to them. Although much is controversial in this literature, proponents of the various mainstream approaches today widely agree that tracking and evaluating each other’s beliefs and reasoning essentially involves higher-order thoughts. This is most obvious for the variety of so-called theory-theory, according to which all tracking of another’s beliefs and reasoning consists in explicit higher-order thinking about the other’s beliefs and reasoning, based on the attributor’s explicit (albeit

66

R. Loeffler

perhaps unconscious) knowledge of a psychological theory of the mind (e.g. Gopnik and Meltzoff 1997; Scholl and Leslie 1999). But it is also a feature of most versions of the simulation theory, the most prominent alternative approach, according to which tracking another’s beliefs and reasoning is a matter of simulating his doxastic and inferential states and activities by using one’s own mind as the simulator. At first glance, the simulation theoretic approach seems to explain our ability to track another’s beliefs and reasoning not in terms of our ability to form higher-order thoughts about his mind, but rather in terms of the ability to put our own minds temporarily into a frame similar to his, and it would thus seem to have the potential to offer a positive theoretical underpinning of Brandom’s merely negative concept of implicit scorekeeping. However, for various technical reasons, most simulation theorists concede that, in the end, reading another’s mind, while largely a matter of mental simulation, does also essentially involve higher-order thoughts about the other’s mental states and processes (e.g. Currie and Ravenscroft 2002; Goldman 2006; Heal 2003. Cf. however Gordon 1995, 1996). Given this concession, most simulation theorists and many theory-theorists presently advocate hybrid approaches combining theory-theoretic and simulation theoretic elements, with much of their debate concerning the degree to which each element contributes to our overall mindreading ability (e.g. Botterill and Carruthers 1999; Nichols and Stich 2003; Currie and Ravenscroft 2002; Goldman 2006; Heal 2003). Anyway, the concession that mindreading via mental simulation is mediated by higher-order thought renders the simulation theoretic framework unsuitable as an underpinning of Brandom’s concept of implicit scorekeeping after all. Thus, from the perspective of present mainstream approaches to the psychology of mindreading, Brandom’s merely negative conception of implicit scorekeeping, so pivotal for his attempt to explain reasoning in terms of mutual scorekeeping, is a non-starter. If tracking and assessing each other’s doxastic cum inferential activity essentially involves higher-order thoughts, Brandom’s use of a merely negatively characterized conception of implicit scorekeeping as an explanatory primitive is deeply problematic. In the reminder, I shall outline a strategy for those sympathetic to Brandom’s project to reply to this challenge.

4.4  Belief as a We-Attitude Some contributors to the contemporary literature on collective intentionality prominently argue that we humans are able to act based on two fundamentally different, mutually irreducible kinds of intentions: I-intentions and we-intentions. If an individual acts based on we-intentions, she acts in some sense as a member of a group of rational agents – “us together” – that is, the action is in some sense on behalf of the group or contributes to the overall group action, and is recognized by the agent as such. On the other hand, actions based entirely on I-intentions are in pursuit of the individual agent’s merely personal goals; the agent acts as a private individual, for merely personal reasons, in abstraction from her membership in any group.

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

67

These contributors to the collective intentionality literature moreover argue that an individual’s actions as a member of, or on behalf of, a group cannot be fully explained in terms of ordinary intentions on her and other group members’ part plus mutual belief – I-intentions plus mutual belief – but must be understood in terms of intentions of a second, distinctly intersubjective kind: we-intentions.3 Yet what are we-intentions? Some of the contributors in question hold that the group as such is literally a subject of agency of its own, above the individual human agents making up the group, and that we-intentions are intentions formed by this group subject (e.g. Gilbert 1989, 2006, 2014; Pettit 2001a, b, 2003; List and Pettit 2011). Yet most insist that the only cognitive subjects involved in group actions are the individual we-intending human members of the group. Agreeing, moreover, that we-intentions are literally propositional attitudes  – cognitive attitudes taken by a subject towards propositional contents that p  – this majority then splits into two camps. Proponents of content-based approaches locate the distinctive, intersubjective feature of we-intentions in the content of these intentions, paradigmatically in certain first-person plural pronominal elements plus further conceptual elements representing joint agency or goals (e.g. Bratman 1999), while proponents of mode-­ based approaches locate it in the attitude or mode of we-intending, as opposed to the content of the intention (e.g. Tuomela and Miller 1988; Tuomela 2003, 2005). On the mode-based approaches, we-intentions are we-attitudes: qua cognitive attitudes they have a certain intersubjective, first person plural dimension built into them that renders them apt to yield group agency. Without dwelling more on we-intentions, and setting aside for present purposes the debates between their proponents and opponents, I would like to shift immediately to belief and recommend that we think of beliefs as propositional we-attitudes too, in ways that mutatis mutandis accord with mode-based approaches to we-­ intentions. The point, apropos my goal to develop a positive conception of implicit scorekeeping, is that if a certain intersubjective, first-person plural dimension is built into the belief attitude then believing that p oneself may ipso facto involve recognizing others – everyone of “us” – as believers too. Yet since beliefs that p are conceptually contentful, representational states only in virtue of their content that p – the attitudinal aspect of a belief that p is not some extra conceptual content added to the content that p, but something else entirely – this treatment of others as believers by believing that p oneself, while higher-order in the sense that it is a recognition of others as believers, is not a matter of higher-order thoughts in the sense in which I use the term here, and is, accordingly, implicit. My expository strategy in the reminder is to develop this conception of belief and the ensuing conception of implicit scorekeeping through a sequence of eight proposals about belief and Brandomian discursive activity. With each new proposal a more refined picture of belief and discursive activity will emerge, which will allow us to address some of the questions that may have arisen at previous stages. Some of these eight proposals will be more controversial than others. Still, I cannot

 For an overview of these debates cf. Schweikard and Schmid (2013).

3

68

R. Loeffler

sufficiently defend any of them here, but hope instead that the proposals in toto will yield  an attractive picture of belief and implicit scorekeeping that merits further exploration. My first proposal is to take the familiar thought seriously that belief is, literally, a propositional attitude and, more specifically, an attitude of holding true.4 For a subject to believe that p, that is, for her to take a token belief attitude towards a content that p, is for her to ipso facto treat p as true. Simply taking the belief attitude towards p suffices for the subject to take p as true. Yet setting relativistic treatments of truth aside, truth is objective – in two senses, actually. First, truth is objective in the realist sense that p is true if and only if p is the case, independently of what anyone or everyone may think. Second, truth is objective in the intersubjective sense that p is true if and only if p is true for everybody at all times, again independently of what anyone or everyone may think. In the reminder, I want to focus exclusively on this second, intersubjective sense of objectivity. My second proposal is that, given that believing is holding true, having any beliefs at all involves having a sense of truth as objective in this interpersonal sense. Specifically, I propose that believing that p – holding p true by taking the belief attitude towards p – intrinsically involves a sense on the believer’s part that p is true for everybody, that is, true for every rational, concept-mongering, doxastic and epistemic subjects at all, or (setting things up for Brandomian purposes) for everybody with whom the subject could in principle discourse. According to this proposal, a certain dimension of interpersonal recognition, extending to all rational beings, is thus intrinsic to believing that p. This dimension is intrinsic to believing that p, in that it is simply a matter of taking the belief attitude towards p, not a matter of further cognitive activity on the subject’s part in addition to believing that p – such as accompanying higher-order thinking that p is true for everybody or some kind of accompanying mental simulation. Of course, we can and do form higher-order thoughts and we may also engage in mental simulation. Indeed, I entirely agree that forming higher-order thoughts plays a central role in our social life (cf. below). But I propose that no accompanying higher-order thoughts or mental simulations are responsible for yielding the subject’s sense, in so far as she believes that p, that p is true for everybody. Even if p is a first-order content concerning some humdrum non-social state of affairs that p, the subject, just by believing that p, still ipso facto treats p as true for everybody, unaided by higher-­ order thoughts or mental simulation. My third proposal is to think of this interpersonal dimension  – the subject’s sense, in believing that p, of p as true for everybody – as a kind of normative attitude: a treatment by the subject of p as to be held true (endorsed) jointly by everybody, or, more precisely, a treatment of p as something everybody is doxastically committed to. (I shall clarify this sense of “commitment” below). The proposal is that, just by taking the belief attitude towards p, the subject ipso facto treats p as  For overview, cf. Schwitzgebel (2015). Alternatively, one may motivate the idea by starting with the thought that belief allows for interpersonal disagreement. Sellars mutatis mutandis pursues this strategy by arguing that moral judgments, given that they are, as he thinks, intentions and allow for interpersonal disagreement, must be we-intentions (cf. Sellars 1968: Ch. 7). 4

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

69

what everybody is committed to hold true, or, since belief is a we-attitude on the present proposal, as what “we together” are jointly committed to hold true, where the “we” includes all rational beings. I note in passing that cashing out the believer’s sense of truth as objective in the second, intersubjective sense as a normative attitude towards everybody echoes the position that truth itself, or anyway true belief, is intrinsically valuable. On this view, we should ceteris paribus strive to acquire true beliefs and abandon false ones just because the former are true and the latter false, not merely because of instrumental or prudential considerations. However, my third proposal, that believing that p intrinsically involves treating p as to be held true jointly by “us together”, or as what “we together” are jointly committed to, does not imply this view. In general, the proposal concerns what one does in holding p true by taking the belief attitude towards p, not the value of truth or true belief. In general, I don’t think this proposal has implications regarding the nature of truth itself (except, perhaps, that truth is objective). It does not imply whether truth is a matter of correspondence, coherence, deflation, or something else, nor whether true belief is intrinsically valuable  – although it does echo this last view. Before elaborating further, a few passing remarks on how the proposed treatment of belief as a we-attitude relates to we-mode approaches to we-intentions. First, while all these approaches stipulate we-intentions in addition to I-intentions, no such contrast exists for belief, on the present proposal. All beliefs are we-attitudes. After all, all beliefs are instance of holding true, and therefore (on the present proposals) treatments of the content believed as to be endorsed by “us jointly together”, where this treatment is built into the belief attitude taken towards that content.5 Second, while, in debates on we-intentions, the “we” is usually restricted to the members of a more or less exclusive group, the sense of “we” involved in belief is all-inclusive, extending to all rational beings. This is again due to the fact that believing that p is holding p objectively true, hence true for everybody at all times. However, the idea of an unrestricted “we” is not unheard of in debates on collective intentionality. In fact, Wilfrid Sellars’ classical account of moral judgments as a type of we-intention, which is a chief source of inspiration for some proponents of contemporary mode-based approaches of we-intentions (Tuomela and Miller 1988), centrally employs the idea of a “we” that unrestrictedly includes all rational beings – precisely to account for moral objectivity, that is, to legitimize the use of the concept of truth in ethical discourse (e.g. Sellars 1968: 222). My proposal regarding belief follows Sellars’ approach to moral judgments in this regard and is inspired by it. Third, proponents of mode-based accounts of we-intentions usually explicate the stipulated interpersonal dimension of we-intentions in descriptive terms, as a sense  There is, however, the issue of a potential contrast between individual group members’ beliefs and the collective beliefs of the group, discussed especially by Tuomela (e.g. 2007) and Gilbert (e.g. 2014). I have to leave these issues for another occasion. But in any case, the difference between group beliefs (if any) and group members’ beliefs would not well mirror, on my theory, the difference between we-intention and I-intention since, on all accounts, I-intentions lack the kind of interpersonal dimension that both beliefs (on my account) and we-intentions (on all accounts) have. 5

70

R. Loeffler

on the agent’s part that “we jointly perform action X” (e.g. Tuomela 2007: 93–4; Searle 1990: 406–8), rather than in normative terms (“we should jointly perform X”). This reflects the fact that they are mainly interested in explaining actually occurring group actions. By contrast, I propose to explicate the interpersonal dimension of belief in terms of a normative attitude (“we should jointly endorse p”). That is, the proposal is not that believing that p involves treating p as something that all of us in fact jointly endorse, but treating it as a normative constraint on every doxastic perspective, regardless whether anyone else endorses p in fact. Fourth, it is controversial among proponents of mode-based approaches to we-intentions whether an individual agent can all by herself we-intend to X, without other members of the group we-intending to X as well. Some argue that this is possible (e.g. Searle 1990), while others disagree (e.g. Tuomela 2005: 333, 339; Schmid 2003). Since it is clearly possible for someone to believe that p even though no one else believes that p, I am mutatis mutandis committed to the minority view on this issue: although belief is a we-attitude, it is certainly possible to believe that p alone. My treatment of the interpersonal dimension intrinsic in believing that p as a normative attitude taken towards everybody, rather than a feature of belief that could be fully explicated in descriptive terms, should make this commitment palatable and, indeed, natural. Believing that p involves treating p as to be endorsed by us jointly together, yet doing so does not imply that that the believer treats anyone else as meeting this normative expectation in fact. The normative expectation intrinsic to believing that p oneself, that p is to be endorsed by everybody, is often unreasonable or otherwise faulty. The subject may lack good reasons for believing that p, and thus should stop endorsing p rather than continuing to do so and thereby normatively expect everybody to endorse p too. Or p may be false, in which case the implicit normative expectation on everybody intrinsic in a subject’s believing that p is still faulty, even if the belief is supported by good reasons. Moreover, even if the subject has good reasons to believe that p and p is true, if we think of the normative attitude she thereby takes towards everybody as an attribution to everybody of an obligation to endorse p, that normative attitude (though not the belief itself) is still unreasonable for many of us. The normative expectation is unconditional, that is, formed without regard for anyone else’s particular epistemic resources, conceptual repertoire, cognitive capacities, deeply held convictions, interests, or plans. After all, if p is true, p is true for everybody, irrespective of anybody’s de facto experiences, knowledge, beliefs, conceptual repertoire, cognitive capacities, interests, and projects. Yet since others often cannot endorse p unconditionally, but only after potentially extensive doxastic, epistemic, semantic, etc. stage setting, and since it is furthermore often unreasonable to require others to submit to such stage setting, given their plans, interests, motivations, etc., the unconditional normative attitude taken by the subject believing that p towards indiscriminately everybody is per se often unreasonable, if we think of it as an attribution to everybody of an obligation to endorse p. However, we may instead think of it as an attribution of a commitment, in something like Brandom’s technical sense. This is my fourth proposal. Commitments are much like obligations. Both are normative statuses that constrain our cognitive or

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

71

social lives, and when we acknowledge or act based on them our doing so is always in this respect right (though it may be mistaken in other respects). However, whereas I’m obliged to do x only if I can do x – “ought” implies “can” – I can be committed to things although I cannot acknowledge commitments to all of them. For example, I’m committed to every implication of what I believe, and those commitments normatively constrain what else I may properly believe. But since I can’t draw every inference of what I believe, I’m not obliged to acknowledge each of those inferentially inherited commitments. Applying the distinction between obligation and commitment to the present issue, we may think of the interpersonal normative attitude intrinsic in believing that p not as an attribution of an obligation to everybody to endorse p but as an attribution to everybody of a doxastic commitment to p. From the viewpoint of the subject believing that p, “we together” are unconditionally doxastically committed to p. And if the belief that p itself is true, this attribution of an unconditional doxastic commitment to p to everybody is to this extent appropriate, whether or not others are in an epistemic, semantic, cognitive, or conative position to actually endorse p sooner or later and, indeed, whether or not the subject herself has good reasons to believe that p. My fourth proposal, in short, is that the interpersonal normative attitude intrinsic in believing that p is a treatment by the subject of p as something everybody is in this sense doxastically committed to, not something everybody is obliged to endorse. Reflective, mature believers often realize that others cannot live up, now or ever, to the doxastic commitments they attribute to them in this way, though unreflective ones will often fail to do so – think of authoritarian leaders, controlling parents, or didactically untalented, nerdy college professors. How is such realization achieved? My fifth proposal is that higher-order thoughts play a central role here. Although believing that p intrinsically involves treating everybody as unconditionally doxastically committed to p, the subject can, while continuing to believe that p, recognize others as not in a position to endorse p, now or ever, by forming collateral higher-­ order beliefs about their present de facto beliefs, knowledge, conceptual repertoire, cognitive capacities, interests, plans, etc. If the subject, by forming corresponding higher-order thoughts while continuing to believe that p, takes the truth of p to matter for those others, then what they really should do, according to her, is to improve their doxastic, epistemic, semantic, cognitive, or conative situation so as to be in position to endorse p as well. Yet at the same time the subject may regard them as unable to live up to this commitment, presently or for good, by forming collateral higher-order thoughts of them as in their specific current doxastic, epistemic, semantic, and conative situation, and thus at least presently incapable to endorse p. More generally, my fifth proposal is that while higher-order thoughts are not essential for recognizing others as doxastic and epistemic subjects – we also do so simply by forming any first-order beliefs that p ourselves and by thereby holding everybody, including them, as doxastically committed to p – higher-order thoughts still play an important role in fully recognizing others as doxastic and epistemic subjects. Higher-order thoughts play a central discourse-guiding and discourse-­ stabilizing role in situations of teaching and learning, and in navigating disputes. If we form them as teachers, they help us see which further epistemic, semantic,

72

R. Loeffler

cognitive, or conative resources we would have to provide to bring our students into position to live up to the doxastic commitments we attribute to them simply by having a system of beliefs ourselves. If we form them as student, they help us identify others as potentially in possession of additional epistemic, semantic, cognitive, or conative resources that, if we had them ourselves, would improve our own doxastic, epistemic, semantic, cognitive, or conative situation. And if we form relevant higher-order thoughts in disputes regarding p – where we believe that p, hence treat everybody as doxastically committed to p, and yet others manifestly refuse to endorse p (either by remaining agnostic or by advancing beliefs contradicting p) – they can help us not only to sustain our own convictions in the face of the dispute, rather than to cave or become confused (a point to which I shall return), but also to understand the other’s reasons and motives, thus enabling us not only to make sense of their refusal to meet our normative expectations as we continue to believe that p, but also, potentially, to see them as offering reasons for us to alter our doxastic perspective. Obviously, forming higher-order thoughts that he believes that p is thus one way for us to treat another as de facto believing that p – as opposed to treating him as doxastically committed to p. Yet my sixths proposal is that there is also plausibly a second way, unmediated by higher-order thoughts. A subject believing that p, hence treating “us together” as jointly doxastically committed to p, may simply recognize particular others as living up to this normative expectation, hence as de facto endorsing p jointly with herself, by detecting outward linguistic or non-linguistic signs manifesting the others’ endorsement of p – their nods or “yes”es in response to the subject’s asserting that p, assertions that p on their parts, their manifest perceptual attention to the fact that p (if the belief that p is perceptual), etc. That is, my sixth proposal is that believing that p oneself, hence ipso facto treating everybody as doxastically committed to p, plus detecting outwards signs by particular others manifesting their endorsement of p, suffices for the subject to treat them as (rightfully) de facto endorsing p jointly with her – unmediated by higher-order thoughts that they (rightfully) believe that p and thanks to the built-in interpersonal dimension intrinsic to the subject’s own belief that p. Recall that my present goal is to outline a positive conception of attributing beliefs that p to others “implicitly in practice”, unmediated by higher-order thoughts that they believe that p. I take it that I have now reached this goal. Key is the idea that beliefs that p are cognitive we-attitudes, inspired by we-mode approaches to we-­ intentions. According to this idea, just by taking the belief attitude towards a content that p, the subject ipso facto treats “us together” as jointly doxastically committed to p, where the “us” includes every rational being as such. Believing that p oneself thus intrinsically involves a normative expectation on indiscriminately everybody to endorse p. And this attribution is implicit in the straightforward sense that it is due to the belief attitude taken towards p, not due to the content that p nor due to collateral higher-order thoughts on the subject’s part that we together are doxastically committed to p. I started with the idea that taking the belief attitude towards p is treating p as true. Moreover, according to standard propositional attitude psychology, a belief that p is a conceptually contentful, representational state only in virtue

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

73

of its content that p, not qua cognitive attitude taken towards p.6 Holding p true by taking the belief attitude towards p is not concatenating p with additional concepts. If we concatenate p with the concept __ is true or the concept everybody is committed to __, we get the contents p is true or everybody is committed to p, leaving open whether we hold p, or the contents we get through the concatenations, true. Thus, holding p true by taking the belief attitude towards p amounts to something entirely different from concatenating p with further concepts. Whatever that is, my key suggestion is that it intrinsically involves treating every rational being (“all of us together”) as doxastically committed to p – considering that truth is objective in the intersubjective sense that what’s true is true for everybody. Of course, in formulating this proposal I use the concepts holding true, being true, being jointly doxastically committed, etc. to characterize what we do when we take the belief attitude towards p. But it does not follow that taking the belief attitude towards p consist in, or involves, employing these or any other concepts. Whatever it is that we do when we take the belief attitude towards p, it is non-conceptual and, in this sense, implicit. If all this is correct, we have now in outline two related positive conceptions of implicitly recognizing others as believers. First, we implicitly attribute to everybody, and hence to any particular others we encounter, an unconditional doxastic commitment to p, simply by believing that p ourselves. Second, we implicitly attribute to particular others de facto (and rightly) held beliefs that p by, again, believing that p ourselves and thereby treating “all of us” as jointly doxastically committed to p, and detecting public signs by those others manifesting their endorsement of p – “yes”es or nods in response to our assertion that p, their assertions that p, signs of their perceptual attention to the fact that p (if the belief that p is perceptual), etc. Detecting such signs in the context of believing that p ourselves suffices, I propose, to regarding them as living up to the doxastic commitment to p that we implicitly attribute to everybody by believing that p ourselves. No higher-order thoughts that they (rightly) believe that p are needed. These two conceptions of implicitly attributing beliefs to others do not capture the whole range of our abilities to treat others as believers. As suggested, many of our more sophisticated ways to treat each other as doxastic and epistemic subjects involve higher-order thoughts and are thus not implicit. Still, as I shall argue now, the developed two positive conceptions of implicit belief attributions, suitably adjusted to Brandom’s theoretical framework, suffice to yield a positive account of implicit scorekeeping in the context of simple, autonomous discourse.

 Pace Schmitz (2017).

6

74

R. Loeffler

4.5  Scorekeeping Implicitly in Practice Recall that, according to Brandom, the most basic form of reasoning is the discursive reasoning that is integral to competently participating in certain simple exchanges of assertions (autonomous discursive practices). It consists in acknowledging successive constellations of assertional and inferential/incompatibility commitments, in systematic interaction with the development of the conversation. Acknowledging assertional commitments concerns individual declarative sentences S and consists in being disposed to assert S in accordance with what one implicitly takes to be the norms of discourse. Acknowledging inferential or incompatibility commitments concerns pairs of sentences S and S′ and consists in treating, or being disposed to treat, S′ as following from S (or vice versa), or S and S′ as mutually incompatible, in the given discursive context, in accordance what one implicitly takes to be the norms of discourse. (In the following, I shall again use “inferential commitment” inclusively, to cover both inferential and incompatibility commitments.) To acknowledge a set of inferential commitments relating some sentence S to some other sentences S′, S″, etc., is to treat S as having a certain inferential role, hence, given inferential role semantics, to interpret S as having a certain propositional meaning that p. To acknowledge an assertional commitment to S itself is, in the context of acknowledging these inferential commitments, to believe that p. Discursive reasoning takes place in the context of mutual scorekeeping. At each stage of the conversation, each participant not only acknowledges a constellation of assertional and inferential commitments but also implicitly attributes to each interlocutor two such constellations, and changes these attributions as the conversation progresses, given what the participant takes to be the norms. First, she attributes to each interlocutor a system of assertional and inferential commitments as de facto acknowledged by him. Second, she attributes to each interlocutor a system of assertional and inferential commitments as to be acknowledged by him – the system that, according to the speaker, the interlocutor really is committed to, whether or not he acknowledges these commitments. According to Brandom’s pragmatism, all scorekeeping in autonomous discourse is implicit, in the sense that it is unmediated by any higher-order thoughts about an interlocutor’s unacknowledged or acknowledged commitments. My goal is to offer a positive conception of implicit scorekeeping, and to indicate how, given this conception, each participant’s intrapersonal activity of acknowledging discursive commitments, because it takes place in the context of mutual implicit scorekeeping, appears to her as governed by objective norms. Obviously, I want to use the developed conception of belief as a we-attitude, suitably adjusted, to reach this goal. My seventh proposal is to regard both acknowledged assertional and acknowledged inferential commitments as we-attitudes in mutatis mutandis the developed sense, this time, however, we-attitudes taken towards, respectively, individual sentences and pairs of sentences rather than towards propositional contents simpliciter. Thus to acknowledge an assertional commitment to S is to be disposed to perform S and it intrinsically involves, given

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

75

the proposed interpersonal nature of the attitude taken towards S, treating S as something everybody – every rational being – is committed to (whether or not anyone else acknowledges it). From the viewpoint of the speaker acknowledging an assertional commitment to S, S is ipso facto such that “we together” are assertionally committed to S. Similarly, to acknowledge an inferential (or incompatibility) commitment regarding S and S′ is to draw, or to be disposed to draw, the inference from S to S′, or vice versa, or to treat, or to be disposed to treat, S and S′ as mutually incompatible, and it intrinsically involves treating the inference or incompatibility as something everybody is committed to. Treating acknowledged assertional commitments as such we-attitudes makes sense in light of general considerations about assertion – although I cannot get into the nuances of the literature on assertion here. Assertion is widely viewed as the kind of speech act that is appropriate only if it advances the sentence asserted, or its content, as true (or as known to be true, or as justified as true). Accordingly, a competent speaker who acknowledges an assertional commitment to S, thereby treats S, or its content, as (known to be, or justified as) true. If so, similar considerations apply to acknowledging assertional commitments as applied earlier to beliefs as attitudes of holding true. Acknowledging an assertional commitment to S involves treating S as (known to be, or justified as) objectively true, hence as true for everybody, hence as something everybody is committed to, whether or not others acknowledge this commitment. It also makes sense in light of Brandom’s specific theory of assertion, according to which the defining feature of assertion is that asserting S prima facie authorizes every interlocutor or overhearer to re-assert S (Brandom 1994: 170, 174). Among types of speech acts, assertion is peculiar in that making a speech act of this type always authorizes without further ado others to simply repeat that speech act – albeit merely prima facie, that is, absent overriding considerations voiding the authorization. My proposal, that acknowledging an assertional commitment to S involves treating everybody as assertionally committed to S, broadly accords with this characterization of assertion as the kind of speech act that prima facie licenses its own spread through the community.7 Treating acknowledged inferential commitments as such we-attitudes makes sense as well, I think. Our inferential activity is governed by objective norms of inference, according to Brandom, in the Kantian sense of norms binding every rational being.8 Accordingly, it makes sense that acknowledging inferential commitments intrinsically involves treating the inferences to which commitment is acknowledged as something everybody is bound or committed to.

 “Broadly” because my proposal distorts Brandom’s views somewhat. For him, competent participants treat only actual assertions, not just any acknowledged assertional commitments, as having this interpersonal normative force. Moreover, he regards this force as licensing entitlement, not commitment, to re-assert. 8  At least according to the Brandom of Making it Explicit. For qualifications in light of his subsequent work, cf. Loeffler (2018, Ch. 8). 7

76

R. Loeffler

Given the proposed treatment of acknowledging assertional and inferential commitment as we-attitudes, acknowledging a system of such commitments oneself thus intrinsically involves taking a complex normative attitude towards every rational being – every actual or potential interlocutor in discourse. In acknowledging a system of such commitments oneself, one ipso facto treats indiscriminately everybody (“all of us together”) as committed to that very system. This complex normative attitude is yet again implicit and due to the interpersonal character of the attitudes of acknowledging assertional and inferential commitments, not due to higher-order conceptual contents. Neither is this complex normative attitude towards everybody derived from the “contents” of one’s attitudes of acknowledging discursive commitments (the corresponding individual sentences or pairs of sentences, qua syntactic strings), nor from the propositional meanings of those sentences (their inferential roles), nor is it due to accompanying higher-order thoughts that everybody is committed to such and such assertions or inferences. Similar elaborations as made earlier regarding beliefs as we-attitudes apply here. In acknowledging a system of assertional and inferential commitments, the speaker ipso facto treats everybody as unconditionally committed to this system, yet many others are for various reasons presently in no position to acknowledge the commitments attributed to them. Again, normal, mature speakers often recognize others as in no such position, and I propose yet again that forming relevant higher-order thoughts – in the Brandomian framework, acknowledging assertional commitments to meta-representational sentences such as “He doesn’t speak English,” “I haven’t spoken yet, and so he doesn’t know yet what he should think,” “He is convinced of such and such, and so won’t accept it” etc.  – play a central role here. In the Brandomian framework, the ability to form such higher-order thoughts is a matter of competency to participate in corresponding non-autonomous meta-­ representational linguistic practices. Thus a merely autonomous discursive practitioner would be severely limited, according to my proposal in the context of this framework, to recognize others as presently unable to acknowledge the commitments she implicitly attributes to them, and would overcome this limitation only as she gains fluency in the relevant non-autonomous discourses. Still, this initial limitation should not hinder her to participate in autonomous discourse. I have now developed the sought positive conception of implicitly attributing assertional and inferential commitments as to be acknowledged to interlocutors. How about the positive conception of implicitly attributing assertional and inferential commitments as de facto acknowledged? Earlier I proposed that someone who believes that p, and thus implicitly treats everybody as doxastically committed to p, may recognize others as meeting this normative expectation unaided by higher-­ order thoughts, just by detecting outward signs manifesting their endorsement of p. I propose the same move here. Each conversation is awash in outward signs directly or indirectly manifesting an interlocutor’s acknowledgement of particular assertional and inferential commitments. I take it that the combination of a scorekeeper’s implicit treatment of everybody as sharing the assertional and inferential commitments she herself presently acknowledges plus her detection of public signs on some interlocutor’s part manifesting his acknowledgement of some of these

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

77

commitments, suffices for the scorekeeper to treat the interlocutor as de facto (and correctly) acknowledging these commitments jointly with her. Additionally, if the other isn’t a stranger, the scorekeeper will moreover remember an abundance of displays by him of other such signs in their past interactions, which will suffice for her to implicitly treat him as de facto (and correctly) sharing presently many other portions of her current system of acknowledged assertional and inferential commitments as well. And if the other is a stranger, she will remember from past conversations with other interlocutors myriad outward signs indicating general acceptance of many of her present assertional and inferential commitments – say, to demonstrative assertions of the type “That’s a chair,” “That’s red,” etc., and to such inferences as from “That’s crimson” to “That’s red,” etc. Based on these past experiences with varying interlocutors, she will simply take for granted, unaided by higher-order thoughts or present outward signs, that her present interlocutors de facto (and correctly) acknowledge such “trivial” discursive commitments jointly with her.

4.6  A Sense of Objectivity So far, the developed conception of acknowledged assertional and inferential commitments as we-attitudes accounts for one aspect of a speaker’s sense of being governed by objective norms. The speaker, in acknowledging a constellation of such commitments, ipso facto treats everybody as committed to that very constellation and, in this sense, as bound by it. But it fails so far to account for another aspect of a speaker’s sense of being governed by objective norms: the sense that her presently acknowledged constellation of commitments may be mistaken in some respects. That is, the developed conception of acknowledged discursive commitment as we-­ attitude does so far not account for the speaker’s sensitivity to the contrast between seeming objectively right herself and being objectively right. It accounts for her sense of discursive authority, that is, her sense that, in acknowledging a constellation of discursive commitments and thereby seeing things a certain way, she authorizes or requires everybody to see things that way too. But it fails so far to account for the speaker sense of discursive responsibility, that is, her sense that her presently acknowledged constellation of discursive commitments may be mistaken in certain respects, thus that standards of correctness govern her presently acknowledged constellation of discursive commitments that are in place independently of her presently acknowledging this constellation of commitments. My eighth and last proposal aims to account for this sense of being potentially mistaken oneself. The proposal is that in discourse, every participant’s assertions and other outward signs manifesting or expressing aspects of his or her system of acknowledged assertional and inferential commitments (his or her own assertions, or his or her “yes”es, “no”s, nods, headshakes, signs of hesitancy or neutrality, etc. in response to the speaker’s assertions) come, and are recognized by each participant as coming, with the same unqualified normative pressure on everybody with which each participant takes his or her own acknowledged discursive commitments,

78

R. Loeffler

and their manifestations and expressions, to come. For example, if an interlocutor asserts S, S is advanced, and is understood by the speaker (the addressee) as advanced, as something “we together” are committed to, hence as something that is to be accepted by everybody. The performance thus places, and is recognized by the speaker as placing, everybody under the unconditional normative expectation to acknowledge an assertional commitment to S. Similarly, if an interlocutor says “no” to the speaker’s assertion of S′, his “no” in this context demands, and is recognized by the speaker as demanding, everybody’s non-acknowledgement of an assertional commitment to S′. Thus, immediately after this “no,” speaker and interlocutor find themselves, and everybody else, as jointly faced with conflicting demands – that commitment to S′ is to be jointly acknowledged by “us together” (given the speaker’s assertion of S′) and not to be jointly acknowledged by “us together” (given the interlocutor’s “no”) – which puts them, and everybody else, under the obligation to ceteris paribus hash our the conflict and to align their attitudes regarding S′ one way or another. And if the interlocutor furthermore advances an assertion of S*, and speaker and interlocutor jointly acknowledge both an incompatibility commitment regarding S* and S″ and an inferential commitment from S′ to S″, and the speaker (though not the interlocutor) moreover acknowledges an assertional commitment to S″ in consequence of acknowledging assertional commitment to S′, S* is not only advanced (and understood by the speaker as advanced) as to be accepted together, but also, in the context of these collateral acknowledgements, as a reason against acknowledging commitment to S′ together. In general, according to this picture, from every participant’s viewpoint, her distinct system of acknowledged discursive commitments and anyone’s assertions and other outward signs manifesting or expressing aspects of his distinct system of acknowledged discursive commitments, come with the demand on “all of us together” to be accepted. Accordingly, whenever two participants’ respective systems of acknowledged discursive commitments manifestly differ  – one acknowledges a commitment that the other doesn’t acknowledge in general, or one acknowledges a commitment that is acknowledged by both as incompatible with a commitment the other acknowledges in particular – both participants ipso facto find themselves, and everybody else, called upon to align their respective systems of acknowledged discursive commitments one way or another, and to thereby establish the interpersonal agreement that their respective implicit normative demands on indiscriminately everybody calls for. In light of each participant’s frequent experience of non-compliance and disagreement in discourse, of joint efforts to re-align, and of the frequent sense that, once re-alignment has been achieved, one’s own, altered perspective has actually improved, each participant will have a sense that many – indeed, perhaps any – discursive commitments she presently acknowledges, although she treats them as to be acknowledged by “us together”, may in principle turn out to be nothing anyone has actually ever been committed to, hence nothing that she herself should ever have acknowledged commitment to in the first place. In other words, each participant will have a sense that her presently acknowledged system of discursive commitments

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

79

may be mistaken in some – indeed, perhaps any – respects, hence of the contrast between seeming objectively right and being objectively right.

4.7  Conclusion The proposed marriage between Brandom’s pragmatist theory of discursive rationality and we-mode accounts of propositional attitudes raises many questions. Regarding Brandom’s theory, the marriage clearly accords with his favored I-thou approach to norms of reason, according to which what’s really correct is not determined by some property of the community of discursive beings, but is rather a matter of ongoing, potentially open-ended, piecemeal negotiation and re-negotiation in the context of the larger history of discoursing about the issue at hand – with each speaker being potentially an important contributor and each participant having a merely formal sense of objective correctness (Brandom 1994: 599–601; 2009: Ch. 3). But the proposed marriage calls for a revision of some important details of Brandom’s specific theory of scorekeeping as keeping of “separate books”, which may affect other aspects of Brandom’s theory  – most importantly his favored account of communicative success given semantic holism – and thus invites more reflection.9 Moreover, the proposed theory of belief as a propositional we-attitude, combined with the role this theory assigns to higher-order thoughts, raises many questions in pragmatics (concerning, for example, the mechanisms of recognizing common ground), the psychology of mindreading (concerning, for example, the psychology of mindreading in situations of teaching, learning, and disagreement, or the well-­ documented difficulty of ordinary young children to attribute false beliefs to others), and the debates on collective intentionality (concerning, for example, the nature of group belief, or the question how the developed theory of belief may contribute to our understanding of group agency). Obviously, I must leave all these questions for other occasions. Acknowledgements  Many thanks to Chris Gauker, Ryan Hebert, Ladislav Koreň, Jaakko Reinikainen, and Preston Stovall for helpful comments and suggestions regarding earlier drafts of this chapter.

 For a critique of these details, and this account, cf. Loeffler (2014, 2018: 193–9).

9

80

R. Loeffler

Bibliography Botterill, G., & Carruthers, P. (1999). The philosophy of psychology. Cambridge: Cambridge University Press. Brandom, R. (1994). Making it explicit. Reasoning, representing, and discursive commitment. Cambridge, MA: Harvard University Press. Brandom, R. (2008). Between saying and doing: Towards and analytic pragmatism. New York: Oxford University Press. Brandom, R. (2009). Reason in philosophy. Animating ideas. Cambridge, MA: Harvard University Press. Bratman, M. (1999). Faces of intention: Selected essays on intention and agency. Cambridge: Cambridge University Press. Currie, G., & Ravenscroft, I. (2002). Recreative minds: Imagination in philosophy and psychology. Oxford: Oxford University Press. Gilbert, M. (1989). On social facts. Princeton: Princeton University Press. Gilbert, M. (2006). A theory of political obligation. Oxford: Oxford University Press. Gilbert, M. (2014). Joint commitment. How we make the social world. New  York: Oxford University Press. Goldman, A. I. (2006). Simulating minds: The philosophy, psychology, and neuroscience of mindreading. Oxford: Oxford University Press. Gopnik, A., & Meltzoff, A. N. (1997). Words, thoughts, and theories. Cambridge, MA: Bradford Books/MIT Press. Gordon, R. M. (1995). Simulation without introspection or inference from me to you. In M. Davies & T. Stone (Eds.), Mental simulation: Evaluations and applications – Reading in mind and language (pp. 53–67). Oxford: Blackwell Publishers. Gordon, R. M. (1996). ‘Radical’ simulationism. In P. Carruthers & P. K. Smith (Eds.), Theories of theories of mind (pp. 11–21). Cambridge: Cambridge University Press. Heal, J. (2003). Mind, reason and imagination. Cambridge: Cambridge University Press. Kant, I. (1998). Critique of pure reason (P.  Guyer & A.  Wood, Eds.). Cambridge: Cambridge University Press. List, C., & Pettit, P. (2011). Group agency: The possibility, design, and status of corporate agents. New York: Oxford University Press. Loeffler, R. (2005). Normative phenomenalism: On Robert Brandom’s practice based explanation of meaning. European Journal of Philosophy, 13, 32–69. Loeffler, R. (2014). Belief ascriptions and social externalism. Philosophical Studies, 167, 211–239. Loeffler, R. (2018). Brandom. Cambridge: Polity Press. Nichols, S., & Stich, S. P. (2003). Mindreading: An integrated account of pretence, self-awareness, and understanding of other minds. Oxford: Oxford University Press. Pettit, P. (2001a). A theory of freedom: From the psychology to the politics of agency. New York: Oxford University Press. Pettit, P. (2001b). Collective intentions. In N. Naffine, R. J. Owens, & J. Williams (Eds.), Intention in law and philosophy (pp. 241–254). Farnham: Ashgate. Pettit, P. (2003). Groups with minds of their own. In F.  Schmitt (Ed.), Socializing metaphysics (pp. 167–193). Lanham: Rowman and Littlefield. Schmid, H.  B. (2003). Can brains in vats think as a team? Philosophical Explorations, 6(3), 201–217. Schmitz, M. (2017). What is a mode account of collective intentionality? In G. Preyer & G. Peter (Eds.), Social ontology and collective intentionality. Critical essays on the philosophy of Raimo Tuomela with his responses (pp. 37–70). Cham: Springer. Scholl, B.  J., & Leslie, A.  M. (1999). Modularity, development and theory of mind. Mind and Language, 14(1), 131–153. Schweikard, D. P., & Schmid, H. B. (2013). Collective intentionality. Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/collective-intentionality/. Accessed 1 May 2018.

4  Implicit Scorekeeping: A We-Mode Account of Belief and Interpretation

81

Schwitzgebel, E. (2015). Belief. Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/ entries/belief/. Accessed 30 May 2018. Searle, J. (1990). Collective intentions and actions. In P. Cohen, J. Morgan, & M. Pollack (Eds.), Intentions in communication (pp. 404–415). Cambridge, MA: Bradford MIT. Sellars, W. (1968). Science and metaphysics. Variations on Kantian themes. Dordrecht: D. Reidel. Tuomela, R. (2003). The we-mode and the I-mode. In F. Schmitt (Ed.), Socializing metaphysics: The nature of social reality (pp. 93–128). Lanham: Rowman and Littlefield. Tuomela, R. (2005). We-intentions revisited. Philosophical Studies, 125, 327–369. Tuomela, R. (2007). The philosophy of sociality. The shared point of view. New  York: Oxford University Press. Tuomela, R., & Miller, K. (1988). We-intentions. Philosophical Studies, 53, 367–389.

Part II

Philosophical and Empirical Cross-Pollination

Chapter 5

Normative Mindshaping and the Normative Niche Jaroslav Peregrin

Abstract  Members entering a community are inevitably formed by its “cultural framework” (via the process of enculturation). The “cultural framework”, in turn, is produced by the members of the community. The nature of this dialectical movement – producing the framework while being produced by it – has long been investigated; however, it is only recently that some scholars have started to appreciate the centrality of rules and norms for an adequate description of this phenomenon. In this paper I argue that to understand it we must give pride of place to norms at a radically foundational level – we must realize how deeply normative we as humans are. I argue that even the most promising accounts of this movement, such as those based on the concept of “mindshaping” or on the idea of “social niche construction” must be seen as essentially normative enterprises. Keywords  Normativity · Mindshaping · Niche construction · Rule · Culture

5.1  Introduction It is clear that various social mechanisms will act upon the minds of the members of any society; that every society generates a “cultural framework” which will have a bearing on any new member being introduced (“enculturated”) into the society. However, what is less clear is how deep this influence cuts, how the framework is produced, and how it wields its influence on newcomers. Naturalistically-minded researchers, until recently, have tended to downplay the influence of enculturation. We humans, they would stress, belong to the animal kingdom and are therefore biological entities; thus our nature is determined by the weighable and measurable factors of biology and genetics, rather than the elusive determinants of culture. Thus, in their classic article, Tooby and Cosmides (1992) criticized what they called the “Standard Social Science Model (SSSM)”, according J. Peregrin (*) Faculty of Philosophy, University of Hradec Králové, Hradec Králové, Czechia e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_5

85

86

J. Peregrin

to which the minds of members of a society emerge as “blank slates” (a term re-­ introduced into this context later by Pinker 2002) which get shaped by the “cultural framework” of the society to become its integral parts. They objected that any “cultural frameworks”, and how they actually influenced individual minds, remained so opaque that this whole picture cannot be taken too seriously. Since this paper was published there have been many developments, and it is now accepted that referring to “cultural frameworks” of this kind need not clash with naturalism, that it might be merely a way of emphasizing the specifically social facets of human animals. Authors, such as Boyd and Richerson (2005), Henrich (2015) and Heyes (2018), have shown how this specific kind of sociality leads to the establishment of a system which parallels the genetic evolution, a system of “culture” that gets handed down from generation to generation in a paragenetic way and which works because social newbies not only come to be immersed into it, but also gradually come to reproduce and upgrade it, providing for future newbies to be immersed into an upgraded version. In this paper I want to contribute to the understanding of the working of the “cultural framework”, arguing that its crucial components are norms and rules. (I use these two terms interchangeably in this paper.) Norms, I argue, are what we learn to be sensitive to during our “enculturation”; they create virtual spaces in which we spend great chunks of our lives after having become fully-fledged heirs of our culture, and they are also what we reproduce and upgrade after having been able not only to follow them, but also to understand, ponder and possibly challenge them. In short, I argue that to understand the working of the “cultural framework” and of both the way we produce it and the way it co-produces us we must understand that we humans are essentially normative creatures.

5.2  Mindshaping In his book, Zawidzki (2013) argues that the broad discussion in current anthropology of the emergence of the human (and possibly of other animals’) capacity of “mindreading”, which is often considered as crucial for our species, tends to ignore the important fact that what was going on during the stages of evolution where “mindreading” was to emerge and flourish was also the rise of what can be called “mindshaping”. According to Zawidzki, mindshaping amounts to “imitation, pedagogy, norm cognition and enforcement, and language-based regulative frameworks, like selfand group-constituting narratives”, its point being “making human minds and behavior more homogeneous and hence easier to predict and interpret” (p. 29). The mainstream story, Zawidzki points out, is that the evolution of us humans accelerated to set us on a track deviating from our animal cousins (developing our sophisticated language, creating culture etc.) all thanks to our increasing ability to read each other’s minds (caused, perhaps, by an increase in our brain size). But this story,

5  Normative Mindshaping and the Normative Niche

87

in Zawidzki’s view, neglects the essential role of our concurrent mindshaping abilities. As a first approximation to understanding the term “mindshaping” as I use it here, consider the following situation. A skilled chess player may be able to “read the mind” of her opponent in that she is able to predict, more or less successfully, his upcoming moves. This is achievable partly because there is only a limited spectrum of possibilities: the mind of the opponent has been formed, by being inculcated with the rules of chess, into a shape where he considers only the possibilities offered by the game. Similarly when she tries to “read the mind” of her peer, the task may be facilitated by the fact that there may also be a kind of game in play that limits the spectrum of thoughts that are available to the peer in the current social situation. Zawidzki’s point is not merely that the mainstream view pays scant attention to the phenomena accompanying the rise of human culture (such as pedagogy, enculturation, social norms etc.). His more crucial point is that such phenomena are being taken as just consequences of the improvement of mindreading, whereas, more fittingly, they should be seen as underlying this improvement. The most basic message of Zawidzki’s book is that our mindreading abilities, instead of underlying the mindshaping ones, in fact piggyback on them. We are good readers of each other’s minds not because we have developed big brains, but because we have managed to shape each other’s minds so that they have become easily readable. Thus Zawidzki writes (p. xii): I argue that the attribution of full-blown propositional attitudes cannot have evolved before sophisticated practices of mindshaping aimed at making us easily interpretable to each other. It is likely that sophisticated mindshaping coevolved in the human lineage with improved versions of sociocognitive capacities that we share with nonhuman primates, such as tracking the goals of conspecifics and anticipating the rationally and informationally constrained behavioral means they select to achieve them. However, as I argue, even highly sophisticated versions of such behavior tracking do not amount to the attribution of full-­ blown propositional attitudes. The capacity to attribute such mental states depends on, and had to await, the evolution of sophisticated mindshaping practices, especially linguistic practices like norm institution and narrative self- and group constitution.

For Zawidzki, mindshaping is closely connected with what has come to be called “niche construction” (Odling-Smee et al. 2003). A niche of an animal species is an environment in which the species lives, seen especially as a distributed collection of resources and dangers relevant for the species. However, some animals not only respond to the selective pressure of their environment, but are able also to modify their environment, thus also modifying the selection pressures to which they respond. This may launch a spiral that, under favorable circumstances, may become largely self-propelling. (Altering the environment can bend the trajectory in the direction of enhanced abilities to modify the environment, which, in turn, can bend the trajectory further …) Our human species is distinguished by the fact that aside of the physical niche construction we can be seen as engaged in the construction of our “social” niche, in fact of the “cultural framework” we mentioned in the beginning of this article. And

88

J. Peregrin

one of the ways to explain the peculiar acceleration of human evolution is precisely as the result of this kind of self-propelling. Zawidzki writes (p. 21): Distinctively human mindshaping is crucial to explaining the success of the hominid sociocognitive syndrome because it constitutes a way of bringing social niche construction under control: unlike other species, we obsessively engage in practices whose raison d’être is social niche construction. Unlike fortuitous niche construction that occurs as a by-product of traits selected for other reasons, human mindshaping enables targeted social niche construction. This is key to understanding mindshaping’s crucial role in the evolution of the human sociocognitive syndrome.

We may imagine that a species which has been selected, among other things, for its ability to avoid a certain kind of danger constantly present in its environment, manages to change the environment so as to dispense altogether with this particular danger. In this way, it influences its own evolutionary trajectory. Now this change may result in a further modification of the environment, effecting a further change of the selection pressures and a further deflection of the trajectory. And the idea is that this is what happened to us humans, when our evolution accelerated and achieved an “escape velocity”, enabling us to part ways with other kinds of animals. Take language. The mainstream story is that due to some biological development of our brains, we became capable of more effective and more articulated thinking than other animals, and this brought about the usefulness of communicating our thoughts (viz. language), which we hence developed. But an entirely different story is possible. At some point, the sounds we emitted became such an important part of our niche (helping us predict what will happen around us) that we not only came to be selected according to our sensitivity to it, but we also came to influence it, to construct our linguistic niche. Such a view of language emergence and development has been put forward, e.g., by Rouse (2015). He writes (pp. 119–120): Language … initially emerges not as the product of enhanced internal capacities of a larger hominid brain but instead as a perceptually salient, developmentally effective, and selectively important behavioral dimension of the developmental and selective environment of some hominid apes. Vocal expressiveness and its behavioral integration into a transformed way of life persisted as an integral part of these organisms’ ecological heritage only through its development and reproduction in each succeeding generation.

Imagine that many of our conspecifics came to emit a similar kind of sound when they detected a danger. Becoming sensitive to such displays of others would likely be useful; and, furthermore, it would be natural to work towards making the sound into a wholly reliable indicator of danger, getting other individuals to emit it always in cases of danger and not to emit it otherwise. Thus, it is not difficult to see that building this kind of regularity into our niche could be advantageous. (And, needless to say, it might be seen as a case of mindshaping.) In a similar way, many other uniformities of our communal life may be helpful, and benefits could be reaped by encouraging and producing them. Thus, linguistic, and more generally social, niche construction would tend to make the social landscape more homogenous and more perspicuous  – optimally more accommodating and more hospitable. This, obviously, can be achieved by

5  Normative Mindshaping and the Normative Niche

89

setting up standards and getting each other to comply with the standards. And we can do this because we have developed into essentially normative creatures – we are apt and diligent producers and consumers of norms, which thus become an important determinant of the world in which we live. (I discussed the details of how norms fuel cultural evolution elsewhere – see Peregrin 2014a.)

5.3  Cognitive Niche vs. Cultural Niche Pinker (2010) also talks about the human niche; and as he thinks that our specifically human way of life is primarily the result of the improvement of human cognition, he talks about a cognitive niche (8993–4): In biology, a “niche” is sometimes defined as “the role an organism occupies in an ecosystem.” The cognitive niche is a loose extension of this concept, based on the idea that in any ecosystem, the possibility exists for an organism to overtake other organisms’ fixed defenses by cause-and-effect reasoning and cooperative action  – to deploy information and inference, rather than particular features of physics and chemistry, to extract resources from other organisms in opposition to their adaptations to protect those resources. These inferences are played out internally in mental models of the world, governed by intuitive conceptions of physics, biology, and psychology, including the psychology of animals. It allows humans to invent tools, traps, and weapons, to extract poisons and drugs from other animals and plants, and to engage in coordinated action, for example, fanning out over a landscape to drive and concentrate game, in effect functioning like a huge superorganism.

This conception is criticized by many of those who think that it is not the improvement of cognition that came first and underlies our peculiar kind of human “cultural” sociality, but that rather it is the sociality, in the form of social niche construction, which boosted the cognition. Thus, Boyd et al. (2011) write (p. 10919): It seems likely that the average human is smarter than the average chimpanzee, at least in domains like planning, causal reasoning, and theory of mind. However, we do not think this is sufficient to explain our ecological success. The cognitive niche hypothesis overestimates the extent to which individual human cognitive abilities allow people to succeed in diverse environments and misunderstands the role that culture plays in a number of important ways. We suggest, instead, that our uniquely developed ability to learn from others is absolutely crucial for human ecological success. This capacity enables humans to gradually accumulate information across generations and develop well-adapted tools, beliefs, and practices that no individual could invent on their own. We have entered the “cultural niche,” and our exploitation of this niche has had a profound impact on the trajectory of human evolution.

Similarly Laland and O’Brien (2011, p. 191): Niche-constructing species play important ecological roles by creating habitats and resources used by other species and thereby affecting the flow of energy and matter through ecosystems – a process often referred to as “ecosystem engineering.” An important emphasis of niche construction theory (NCT) is that acquired characters play an evolutionary role through transforming selective environments. This is particularly relevant to human evolution, where our species has engaged in extensive environmental modification through cultural practices. Humans can construct developmental environments that feed back to affect how individuals learn and develop and the diseases to which they are exposed.

90

J. Peregrin

It is clear that niche construction can come in various forms, from very simple to highly complex. Rearranging simple elements of the environment (like stones or branches) is quite simple and can be done by animals without excessive sophistication. However, the more complex the elements are, and the more complex their rearrangement aims to be, the more sophistication and specific skills are required. (Dealing with plants, for example, is less easy than dealing with stones, for plants tend not to stay in the form in which we put them). And, needless to say, if the niche we are dealing with consists of other humans, the difficulties associated with its rearrangement bourgeon. Now the idea, in a nutshell, is that we humans have succeeded in mastering even this form of niche construction, that mindshaping is its principal tool, and that norms are the principal tool of mindshaping. You can move a stone and it stays where it is; you can pull up the weeds, but you have to repeat it regularly; and you can organize a society, but only by means of a kind of constant influence on its members. And the idea is that rules and norms can be seen precisely as the tools we humans have developed for this very purpose. This, I suggest, is what has led to the creation of our cultural niche alternative. Thanks to it, we humans are smart not only because of our genetic endowment, but also because of the niche, which we hand down from generation to generation. (This, of course, would not defy a naturalistic explanation – the only proviso would be the ability to cope with the sheer complexity of the corresponding naturalistic picture, taking into account the peculiarities of our species’ sociality which has allowed us to establish our cultures as frameworks which also provide for “cultural inheritance”.) However, it is crucial to explain how the perpetuation of the cultural niche works; and here again, I think, we must turn our attention to what I think is its more forerunning element – our ability to bring rules into being and to follow them. Therefore I propose we characterize our human niche as first and foremost a “normative niche”.

5.4  Rules and Normativity I have frequently pointed out that an extremely inspiring philosopher with regard to the normative dimension of human sociality and its perpetuation was Wilfrid Sellars (Peregrin 2010, 2011, 2014a). His analysis of the concept of rules, and of rule following, not only demonstrates the crucial role rules play within the human world, but can also be taken to indicate why and how rules appeared and spread in human communities. Sellars (1969) anatomized the dialectics of what he called “ought-to-­ be’s” and “ought-to-do’s”. Return, for an illustration, to the example of chess. Suppose Boris plays chess and his king is checked. In this situation Boris ought to move so that his king is no longer checked is an ought-to-do, a prescription addressed to Boris. Of course, it can do its work only if Boris is able to understand it; hence Boris must be a creature possessing the concepts out of which the ought-to-do is composed. On the other

5  Normative Mindshaping and the Normative Niche

91

hand, consider the general prescription The king ought not stay checked. This is a case of an ought-to-be and of course it does not presuppose that the king understands it. The prescription, obviously, is not addressed to the chess piece; it is, as it were, “free floating” and should be picked up by those agents who are able to use it to infer some ought-to-do’s addressed to them. This is the case, for example, of Boris: as The king ought not stay checked and his king is checked, he infers that He ought to move so that his king is no longer checked. Sellars is concerned with the workings of language: speakers of a language grasp certain ought-to-be’s concerning language in the form of certain patterns displayed by the ways words are handled. Because the speakers grasp the patterns as ought-to-­ be’s, they try to bring them into being (because ought-to-be’s entail the ought-to-­ do’s which aim at bringing about the ought-to-be’s) by getting the language novices to instantiate these regularities. Thus the linguistic behavior of the novices comes to instantiate the patterns in question.1 The same can be said about many systems of rules other than the linguistic ones: the point is that what we need in order for a rule to perpetuate is that the novices taught to follow the rule not only instantiate the pattern prescribed by a rule, but also grasp the pattern as an ought-to-be. If this grasping were somehow brought about by being squeezed into the pattern, we would have a virtuous circle which would propagate the rules from generation to generation: by becoming a follower of a rule one would become, in one sweep, a tutor of the rule. And as rules do seem to be perpetuated in this way, it seems that something like this must actually be happening. Let me stress how nontrivial this step is. Suppose that I am trying to scramble through some dense forest, where it is difficult to find any way through. Here, the forest is restraining me and it may be that it forces me into an almost unique path. I submit, and move the way which the forest permits; but certainly I do not join the forest in restraining others who want to get through it. Being restrained by nature is something one has to face; of course it is not something with which one feels any need to join forces. Compare this with what happens when I learn a language (or, for that matter, another social activity). I may try various ways of doing it (for example I may try to emit various kinds of sounds), but I get diverted, by my tutors, from many of them, and am left with only a few – with only a restricted path to move through the “forest of language”. But in this case, surprisingly, and unlike in the physical forest case, I come to mimic the activity of my mentors and I start to mentor other would-be speakers of the language. This is a crucial difference betraying that we come to distinguish the patterns in social reality from our patterns in the natural world; the difference is that we understand many such patterns in our social world as ought-to-be’s. My conjecture (and this is something which differentiates my view from those of other inferentialists, notably Brandom) is that the Sellarsian notion of ought-to-be points to the minimal element laying the foundation of any kind of normativity: the

 For a further anatomization of rules of language see Sellars (1949, 1954).

1

92

J. Peregrin

attitude of “holding as ought-to-be”, or “holding correct”.2 (This attitude can perhaps be seen as a generalization of that of “holding true” in Davidson 1990). This, we can say, is the only “unexplained explainer” of the theory – I simply assume we humans, in contrast to other animals, have come to acquire the ability to assume this kind of attitude.3 As a consequence, it is this specific kind of attitude which is the source of our culture and our very specific kind of sociality. From the viewpoint of Zawidzki’s notion of mindshaping as the container term for “imitation, pedagogy, norm cognition and enforcement, and language-based regulative frameworks, like self- and group-constituting narratives”, the thesis is that these disparate activities have a common denominator and that mindshaping, after all, is not merely a container notion. The idea is that while many animals certainly assess the behavior of their conspecifics in the sense that they try to divert them from certain ways and encourage them in others, their main concern is the resultant behavior of the conspecifics towards themselves. We humans, on the other hand, have become capable of assessing the behavior as such, largely independently of who, whom or what is its source and target. We raise ourselves above our parochial individual perspective and reach a more “impartial” stance, enabling us to see the behavior in question as independent of any particular source or target.4 My idea, presented in the papers mentioned above, is that the peculiar “operating in normative mode” is the ability to perceive the kind of social coercion which amounts to enculturation, as not only something that is to give me a direction, but rather something that ought to be, generally. In a recent empirical study (Schmidt et al. 2016), the fact that human infants do indeed come to acquire precisely this kind of stance is confirmed (p. 1360): Three-year-old children are promiscuous normativists. In other words, they spontaneously inferred the presence of social norms even when an adult had done nothing to indicate such a norm in either language or behavior. And children of this age even went so far as to enforce these self-inferred norms when third parties “broke” them. These results suggest that children do not just passively acquire social norms from adult behavior and instruction; rather, they have a natural and proactive tendency to go from “is” to “ought.” That is, children go from observed actions to prescribed actions and do not perceive them simply as

 See Peregrin (2014b, Chapter 4).  The list of primitive notions in the theory of Brandom as well as in those of some other inferentialists is more complex: especially it includes the concepts of commitment and entitlement. In contrast to this I think that these notions are reducible to the more primitive concept of holding correct: to be committed (to do something), for example, is to be held for committed, which, in turn, is to be required to carry out certain actions. This means that a person is committed to something iff certain actions of the person are taken to be correct. 4  The “impartiality”, of course, is not to be understood so that the correctness of an action must be assessed independently of other actions. An action might be correct as a successor or a predecessor of other actions. Thus, an action of a person may be correct (because of other actions of the person), while the same action carried out by a different person may be incorrect. The “impartiality” means that it is disregarded which concrete persons are behind the actions – especially whether it is the assessor or somebody else. 2 3

5  Normative Mindshaping and the Normative Niche

93

guidelines for their own behavior but rather as objective normative rules applying to everyone equally.

Seen from the phylogenetic, rather than ontogenetic perspective, Henrich (2015) diagnoses a very similar phenomenon (p. 188): Over our evolutionary history, the sanctions for norm violations and the rewards for norm compliance have driven a process of self-domestication that has endowed our species with a norm psychology that has several components. First, to more effectively acquire the local norms, humans intuitively assume that the social world is rule governed, even if they don’t yet know the rules. … Second, when we learn norms we, at least partially, internalize them as goals in themselves. This internalization helps us navigate the social world more effectively and avoid temptations to break the rules to obtain immediate benefits.

5.5  Acting as Persons It is important to realize that from our human viewpoint, norms do not merely make up a scaffolding that helps us read each other’s minds in the sense of estimating their future behavior, consequently better coordinating with each other and thus establishing a system of “cultural inheritance”. Rather, it is the normative scaffolding itself that licenses us to live in a world that we perceive as “meaningful”, in a world where there are not only moving bodies causing various effects, but also persons acting for reasons to achieve goals. The basic idea is that a system of norms, if put together in a suitable way, may constitute a space in which we can carry out new actions unheard of before. (And there is a sense of the term action in which it is only within such a normative space that we can carry out actions at all.) Hence, the framework of the rules of chess allows us to check the opponent’s king, that of the rules of language allow us to assert that the sun is shining, while the framework of rules making up a university allows us to pass an exam. This idea was tabled by Brandom (1979) and it leads to what Steiner and Stewart (2009, p. 530) call heteronomy: Becoming socialised is achieved by becoming heteronomous: it involves knowing that the behaviours one produces have to be performed in a certain way, and acting accordingly. Abiding by norms is a relational property of agents: it depends on the existence of these norms independently of the agent (this existence consists in their following and practical acknowledgement by a community of agents), and on the fact that the performances of the agent are recognised by other agents as being sensitive (and not randomly conforming) to these norms. Unlike natural laws or biological norms, it is quite possible for an agent to behave in a way that does not respect a social norm; the sanction is no more, and no less, than that the behaviour in question will not be recognised as a socially meaningful and appropriate action.

Thus, various kinds of frameworks of rules have come to take part in constituting the world in which we live: they are our niche into which we are born and, later in our lives, we are destined to contribute to its reweaving. This brings us back to the mechanism allowing us produce the “cultural framework” which then co-produces us. Once we recognize that the “cultural framework”

94

J. Peregrin

crucially involves a system of frameworks of rules, we can easily see how this works. We have become able to “hold” certain kinds of behavior in certain circumstances “as correct”, while others “as incorrect”. Once these “normative attitudes” came to resonate throughout a community, there appeared something akin to rudimentary “implicit rules”: certain ways of behavior came to be (taken as) generally correct or appropriate within the community. The rudimentary rules that were established by such attitudes were then, on the one hand, fortified and elaborated, while, on the other hand, becoming composed into frameworks that constituted rudimentary communal institutions. The most important of such institutions was language: here a framework of rules provided for the possibility of making various kinds of meaningful utterances. Once language was in place, the whole enterprise of establishing, maintaining and abiding by the normative institutions changed its character – it became possible to make the rules explicit, to constitute new rules by means of explicit stipulation, and, on a still more advanced level, to discuss the pros and cons of various rules. Many institutions we take for granted would be inconceivable without the support of language. There is nothing mysterious about the normative frameworks and institutions established in this way. At the same time, there is little doubt that we do live in a system of such frameworks and institutions – so many things we do in our lives are not merely physically or biologically characterizable pieces of behavior, they are rather actions which can be carried out only within their normative context. Hence the “cultural framework” understood in this way is indeed something quite crucial, and yet nothing that would be suspicious from the viewpoint of science or philosophy. Learning to live within this system of normative frameworks and institutions, which one does during the process of education and enculturation, is a kind of mindshaping. Undergoing mindshaping is learning to respect not only the limits of one’s world as posed by nature, but also the limits as posed by other people and the wider society (the ought-to-be’s). But becoming not only a “mindshapee”, but also a “mindshaper” (which is ultimately part and parcel of undergoing mindshaping) one must learn to distinguish between the two kinds of limits (to take the ought-to-be’s for what they are, namely something that yields us ought-to-do’s to be followed). It is only the latter, the limits posed by social reality, which have the qualities of a rule, viz. something to which one must not only yield, but which one must take part in supporting.5

 At first, it would seem the apprentice may perceive all kinds of limits to her world as on a par: the fact that she cannot as on a par with the fact that she may not. It is only later (perhaps at about 3 years of age, as Schmidt et al. 2016, suggest) that she comes to distinguish between the “hard” cannot and the “soft” may not - the latter being “soft” not only in that there are ways of violating it, but also in that to hold, it needs the support of people, including the apprentice. Also, she becomes more sensitive to the often quite mild “social friction” that marks the presence of a rule. And, as a culmination of the process, she may come to reflect on the rules constituting the “cultural framework” and their appropriateness. 5

5  Normative Mindshaping and the Normative Niche

95

5.6  Normative Niche and Collective Intentionality Some philosophers and scientists have proposed that the ultimate trigger of our specifically human evolutionary trajectory leading us to our self-propelling cultural niche, was the emergence of the ability of shared or collective intentionality (thus, e.g., Tomasello et  al. 2005, claim that it is the ability to share intentionality that constitutes “the crucial difference between human cognition and that of other species”). To what extent is the proposal discussed here compatible with this claim? It is quite clear that once a human community fastens down its “normative niche” (i.e. once its members come to spend most of their lives within a system of normative frameworks), we can account for this in terms of collective intentionality – it is clear that the rules are “commonly accepted” by the community (in a sense which goes beyond mere concurrent individual acceptances). The question, however, is whether such a “common intention” must be one of the very basic building blocks of the whole edifice; and whether it is the only (or at least the most basic) building block. Tomasello et al. (2012, pp. 673–4) write: Our hypothesis, which we call the Interdependence Hypothesis, is that at some point humans created lifeways in which collaborating with others was necessary for survival and procreation (and cheating was controlled by partner choice). This situation of interdependence led inevitably to altruism, as individuals naturally wanted to help the collaborative partners on whom they depended for, for example, foraging success. Moreover, interdependent collaboration also helps to explain humans’ unique forms of cognition and social organization, since it is collaboration, not altruism, that creates the many coordination problems that arise as individuals attempt to put their heads together in acts of shared intentionality to create and maintain the complex technologies, symbol systems, and cultural institutions of modern human societies.

This proposal is meant as an alternative to the “altruism first” views, proposing that it was altruism (yielded by the kin selection mechanism or sustained by reciprocation) that founded cooperation; Tomasello and his collaborators propose that cooperation developed in its own way, not presupposing (but possibly yielding) altruism, and that it was shared intentionality that played a crucial role in this. Tomasello (2014) writes that while the intelligence of primates is utterly “Machiavellian”, i.e. oriented towards “competition or exploitation of others”, our human intelligence aims at “cooperation or communication with others”. This may be true (and there is, to be sure, a sense in which it is obviously true); and it is also quite clear, from the empirical studies, that shared intentionality is something that distinguishes us from the primates and that must have played a role in our development. The question, however, as I see it, is whether it was really shared intentionality that formed the bridge from individualistic to cooperative intelligence. Perhaps there is also another possibility to be taken into account. The view put forward here indicates that there is also an alternative to the proposal that it was shared intentionality that made us cross the Rubicon separating the “individualists” from the “cooperators”. This alternative possibility is that the crucial point was the specific kind of attention that we started to pay to each others’

96

J. Peregrin

behavior, the peculiar way in which we started to practically discriminate between behaviors that are “correct” or “incorrect” not in respect of their influence directly on us as individuals, but from an “impartial standpoint”. It was behavior as such, not its impact on the assessor that started to be seen as desirable or undesirable. I think that reaching an “impartial standpoint” (or “agent-neutral thinking”, as Tomasello 2014, would call it) may be seen not as a result of overcoming the “Machiavellian” stance, but as its continuation (if not culmination). It seems to me that in their effort to exploit everything around them including their conspecifics, our ancestors might have come to the conclusion that it is profitable to concentrate on the usefulness of actions independently of who or whom is their source or target. Let us return to our example of the tribe whose members tend to emit a certain sound when they are in danger. Sounds of this kind may become a part of their niche as accompanying certain kind of situations. And it would certainly foster the survival of the members of the tribe if at least some of the sounds were generally reliable indicators of the situations; which may lead, under certain circumstances, to the members coercing each other to emit them always only in the appropriate situations. This requires emitting the sounds “appropriately” independently of who is the emitter. And this, it seems to me, can still be an outcome of the “Machiavellian” stance, though this is also a clear rudiment of the normative attitudes that, according to me, in their more mature form build up rules. However, it is clear that the ability to assume normative attitudes towards actions is not enough to produce the fully-fledged normative niche. The attitudes must, as I put it, come to resonate throughout the community. Does this mean that it is enough that they just coincide, or do they have to be assumed collectively in some deeper sense, presupposing shared or collective intentionality? I have already stated that I take for granted the fact that shared intentionality is a distinctively human innovation which plays a substantial role in the way we humans co-exist and cooperate; I take it to be an empirically ascertained fact, not warranting challenge from a philosophical ivory tower. Let me just point out that the theory I am proposing does not presuppose this in any obvious way. We can very well imagine that rules evolve from the mere resonance of normative attitudes via their reflective coordination to a collective sustainment of the rules, which is however a product of the coordination of the normative attitudes, rather than their presupposition. In a sense, the problem might be seen as that concerning the relative primitiveness of cooperation and coercion with respect to each other. One possibility is that the emergence of shared intentionality makes us switch from the coercion dictated to us by our “Machiavellian” intelligence to a brand new kind of intelligence that is essentially cooperative; another possibility is that the cooperation emerges directly from coercion, once it turns out that the coercion might be especially effective when wielded from an “impartial standpoint”. The latter option conjectures that what comes out of the selective pressures of the environment is the tendency to control one’s conspecifics in such a way as to force them into the emerging mold of what is correct and what is not.

5  Normative Mindshaping and the Normative Niche

97

5.7  Conclusion We, humans, live in largely a normative world; or perhaps in a motley of normative micro-worlds. The introduction of infants into a human community consists, to a large part, in teaching them how to recognize, to respect, and also to support the rules that make up our common “virtual” worlds. Our ability to “operate in the normative mode” provides for a suitable engine on which these worlds “run”: they enable us to become the inhabitants of the worlds as well as their wardens in one sweep. This account, I think, explains the self-perpetuating character of our culture. The normative niche tends to shape the minds of all novices joining the society; they become sensitive to the kind of social friction that indicates the presence of rules, thus learning to live within the system of normative frameworks that constitute it. They learn to respect the norms, and at the same time also to sustain them. In this way enculturation produces not only subjects conforming to the norms, but also those enforcing them, which guarantees its continuation. The enculturation, due to the specific character of the normative niche, also produces persons, who carry out various kinds of actions made possible by the individual normative frameworks. In particular, it produces discursive persons or speakers, who enter the most important normative framework, the framework of a natural language, viz. the framework providing for the possibility of carrying out meaningful utterances. Entering this framework completely rebuilds our ability to operate within other frameworks and especially to create and sustain new frameworks. Acknowledgements  Work on this chapter was supported by the joint Lead-Agency research grant between the Austrian Science Foundation (FWF) and the Czech Science Foundation (GAČR), Inferentialism and Collective Intentionality, GF17-33808 L. I am grateful to Javier Gonzáles de Prado Salas, Jeremy Koons, Ladislav Koreň, Glenda Satne, Vladimír Svoboda, and Preston Stovall for valuable critical comments.

Bibliography Boyd, R., & Richerson, P. J. (2005). Not by genes alone. Chicago: University of Chicago Press. Boyd, R., Richerson, P. J., & Henrich, J. (2011). The cultural niche. PNAS, 108, 10918–10925. Brandom, R. (1979). Freedom and constraint by norms. American Philosophical Quarterly, 16, 187–196. Davidson, D. (1990). Representation and interpretation. In W. H. Smith & K. V. Wilkes (Eds.), Modeling the mind (pp. 13–26). Oxford: Clarendon Press. Reprinted in Davidson, D. (2004). Problems of rationality. Clarendon Press, Oxford, pp. 87–100. Henrich, J. (2015). The secret of our success: How culture is driving human evolution, domesticating our species, and making us smarter. Princeton: Princeton University Press. Heyes, C. (2018). Cognitive gadgets: The cultural evolution of thinking. Cambridge, MA: Harvard University Press.

98

J. Peregrin

Laland, K. N., & O’Brien, M. J. (2011). Cultural niche construction: An introduction. Biological Theory, 6, 191–202. Odling-Smee, F.  J., Laland, K.  L., & Feldman, M.  W. (2003). Niche construction. Princeton: Princeton University Press. Peregrin, J. (2010). The enigma of rules. International Journal of Philosophical Studies, 18, 377–394. Peregrin, J. (2011). Creatures of norms as uncanny niche constructors. In T. Hříbek & J. Hvorecký (Eds.), Knowledge, value, evolution (pp. 189–198). London: College Publications. Peregrin, J. (2014a). Rules as the impetus of cultural evolution. Topoi, 33, 531–545. Peregrin, J. (2014b). Inferentialism: Why rules matter. Basingstoke: Palgrave. Pinker, S. (2002). The blank slate. London: Penguin. Pinker, S. (2010). The cognitive niche: Coevolution of intelligence, sociality, and language. PNAS, 107, 8993–8999. Rouse, J. (2015). Articulating the world. Chicago: University of Chicago Press. Schmidt, M. F., Butler, L. P., Heinz, J., & Tomasello, M. (2016). Young children see a single action and infer a social norm: Promiscuous normativity in 3-year-olds. Psychological Science, 27, 1360–1370. Sellars, W. (1949). Language, rules and behavior. In S. Hook (Ed.), John Dewey: Philosopher of science and freedom (pp. 289–315). New York: Dial Press. Sellars, W. (1954). Some reflections on language games. Philosophy of Science, 21, 204–228. Sellars, W. (1969). Language as thought and as communication. Philosophy and Phenomenological Research, 29, 506–527. Steiner, P., & Stewart, J. (2009). From autonomy to heteronomy (and back): The enaction of social life. Phenomenology and Cognitive Science, 8, 527–550. Tomasello, M. (2014). A natural history of human thinking. Cambridge, MA: Harvard University Press. Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences, 28, 675–690. Tomasello, M., Melis, A. P., Tennie, C., Wyman, E., & Herrmann, E. (2012). Two key steps in the evolution of human cooperation. Current Anthropology, 53, 673–692. Tooby, J., & Cosmides, L. (1992). The psychological foundations of culture. In J.  Barkow, L. Cosmides, & J. Toby (Eds.), The adapted mind: Evolutionary psychology and the generation of culture (pp. 19–135). New York: Oxford University Press. Zawidzki, T. W. (2013). Mindshaping: A new framework for understanding human social cognition. Cambridge, MA: MIT Press.

Chapter 6

Collective Intentionality, Inferentialism and the Capacity for Claim-Making Glenda Satne

Abstract  Some of our linguistic practices are special in that they involve claims about how things stand in the world. These judgments are thought to be true or false with respect to what they are about. The nature of these practices of claim-making has been studied by philosophers and psychologists alike. Furthermore, important strands in evolutionary studies have relied on both psychological and philosophical theories for addressing the question of how these practices evolved in the hominins lineage, claiming that engagement in such linguistic practices is humans’ exclusive heritage. This paper aims to show that (some of) the conceptual tools provided by collective intentionality and inferentialist theorists can productively complement each other in pursuing this theoretical endeavor, namely the elucidation of human practices of claim-making. In particular, I argue that (i) Brandom’s inferentialist account of basic linguistic practices in interpretational terms is problematic and that these problems can be addressed via appeal to the collective intentionality theorist’s toolbox, and (ii) Inferentialists resource to linguistic norms in understanding meaning and claim-­ making can offer crucial tools for understanding the emergence of human specific cognitive capacities of objective judgement, tools that are needed to complement the accounts offered by collective intentionality theorists. The main aim of the paper is to provide an account of the evolution and development of human-specific abilities of claim-making which combines resources from both approaches in order to understand the nature and crucial role of shared activities in their emergence. Keywords  Collective intentionality · Inferentialism · Claim-making practices

G. Satne (*) Philosophy, School of Humanities and Social Inquiry, Faculty of Arts, Social Sciences and Humanities, University of Wollongong, Wollongong, NSW, Australia e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_6

99

100

G. Satne

6.1  Introduction Some of our linguistic practices are special in that they involve claims about how things stand in the world. These judgments are thought to be true or false with respect to what they are about. The nature of these practices of claim-making has been studied by philosophers and psychologists alike, asking questions such as: How are these practices best understood conceptually? What are the cognitive mechanisms that underlie the capacity to make and assess this kind of judgments? Furthermore, important strands in evolutionary studies have relied on both psychological and philosophical theories for addressing the question of how these practices evolved in the hominins’ lineage, claiming that engagement in such linguistic practices is humans’ exclusive heritage. In the philosophical domain, social claim-making practices have been at the center of what prima facie might seem two very different domains of research. On the one hand, some philosophers interested in the question of how to understand the sort of meaning that objective judgments have, argue that meaning is to be understood as instituted within social practices. These theories attempt to describe conceptually the central features of such practices. Following the suggestion put forward by the editors of this book, I will call these authors ‘inferentialists’. Inferentialism is a theory of meaning. Its advocates seek to explain meaning in terms of its institution through normative practices of giving and asking for reasons that are inferentially articulated. The view is inspired by Sellars (1956) and Brandom (1994) (see also Kukla and Lance 2009; Peregrin 2012 for views along the same lines). Its central thesis is that meaning emerges in the context of a normative social game, in which social rules, roles, statuses, etc. are instituted and enacted through the mutual attribution of normative statuses articulated in terms of reasons. According to inferentialism, the meaning of words and sentences—and by extension mental and other symbolic items—are given by the inferential roles that linguistic items acquire in the social game of giving and asking for reasons. On the other hand, theories of collective intentionality that focus on joint action and shared intentionality have also attempted to understand the nature of the social practices at issue. One important application of collective intentionality theories concerns precisely the explanation of the evolution of social understanding and even the emergence of human capacities for objective thinking (Tomasello 2014). A promising idea behind this approach is the ‘Cooperative Evolutionary Hypothesis’ (CEH), namely, the idea that humans’ capacity for social cooperation is at the heart of their ability to understand others’ mental states and behavior, leading to an explanation of how humans came to share thoughts and language (see e.g.  Satne and Salice 2020; Sterelny 2012). Collective intentionality theories seek to shed light on a set of distinctive collective phenomena from group action to group feeling and thinking. According to these views, intentionally doing something together with other agents involves joint or shared intentions which do not straightforwardly reduce to the individual intentions and goal-directed actions of the individual agents making up the group (Schmid 2009; Schweikard and Schmid 2013). Following Schweikard and Schmid (2013),

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

101

the different positions within this domain can be classified into three views: the Content view, the Mode view and the Subject view. The Content view holds that what is collective in collective intentionality is a kind of content. The central idea behind this approach is that the contents of the individuals’ mental states overlap and interlock in ways that are sufficient for joint action. Bratman (1992, 2013) and Butterfill (2012) are among the main representatives of this view. The Mode view, whose main inspiration is the work of Searle (1990), explains collective action as a special, plural kind of intentionality. It distinguishes between an individual mode of intentionality and a collective mode, a distinction that applies to intentional states across the board. Finally, the Subject view stresses the role of the subject, claiming that what makes an action or a mental state collective is that it is endorsed, entertained or carried out by a plural subject. Both Schmid’s (2009) and Gilbert’s (2014) are views of this kind. Despite the differences between the three approaches, they all agree that shared intentionality is key to understand the nature of social roles, statuses, rules, institutions and practices. Thus, while both inferentialism and theories of collective intentionality, in their different flavors, converge in giving pride of place to the role of shared activities in understanding social reality in general and practices of claim-making in particular, they diverge remarkably in their approaches. They are different not only in their aims but also in the theoretical strategies and conceptual tools they use for pursuing them. Furthermore, both collective intentionality theories and inferentialist ones have been called on in accounting for the evolution of social understanding and human specific cognitive abilities of claim-making. From the collective intentionality perspective, Tomasello (2008, 2014), Tomasello et al. (2005), Butterfill (2012), Pacherie (2013) and Tollefsen and Dale (2012) offer accounts of the evolution of human specific forms of cognition. From the inferentialist perspective, Hutto and Satne (2015), Peregrin (2014) and Rouse (2015), have recently attempted to explain the emergence of uniquely human complex cognitive capacities of objective thinking in terms of normative social practices. In this respect, inferentialism and theories of collective intentionality can be seen as rival theories, as different ways of cashing out the idea that social interaction plays a central role in explaining the emergence of human-specific cognitive abilities, including the ability to think and talk objectively. This paper aims to show that (some of the) conceptual tools provided by collective intentionality and inferentialist theorists can productively complement one another in pursuing this theoretical endeavor, namely the elucidation of the sort of social practices that can be thought to lie at the heart of human practices of claim-­ making. In particular, I argue (i) that Brandom’s inferentialist account of basic linguistic practices in interpretational terms is problematic and that these problems can be addressed via appeal to the collective intentionality theorist’s toolbox, and (ii) that Inferentialists resource to linguistic norms in understanding meaning and claim-making can offer crucial tools for understanding the emergence of human specific cognitive capacities of objective judgement, that can complement the accounts offered by collective intentionality theorists. The main aim of the paper is to provide an account of the evolution and development of human-specific abilities

102

G. Satne

of claim-making which combines resources from both approaches in order to understand their nature and the crucial role of shared activities in their emergence. The second section of the paper presents and discusses inferentialist and collective intentionality views vis-à-vis the challenge of providing an account of ‘minimal or basic’ linguistic practices, the kinds that can be thought to be at the basis of our ability for claim-making. The third section puts forward an alternative view of collective intentionality—dubbed ‘evolving collective intentionality’—that combines theoretical resources from both approaches and outlines the developmental and evolutionary trajectories of human specific capacities making use of such an enlarged toolbox. This section illustrates these trajectories with empirical literature from developmental psychology and evolutionary anthropology.

6.2  Inferentialism vs. Collective Intentionality Collective intentionality theories underscore the role of shared goals in joint action. Inferentialism, on the other hand, elucidates rational agents’ linguistic performances and thoughts—both their contents and pragmatic force—in terms of practices of mutual attribution of normative statuses, what Brandom (1994) calls practices of ‘deontic scorekeeping’. Brandom (1994) advances two important claims that can be thought to straightforwardly conflict with some of the core ideas advocated by collective intentionality theorists. First, he argues that we should not give an account of social practices in terms of individuals’ contentful states, prior to engaging in such practices. All collective intentionality theories, in their three different flavors, the Subject, Content and Mode view, employ contentful intentional states to account for shared intentionality and are thus committed to the theoretical move that Brandom is criticizing here. The contention against this move is that social practices are required in the first place to make sense of the contents of intentional states. The argument runs as follows. If content is required in the first place in order to interact socially then content should be possible before and prior to engaging in social practices. Brandom follows Wittgenstein’s famous rule-following discussion to criticize this view, claiming that involvement in social practices is required to make sense of the conditions for the correct application of such contents for otherwise “seeming to be right is the same as being right” (Wittgenstein 1953: 202/253). If contents were possible prior to our involvement in social practices, then no conditions of satisfaction/appropriateness would apply to those contents. Yet, if they lacked such conditions, they could not be considered contents at all, for if everything “can be made to accord” (Wittgenstein 1953: 201) with a content, “there would neither be accord nor conflict here” (ibid), and thus no determinate content assigned to the linguistic or mental acts at issue.1 Moreover, and even more important for us in this context, to assume

 For a detailed discussion of these arguments, see Wittgenstein (1953: 186–242) and Brandom (1994: 20ff). 1

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

103

that contents are there prior to social interaction forecloses the possibility of giving an account of content in terms of social interaction. If having contentful states is needed for social interactions to take place, then surely an account of joint action cannot help with the task of explaining how contents are possible in the first place. Second, Brandom argues that it is a mistake to account for social normative practices taking as a starting point the perspective of a ‘we’ or plural subject. This places Brandom’s view against the Mode and Subject approaches to collective intentionality. Brandom (1994: 38–39) points out that the ‘we’ approach to social norms and social statuses, mystifies the nature of the authority of norms by attributing authority to what ‘the community’ takes as correct, thus modeling this authority after that of a ‘super-person’. Brandom claims that ‘being a member of a community’ and ‘following communitarian norms’ are normative notions that have to be elucidated themselves through an intersubjective account of normativity. Brandom proposes instead to understand social practices and meaning on the basis of normative attitudes that institute the kind of norms that make both contents and we-perspectives possible: i.e. practices of intersubjective normative scorekeeping. Brandom follows Davidson in that an elucidation of contentful practices requires adopting the point of view of an interpreter (Davidson 1973; see also Brandom 2010: 33–4). According to both Brandom and Davidson, we cannot make sense of meaning without acknowledging the perspectival character of meaning-­ attribution within human social practices. It is only through the eyes of the interpreter that we can make sense of the contentful practices in which human beings are engaged. This is what Brandom (1994: 37-2, esp. 39) calls the ‘I-Thou’ relation. The I-We relation can only be understood as derivative from I-Thou interpretational relations between individuals. Thus, according to Brandom, our practice of understanding each other is not dependent on “the existence of any form of preexistent, determined, ‘internalised’ agreement” (Malpas 2011: 261). Brandom makes clear that the interpretative stance that normative practices require can be thought to be essentially equivalent to that of an external observer of the practice: […] this sort of external interpretive stance – what one must do, how one must treat an alien community in order thereby to count as taking them to be making assertions and inferences – is seen to be equivalent to an internal scorekeeping stance within a discursive community. That is, one must adopt toward the practitioners […] the same sort of attitude one both takes them to adopt towards each other and adopts towards one’s own discursive fellows […] In short, the stance in question is a translational-interpretive stance. (Brandom 2010, 33–34, my emphasis; see also Brandom 1994, 659, n 50)

There is much to agree with in the view that Brandom puts forward. The idea that we should not start from fully contentful states if we want to account for the phylogenetic origins of human cognitive capacities has independent grounds in evolutionary theory and is prima facie congenial with Brandom’s inferentialist ideas, despite the different aims at stake. We can see this by comparing it to one of the main developmental/evolutionary collective intentionality accounts of human specific cognitive abilities belonging to the Content view, the view offered by Tomasello (2014), who takes inspiration from Bratman’s (1992) Content view of shared intentionality.

104

G. Satne

6.2.1  Tomasello’s Content View of Shared Intentionality Tomasello’s work represents a very popular strand in developmental and evolutionary studies of human cognitive abilities. He agrees with inferentialists that [Thinking] is a solitary activity all right, but on an instrument made by others for that general purpose, after years of playing with and learning from other practitioners […]. Human thinking is individual improvisation enmeshed in a sociocultural matrix. (Tomasello 2014: 1)

When it comes to account for the evolutionary origins of such a capacity, Tomasello relies on the collective intentionality theorists’ toolbox. Tomasello (2014) distinguishes two forms of social intentionality that correspond to a two-step evolutionary/developmental story. Both are forms of interaction and collaboration that are said to have paved the way for the emergence of human-specific forms of cognition, including the capacities for claim making and objective thinking. First, there is joint intentionality. This form of intentionality characterizes short-­ lived spatiotemporally located second-personal interactions between particular individuals. These are small-scale collaborations, triadic interactions in which the individuals involved share goals and jointly attend to situations and objects in the environment (Tomasello 2014: 48). According to Tomasello, this form of intentionality does not involve any grasp of objectivity in the sense of universal validity on the part of the interacting agents. Rather, individuals involved in these kinds of collaborations have a perspectival understanding of the world that encompasses a ‘sense’ or understanding of other perspectives towards the same objects. This capacity is claimed to have emerged in the hominin’s lineage and to characterize Homo Heidelbergensis’ social exchanges (Tomasello 2014). It is also characteristic of young infants’ social interactions from around 14 months onwards (Warneken and Tomasello 2007). Second, there is collective intentionality. This kind of intentionality is at play when agents engage in large-scale collaboration. This form of engagement between agents goes beyond the here and now characteristic of joint intentionality. It involves the establishing—or the background presence of an already established—cultural common ground (e.g., institutions, conventions, and so on). This kind of intentionality, according to Tomasello, makes an “agent-neutral point of view” available to the agent (Tomasello 2014: 5/122). This amounts to the point of view of a universal group, which, in this framework, equates to the ability to grasp an objectively valid point of view. This capacity is characteristic of Homo Sapiens (Tomasello 2014), and develops in young children at around 3 years of age (Rakoczy et al. 2008). Thus, on Tomasello’s view, the notion of ‘joint intentionality’ describes the set of abilities that an organism can possess without yet deploying capacities for the manipulation of truth-functional representations of the world and others’ mental states that underpin claim-making practices. It comprises mechanisms for social cooperation that enable individual learning and work in conjunction with mechanisms for the social inheritance of culturally evolved devices to construe a cultural common ground. When it comes to giving an account of joint intentionality,

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

105

Tomasello refers to the work of Bratman (1992). Joint cooperative activities proper to joint intentionality are defined in terms of the satisfaction of Bratmanian conditions for joint intentional action (Tomasello 2014: 38ff). Accordingly, Tomasello states that three conditions for joint intentional action are sufficient for joint intentionality: If you and I are agents, and J is a goal, then: I must have the goal of doing J together with you, you must have the goal of doing J together with me, and we must have “mutual knowledge, or common ground, that we both know each other’s goals”. (Tomasello 2014: 38)

Condition (3) involves meta-representations of other people’s mental states, and hence higher order attitudes. It is to be noted that not every interaction of this kind will involve meta-representations, because there may be sufficient information in the situation available for the agents to know that they both want to J. In such cases, the agents know their mutual goals because they stand on ‘common ground.’ Nevertheless, Tomasello argues that a capacity for recursive mindreading needs to be presupposed on the part of the agents to make sense of the very possibility of them engaging in this kind of interaction. He argues that in conflicting situations, we draw back to recursive reasoning, showing that our capacity of knowing that we both want to J has “an underlying recursive structure” (Tomasello 2014: 38; also 2008). Thus, a suite of mechanisms for understanding the perspective of another agent towards the common goal and the situation need to be presupposed. Tomasello further argues that understanding the perspective of a partner in joint action requires simulating her abductive inferences ahead of time in order to anticipate how one might be understood by the other (Tomasello 2014: 94). Thus, agents capable of this sort of interaction have cognitive capacities that underpin recursive inferences, simulations, ongoing self-monitoring and meta-representations of oneself and others (Tomasello 2014: 5, 9, 143). In this view, social cognition is part and parcel of joint intentionality involving both simulation theory (ST) and theory-theory (TT) of mind. As Tomasello explicitly acknowledges, his conception of shared intentional action is inspired by Bratman’s Content view (Tomasello 2014: 38ff). Despite this, it should be noted that Bratman’s aim is not to give an account of the emergence of human cognitive capacities, so even if this is not a consequence in his view, in Tomasello’s theory, joint intentionality of the Bratmanian kind is possible both for ‘early humans’ -Homo Heidelbergensis (Tomasello 2014)—and for children at around 18–24 months (Warneken and Tomasello 2007). It follows that we need to attribute to the child capable of participating in joint action an understanding of the concept of intention as applied to herself and others as well as common knowledge of their own intentions and of those of others. This involves the ability to grasp and deploy higher-order nested intentional attitudes. But, as many have argued (cfr. Tollefsen 2005; Brownell 2011; Michael et al. 2014; Zahavi and Satne 2015; Satne and Salice 2020), Bratman’s conditions for shared intentional activity are too cognitively demanding for young children to be able to meet them, not only because children engage in this sort of activity quite early  in their lives but, more

106

G. Satne

importantly, because engaging in some shared activities and joint actions may be essential for acquiring such sophisticated knowledge of other persons as agents. Thus, Tomasello’s attempt to give an account of the emergence of human specific cognitive abilities in terms of Bratman’s content view of shared intentional activities proves that the latter is ill-suited for the task. Might we alternatively explain early stages in the development and evolution of human cognitive capacities in terms of the inferentialist’s account of normative scorekeeping? It is important to note that, as in Bratman’s case, Brandom’s aim is not to give an account of the emergence of human cognitive abilities of claim-making. Instead, he offers an account of what he calls ‘Basic Linguistic Practices’ that is supposed to give minimally sufficient conditions for a practice to be understood as involving the exchange of linguistic items that are objectively contentful (see Brandom 1994, esp. ch. 1 and 3). Yet, taken as a conceptual elucidation of sufficient conditions for a practice to involve claims that are objectively valid, Brandom’s account can be thought to provide a theoretical model for minimal linguistic practices of the kind that can be thought to be at issue in understanding the development of human practices of claim-making. As it will be argued in what follows, significant problems haunt Brandom’s account of normative scorekeeping practices making it at the end unsuited for this task.

6.2.2  B  random’s Inferentialist Account of ‘Basic Linguistic Practices’ As illustrated by the quote above, Brandom rejects the idea that there is a distinction between being in conversation with someone (an internal scorekeeping stance within a discursive community) and adopting an external interpretational stance towards someone (the one someone adopts when observing a conversation from a distance). Nevertheless, by conflating both dimensions, his theory depreciates the relevance of a different understanding of interaction, one that may seem to be essential for making sense of the possibility of linguistic practices altogether, the ones at stake in the developmental and evolutionary debates on the origins of human specific cognitive capacities. Let’s see why. Habermas cast the first stone in this debate. Against Brandom’s picture, Habermas (2000) claimed that if we think of the interpreter’s stance from such an externalist perspective, we lose the idea of a language as being a way in which individuals engage in the pursuit of common goals and values. Brandom (2000) responds by claiming that, according to an attributional conception like his own, one can actually engage in social linguistic practices without pursuing common goals or sharing values. According to Brandom, common goals are not necessary to make sense of the possibility of there being linguistic practices altogether (Brandom 2000, 362). Yet, the real problem in my view, is one that Habermas indirectly points to but does not develop in detail: namely, that, in lacking such a dimension of sharedness—one that is defined by sharing goals, not just the goal of making interpretation

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

107

possible—normative practices turn out to be insufficient to build up a ‘common’, shared practice. This being the case, such practices turn out to miss the mark as a conceptual elucidation of the conditions that are sufficient for a practice to be a ‘basic linguistic practice’, i.e. a practice that is minimally sufficient for objective contents to be in place. This is because, as I argue below, the Wittgensteinian condition of adequacy of making a distinction between ‘seeming to be right’ and ‘being right’ for individual’s doings, that Brandom points to as crucial for meanings to be in place, is not met. Thus, objective normativity is not achieved. The heart of the problem is that in Brandom’s view the ‘attribution (of normative statues)’, the social glue of the game of giving and asking for reasons, is an individual endeavor, not a joint one. Each participant’s moves in the game are beholden to the individual interpreters’ assessments of their significance. This being the case, nothing guarantees that agreement between participants on relevant facts is going to be achieved. The problem seems to be that in a practice like the one that Brandom describes the reasons of each practitioner do not count as reasons for the others by their own sake. It is only when one participant’s reasons are already counted as justified according to the other participants that these reasons are to be taken seriously. Being reasons dependent on individual’s interpretations of their contents, the only way in which the practitioners can get others to recognize their claims is if the others already have the same interpretations of them, that is, if they already agree. It is only if the practitioners do not hold incompatible views that they can convey to each other what they take each other to be entitled to. In this case, the practitioners would agree in the interpretations they have of the claims being made. Reaching agreement will then be a consequence of them already holding the same views, endorsing the same claims, and acknowledging that this is the case. But if the practitioners do hold claims that they themselves see as incompatible—for example, when one of them presents a claim as being in disagreement with the view of the other—there is no way for them to reach agreement. For each interpreter has her own interpretation of the claims being made as the point of departure for ascribing meaning to all the claims being made in the exchange, and in this case the interpretations differ. What is lost in a scenario like this one is the possibility of interlocutors disagreeing, addressing incompatible claims to each other, in a way in which they can get their interlocutors to recognize their reasons. Disagreement in this picture can only be recognized when there was a previous reason “in the interlocutor’s own eyes” to change her commitments, that is, when there was a previous agreement, not a disagreement after all. Yet as we explained above Brandom himself explicitly emphasizes in responding to Habermas that agreement is not necessary for the practices to be sustained (Brandom 2000). According to his proposed inferentialist model, what the situation of normal disagreement amounts to, is that in the eyes of each scorekeeper the others “loses authority” (see Brandom 2010: 33). But, the situation is much worse. It is not only that the practitioners lose authority vis-à-vis each other but that in a situation like this, when there are disagreements, there is no import of the challenges or defensive claims of the practitioners to the way each one keeps their own score, and

108

G. Satne

thus, in such scenario, the only remaining authority would be the one that each individual scorekeeper has on herself. Thus, even if this may come as a surprise in a model such as Brandom’s where the social structure of the practice is essential, such an interpretational practice would be individual and not social. Each person’s reasons would not be inherited intersubjectively; on the contrary, each participant would only have individual sources of justification and she would only attribute entitlement to those assertions of others that were in her eyes indistinguishable from her own. This means that they would be reluctant to accept any challenge raised by others, not willing to change or revise their commitments in the light of other practitioners’ reasons. So, in this interpretative inferential practice each practitioner would simply refuse to recognize any authority but their own and as a consequence they would lose grip on any criteria of correction (any notion of being wrong). To put it in a Wittgensteinian way, in such cases, whatever seemed correct to them would be correct; and hence what the practitioners were doing would not be subject to any norms. In fact, they would lose the possibility of recognizing incompatibilities altogether, losing sensitivity to any cognitive friction and with it the possibility of having meant anything at all. This is Wittgenstein’s argument against the possibility of a private language, one that Brandom accepts when he claims that a practice to be normative has to be social. Thus, contrary to Brandom’s answer to Habermas, as this argument shows, agreement is essential to the game of giving and asking for reasons. If agreement is not reached, then, despite the inferentialist’s best intentions, the practices and the attributed contents end up being only individually held and not socially inheritable. As a consequence, private contents—articulating inferences attributed to one’s self and others—will be presupposed. This is a reductio of the position, since the view starts from the idea that content depends on social practices and there are no private contents.2

6.2.3  T  he ‘Minimalist Program’ as a Platform for the Emergence of Practices of Claim-Making One way to go to address this problem is to go back to the collective intentionality theorist’ toolbox and see if we can find a notion of shared practice, to which agreement is fundamental, to correct and complement Brandom’s account of ‘Basic Linguistic Practices’. The account at issue should not presuppose meaning of the sort we are trying to understand, something that fits Brandom’s Wittgensteinean adequacy conditions as well as the cognitively minimalist constraints that come from the aim of accounting for the development of human practices of

 For a detailed argument regarding this problem in Brandom, see Satne (2017).

2

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

109

claim-making.3 The ‘minimalist program’ (Knoblich et  al. 2011; Butterfill 2012; Butterfill and Sinigaglia 2014) seems to be suited for this task, as it pursues a non-­ cognitively-­demanding conception of joint action that can serve as a platform for the development and evolution of sophisticated human cognitive capacities (cf. Butterfill 2012). The minimalist program thinks of basic joint activities as the overlap of individual distributive goals pursued by the participant agents. Accordingly, its advocates define distributive goals in the following way: […]an outcome is a distributive goal of two or more agents’ actions just if two conditions are met. First, this outcome is one to which each agent’s actions are individually directed. Second, each agent’s actions are related to the outcome in such a way that it is possible for all the agents (not just any agent, all of them together) to succeed in bringing about this outcome. (Butterfill 2012: 849)

There are a number of reasons why this approach cannot offer an adequate account of a shared practice that could be at the basis of both the evolution and acquisition of shared semantic norms and highly cognitively-demanding practices that exploit such norms. The first is that individual goals in this framework are understood as individuals’ mental motor representations that have content of a basic, non-­ propositional kind (Butterfill and Sinigaglia 2014). Thus, a first question for the view is how to explain this kind of content in evolutionary terms. Seen in light of Brandom’s and Wittgenstein’s arguments on normativity and normative conditions for content, there is no possibility for these mental representations to have content since content depends on there being correctness conditions that establish what state of affairs in the environment accords with the goal as set by the mental representation. But as explained above, private contents cannot have conditions of correctness. Brandom and Wittgenstein insist that there are no mental representations of this kind, which is the kind that can establish particular goals for action. A second reason, more important for us in this context, is that if goals are shared only in the sense of distributively overlapping with each other, there is no necessary pressure to come to share goals in a stronger sense, i.e. by having the same contents in mind. After all, the contents of the mental representations of goals are individually specified, and overlap with the individual goals of others is possible even if pursuing very different individual goals (See Blomberg 2015). Thus, the view offers no explanation as to how contents (and contentful norms) come to be shared, if they ever do. Overlap suffices for coordination and distributive joint action but allegedly not for sharing norms and contents. In sum, this view is incompatible with Brandom’s and Wittgenstein’s takes on meaning and normative practices. It is also potentially problematic for a developmental and evolutionary story that tries to account for the origins of content as it gives no account of the basic kind of content it commits to nor of how semantic norms are thought to emerge. That is, minimalist accounts of this  As we discussed above, the Content view championed by Bratman is not suited for the task at issue because its story is too cognitively-demanding. For the same reason, cognitively-demanding accounts of the ‘we’ belonging to the Subject view—such as Gilbert’s (cf. Gilbert 2014), where joint practices depend on participants’ commitments to each other—are not suited for the task. 3

110

G. Satne

sort are unsuited for explaining the emergence of complex shared norms including those of objective claim-making in terms of social practices of the sort they focus on. In the following section I present an alternative view of joint action that stresses collectivity rather than overlap between individual mental states, and that is not cognitively demanding. The aim is providing an account of shared practices that meets the cognitively minimalist requirement without being immediately subject to the problems that affect Brandom’s interpretationist account of ‘Basic Linguistic Practices’.

6.3  Evolving Collective Intentionality In this section, I defend the view that shared intentionality is always collective or we-intentionality. Individuals engage in collective or we-intentionality when they are part of a group, which can be a dyad, that sets out to do something. This is an ‘evolving’ collective view in that instead of distinguishing two forms of shared intentionality—one joint, characterized by I/Thou relations, and one collective, characterized by I/We relations, as Tomasello does—it understands shared intentionality as collective from the outset, having different stages of development (see Satne 2016; Satne and Salice 2020). This view takes as its point of departure the apparent fact that children are immersed in social practices from the beginning of their lives, e.g. interacting with family members, care-givers, and peers, including siblings. The view proposes that the form of intentionality that belongs to these early stages in the emergence of social intentionality is a more primitive, minimal, but already collective form of intentionality. That is, it is a kind of social intentionality that can be thought to underlie and support the development of the full-fledged kind that involves conventions, language, institutions, and objectively valid contents. A primitive kind of group behavior that is not preceded by ‘joint intentionality’ in Tomasello’s sense or by Brandom’s practices of ‘normative scorekeeping’, but rather is integrated with more basic forms of group thinking and feeling. Thus, on the proposed view, collective intentionality admits a variety of degrees, ranging from minimal forms of interaction and mutual understanding, to very complex ones. Minimal collective intentionality in its turn designates special forms of non-­ cognitively demanding group thinking and feeling that are displayed in jointly acting. This minimal variety of collective intentionality is associated with a minimal and unsophisticated understanding of others’ mentality and agency. Such early forms of joint action do not require representations of others’ mental states from the outset, but spring from common background norms or habits that in turn provide a platform for the emergence of representational capacities informing more cognitively sophisticated forms of group behavior (see Hutto and Satne 2015; Satne 2016). The minimal form of collective intentionality at issue is basic also in the sense that it is not parasitic on already stable groups. This accommodates the developmental fact that basic forms of social intentionality are possible on a one-on-one basis.

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

111

In particular, in second-personal interactions between particular individuals that interact with one another in intentional activities, interactions that are particularly pervasive in the first years of life, e.g. when children interact in one-on-one dyads with care-givers (Reddy 2015; Satne and Salice 2020). In this framework, second personal interactions are understood as a form of collective intentionality, always involving the pursuit of common goals between participants, even the very basic ones of maintaining the interaction or being in communication.4 Thinking of collective intentionality as coming in degrees, ranging from minimal to sophisticated kinds, is also suited to explain the early stages in the emergence of cooperative behavior in the hominids’ evolutionary trajectory. In particular, it allows to make sense of social cooperation as underlying modern humans’ fitness and evolutionary success (cfr. Satne and Salice 2020). One important aspect of the evolving collective view, that brings it in contrast to Tomasello’s view, is that it claims that social interaction is possible, and actually begins, without the exercise of perspective-taking. In this view, collective intentionality comes in degrees and progresses from cognitively minimal forms to others that are sustained and scaffolded by social norms, language and conventions. The capacity to act and think within groups expands with time and training, beginning with perspective blindness and progressively moving towards the capacity to explicitly represent foreign perspectives towards ‘the same things’ and to consider judgments as objectively valid independently of any particular standpoint (see Satne 2016 and Tomasello 2016 for debate). It is only with time and exposure to multiple sources of social learning and training- including exposure to ‘foreign’ groups- that a more developed form of collective intentionality develops progressively. Essential to this progression is the emergence/learning of language, which is an external cultural reservoir of rules and norms (Clark 2005, 2006), and other sorts of environmental scaffolds (Sterelny 2012), in the context of which objective contents and the practice of claim-making have its proper place (Hutto and Satne 2015; Satne 2016). It is in the context of practices that are scaffolded by shared linguistic norms that participants become answerable to shared norms that are taken to be universally valid and it is in those same linguistic practices that  judgments are taken to be answerable to how things are with respect to what they are about regardless of the particular standpoint of the judger. In this respect, the evolving view of Collective Intentionality is importantly different from both Brandom’s and Tomasello’s accounts of shared activities. For in the evolving collective view, agreement is fundamental and comes hand-in-hand with the ability to follow shared norms. This capacity is not built upon the capacity to explicitly represent contrasting I-Thou perspectives. Rather, the latter depends on the capacity to represent objective facts, i.e., to adopt an “objective standpoint”, a

 For a detailed argument in favor of the idea that I/Thou relations are always We relations as well, see Rödl (2015) and Satne (2014a). 4

112

G. Satne

capacity that itself crucially depends on the emergence and mastery of linguistic practices governed by special norms of truth-telling.5 Yet some authors claim that at the early stages of the hominins’ evolutionary trajectory there were no groups and thus no group norms in place to be followed and learned (cfr. Tomasello 2014), thus posing a prima facie challenge for the idea that social intentionality can be thought to be collective all the way down, as the evolving collective view claims. However, the minimal form of collective intentionality here outlined is not meant to depend on existing norms or already stable groups. Rather, we can see it as progressively contributing to their development. A few individuals behaving alike under the same circumstances and with similar needs and environmental pressures might have sufficed to get some common practices off the ground which, in turn, fostered the emergence of the capacity to engage in minimal forms of collective intentionality that gave rise to more reliable forms of group binding, those themselves supported by the emergence of shared norms. That notwithstanding, contrary to the claim that at the early stages in the evolution of the hominins there were no groups, it is a much more likely hypothesis that humans’, chimps’ and bonobos’ common ancestors lived in groups, as chimps and bonobos in the wild do, and that the specific shift in hominins’ evolution was on the kinds of social norms and collective engagement that each species is capable of, rather than a turn from individual forms of life to social ones (for other views along these lines see Sterelny 2012; Birch 2017). Thus, in the picture offered by the evolving collective view, minimal forms of shared agency and social interaction help explain the development and evolution of more sophisticated forms of social cognition and shared activity correlated with abilities that are more cognitively demanding. The emergence of social norms is explained as a progressive development from basic forms of collective intentionality. Minimal collective intentionality enables participation in shared activities that provide a platform for the institution and development of social norms, including objectively valid norms in which objective contents have their home. In contrast to Tomasello’s and Brandom’s models, this progression of evolving forms of collective engagement allows understanding the emergence of human specific claim-making practices without committing from the outset to attitudes of perspective-taking, interpretation, or mind-reading abilities. With the aim of substantiating this claim, in what follows a brief outline of the psychological progression of collective intentionality capacities is provided. This is meant as an answer to a ‘how possible’ question, providing the key steps and platforms that give substance to the alternative defended, rather than as a fully developed empirical account of the progression. While more details will be required for the latter, the outline below points to the relevant literature that expands on the main characteristics of each step of the progression.

 For a further elaboration on this point, see Hutto and Satne (2015) and Satne (2014a).

5

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

113

Developmental psychology studies show that the development of the capacity for collective intentionality breaks into several stages. In the case of infants, the ability to coordinate and attune to others’ directed actions and reactions might suffice for making sense of a minimal capacity for collective intentionality, belonging to very simple shared activities in early infancy. Care-givers/infants interactions show that 2- to 4-month-old babies accommodate and bodily anticipate the usual style of their care-givers pickup postures (see Reddy 2015; Reddy et al. 2013b) and 6-month-old babies bodily comply with intentional directives (including linguistic ones, Reddy et al. 2013a; Reddy 2015; see also Rochat 2015). What is characteristic of these forms of interaction is that each participant’s role in the activity is determined and structured by the shared activity in which they—as a group/we/dyad—are taking part, without the participants necessarily planning ahead or intellectually predicting and representing the mental states of their partners (see Reddy 2015; Rochat 2015 and Satne 2020). For older children, the ability to identify with a group is a key pillar for the development of the capacity of sharing more sophisticated goals with others (Rakoczy et al. 2008; Salice and Miyazono 2019). The ability to adopt the group perspective, coupled with abilities for practical reasoning, enables, e.g., the formation of long-­ standing group goals whose achievement requires complex planning and coordination through time, and presupposes the exercise of normative pressure on the individuals to provide the expected contribution (see Satne and Salice 2020 for a detailed account). The progression from a minimal form of collective intentionality, based on bodily coordination and interpersonal attunement, to group-based forms of collective engagement, crucially involves the scaffolding of shared intentional activities by shared cultural norms, and the exercise of practical reasoning, that together make possible for more sophisticated reason-guided collective behavior to emerge.6 Accordingly, the psychological abilities that allow for the institution of norms, which correspond to minimal forms of collective intentionality, are not abilities of mindreading, but dispositions for ‘social conformism’. Following Haugeland, ‘social conformism’ can be described as a set of psychological dispositions to follow others in a community along with corresponding dispositions to monitor and control the behavior of other members of the community, in his words: When community members behave normally, how they behave is in general directly accountable to what’s normal in their community; their dispositions have been inculcated and shaped according to those norms, and their behavior continues to be monitored for compliance. (Haugeland 1990: 406)

 Crucial in this aspect are the folk-psychological narratives in which we are introduced through social practices of story-telling. See Hutto (2008) and Hutto and Gallagher (2008), for a detailed account of the development of practical reasoning through engagement in cultural narratives. 6

114

G. Satne

‘Social conformism’ is underpinned by a set of devices and mechanisms for shaping children’s dispositions to match the behavior of other members of the community and to continue monitoring them for compliance. Socio-conforming abilities have been studied by several researchers under different headings: viz. ‘mind-shaping’ (Zawidzki 2013), ‘mutual adaptation’, ‘folk-­ psychological normative regulation’ (McGeer 2007, 2015), ‘in-group bias imitation’ (Buttelmann et al. 2013), ‘natural pedagogy’ (Csibra and Gergely 2009), ‘normative enforcement’ (Rakoczy et al. 2008) and ‘imitation/emulation’ behaviors (Tomasello 1999; Tomasello et al. 2005; Tomasello and Rakoczy 2003; Tennie et al. 2009, 2012). While many theorists recognize the phenomenon of social conformism, they still assume that it is explained through mechanisms of mind-reading (e.g. Tomasello 1999; Rakoczy et al. 2008; Csibra and Gergely 2009; for a critical discussion see Satne 2014b). Contrariwise, the claim here is that social dispositions to conform such as bodily coordination and intersubjective attunement, matched with corresponding dispositions to impose and monitor norm-responsive behavior on the part of other community members, are sufficient to get a basic normative practice off the ground, one that constitutes a platform for the emergence of mind-reading strategies for social cognition that are transmitted and inculcated through such social conformist practices. These ‘basic’ mechanisms for social conformity get the practice of learning and teaching social norms and group behavior off the ground and do not require individuals to purposefully comply with rules from the get-go. Instead of representing rules and others’ assessments of their actions, individuals are sensitive to others’ assessments through more minimal cognitive pathways such as emotional attunement to others’ approval and disapproval and basic mechanisms for bodily and emotional coordination (cf. Satne 2014b, 2020).7 In sum, the institution of norms in this view is not explained through normative scorekeeping as in Brandom’s picture but as a result of social conformism which does not require in its most basic forms any kind of perspective taking. The child, born with abilities for bodily coordination and intersubjective attunement, slowly learns norms from others and progressively exercises her capacity for practical reasoning in understanding those norms, a capacity that develops in jointly exercising it with others in conversation and other cooperative activities, including activities such as storytelling and role playing. These are the abilities that later on in development  allow for the distinction between perspectives, the kind of attitudes that Brandom’s account of interpretative score-keeping privileges. The interpretive activity of attributing normative statuses to each other—a capacity that makes the  Accounting for the introduction of children into normative practices through social conformism has the additional advantage of avoiding giving too much weight to intricate debates about the emergence of ToM abilities and the nature of their early versions (see Low and Perner 2012). The remarks on developmental trajectories provided here fit well the specialized empirical literature on the development of capacities for understanding others in terms of reasons, placing them at the end of the developmental process, at around 3 to 4 years of age (Ibid.). In this view, infants’ success in allegedly early ToM tasks tracks not ToM abilities but precursors of them, in particular, infants’ abilities to monitor the positions of others for possibly engaging with them in basic coordinated activities (see Satne and Salice 2020 and Satne 2020). 7

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

115

perspectives of a member-of and an outsider-to the linguistic community equivalent (cfr. Brandom 2010)—fundamentally depends on capacities for social conformism, which are in place prior to the emergence of capacities for normative scorekeeping. Social conformism is only achievable within a discursive community, for it is there that one’s dispositions are shaped and monitored for compliance. Perspective-­ taking, an attitude that can be taken towards an alien community, in its turn, is only possible when a perspective has been adopted in the first place, and the capacity to adopt such a perspective depends on in-group attunement and in-group norm-­ abiding practices being already in place and well established. That our norms and social practices are thought to be shaped through social conformism and joint reasoning means, against Brandom, that it is simply not true that the fundamental form of our engagement with others is interpretational. Rather, our primary social interactions, the ones that can be thought as characteristic of ‘basic linguistic practices’, imply ‘sharing’ reasons in a much stronger sense than the inferentialist view makes available, a sense that is irreducibly inaccessible to an interpretational-external stance from which we are only observable or interpretable as doing something without experiencing it together. What is at stake in the case of participatory interaction of the sorts described by the evolving collective view is the acquiring and occupying of a place in a shared space that is shaped by our responsiveness to each other. The participant’s perspective emphasized by the evolving collective view implies that a fundamental agreement is built upon the actual exchanges and mutual responsiveness of the practitioners in shared practices. Moreover, in this view, objective norms, the norms of truth, are already in place, shaping our practices, well before the emergence of our ability of attributing commitments and entitlements to each other.

6.4  Concluding Remarks This paper aimed to show how some of the theoretical assets put forward by collective intentionality theorists and inferentialists can productively complement each other in pursuing a crucial theoretical endeavor, i.e., that of offering key tools for understanding the development and evolution of human specific cognitive capacities of claim-making. And  furthermore, that doing this allows both theoretical approaches to overcome the difficulties that they face when trying to pursue this inquiry by relying on their own resources alone. Brandom’s inferentialism misconstrues the nature of basic social linguistic practices by committing to an interpretational understanding of the attitudes of the agents involved in such practices. Collective intentionality theorists in their turn assume that social practices require on the part of the agents to craft intentions with contents to represent other agents’ intentional states as well as their own, blocking in this way their chances to give an account of the emergence of such contentful attitudes and their intrinsic normativity.

116

G. Satne

The proposal advocated here resists both tendencies finding both problematic for the task at issue. It offers an account of collective engagements that is not cognitively demanding, and yet suffices to be a platform for the emergence of more sophisticated and cognitively demanding shared activities. It also claims that basic collective engagements of the kind infants are capable of do not presuppose the grasp of contents on their part, nor the ability to represent foreign perspectives. Much along the lines of inferentialists, the view proposes to understand the emergence of content in general—and objective content in particular—in terms of the intersubjective institution of social norms. Norms of claim-making are one special kind of those norms, themselves dependent on the learning and mastering of other norms, including those for the use of a natural language. Contrary to Brandom’s Inferentialism, this account treats I-We and I-Thou interactions, not as one explaining the other but as two sides of the same coin, both unfolding together in development. The account focuses not on practices of mutual score-keeping, but on basic forms of social conformism. Social conformism progressively shapes a sense of group membership and gives raise to primitive forms of norm-following upon which sophisticated practices of normative score-keeping, as the ones Brandom describes, are built. In sum, this paper has argued that  theoretical tools both from Inferentialism and collective intentionality theories can be productively combined to account for the natural origins of human practices of claim-making. The core claim of the evolving collective view here presented is that introducing in the theoretical mix a primitive capacity for collective intentionality based on non-cognitively demanding dispositions to social conformism, that lays the ground for the emergence of social norms, may provide the right tools to account for the emergence of objective content. While the view was presented in its bare bones, my aim here was to lay out a possible and appealing path to be further developed.

Bibliography Birch, J. (2017). The philosophy of social evolution. Oxford: Oxford University Press. Blomberg, O. (2015). Shared goals and development. The Philosophical Quarterly, 65(258), 94–101. Brandom, R. (1994). Making it explicit. Reasoning, representing & discursive commitment (MIE). Cambridge/London: Harvard University Press. Brandom, R. (2000). Facts, norms and normative facts: Reply to Habermas. European Journal of Philosophy, 8, 356–374. Brandom, R. (2010). Conceptual content and discursive practice. Grazaer Philosophische Studien, 81(2010), 13–35. Bratman, M. (1992). Shared cooperative activity. The Philosophical Review, 1992, 327–341. Bratman, M. (2013). Shared agency. A theory of planning together. Oxford: Oxford University Press. Brownell, C. A. (2011). Early developments in joint action. Review of Philosophy and Psychology, 2, 193–211. Buttelmann, D., Daum, M., Zmyj, N., & Carpenter, M. (2013). Selective imitation of in-group over out-group members in 14-month-old infants. Child Development, 84(2), 422–428. Butterfill, S. (2012). Interacting mindreaders. Philosophical Studies, 165(3), 841–863.

6  Collective Intentionality, Inferentialism and the Capacity for Claim-Making

117

Butterfill, S., & Sinigaglia, C. (2014). Intention and motor representation in purposive action. Philosophy and Phenomenological Research, 88(1), 119–145. Clark, A. (2005). Word, niche and super-niche: How language makes minds matter more. Theoria, 20(54), 255–268. Clark, A. (2006). Language, embodiment, and the cognitive niche. Trends Cognitive Science, 10(8), 370–374. Csibra, G., & Gergely, G. (2009). Natural pedagogy. Trends in Cognitive Sciences, 13(4), 148–153. Davidson, D. (1973). Radical interpretation. Dialectica, 27(1), 314–328. Gilbert, M. (2014). Joint commitment: How we make the social world. New  York: Oxford University Press. Habermas, J. (2000). From Kant to Hegel: Robert Brandom’s pragmatic philosophy of language. European Journal of Philosophy, 8, 322–373. Haugeland, J. (1990). The intentionality all-stars. Philosophical Perspectives, 4, 383–427. Hutto, D. (2008). Folk psychological narratives. Cambridge, MA: MIT Press. Hutto, D., & Gallagher, S. (2008). Understanding others through primary interaction and narrative practice. In J. Zlatev, T. Racine, C. Sinha, & E. Itkonen (Eds.), The shared mind: Perspectives on intersubjectivity (Converging evidence in language and communication research) (pp. 17–38). Amsterdam: John Benjamins Publishing Company. Hutto, D., & Satne, G. (2015). The natural origins of content. Philosophia, 43(3), 521–536. Knoblich, G., Butterfill, S., & Sebanz, N. (2011). Psychological research on joint action: Theory and data. In B. Ross (Ed.), The psychology of learning and motivation (pp. 59–101). Burlington: Academic. Kukla, R., & Lance, M. (2009). Yo!’ And ‘lo!’: The pragmatic topography of the space of reasons. Cambridge, MA: Harvard University Press. Low, J., & Perner, J. (2012). Implicit and explicit theory of mind: State of the art. British Journal of Developmental Psychology, 30, 1–13. Malpas, J. (2011). Dialogues with Davidson: Acting, interpreting, understanding. Cambridge, MA: MIT Press. McGeer, V. (2007). The regulative dimension of folk psychology. In D. D. Hutto & M. Ratcliffe (Eds.), Folk psychology re-assessed (pp. 137–156). Dordrecht: Kluwer/Springer. McGeer, V. (2015). Mind-making practices: The social infrastructure of self-knowing agency and responsibility. Philosophical Explorations, 18(2), 259–281. Michael, J., Christensen, W., & Overgaard, S. (2014). Mindreading as social expertise. Synthese, 191(5), 1–24. Over, H., & Carpenter, M. (2009). Eighteen-month-old infants show increased helping following priming with affiliation. Psychological Science, 20(10), 1189–1193. Pacherie, E. (2013). Intentional joint agency: Shared intention lite. Synthese, 190, 1817–1839. Peregrin, J. (2012). Inferentialism and the normativity of meaning. Philosophia, 40(1), 75–97. Peregrin, J. (2014). Rules as the impetus of cultural evolution. Topoi, 33, 531–545. Rakoczy, H., Warneken, F., & Tomasello, M. (2008). The sources of normativity: Young children’s awareness of the normative structure of games. Developmental Psychology, 44(3), 875–881. Reddy, V. (2010). Engaging minds in the first year: The developing awareness of attention and intention. In G.  Bremner (Ed.), Handbook of infant development (pp.  365–393). Oxford: Wiley-Blackwell. Reddy, V. (2015). Joining intentions in infancy. Journal of Consciousness Studies, 22, 1, –2, 24-44. Reddy, V., Liebal, K., Hicks, K., Jonnalagadda, S., & Chintalapuri, B. (2013a). The emergent practice of infant compliance: An exploration in two cultures. Developmental Psychology, 49(9), 1754–1762. Reddy, V., Markova, G., & Wallot, S. (2013b). Anticipatory adjustments to being picked up in infancy. PLoS ONE, 8(6), e65289. https://doi.org/10.1371/journal.pone.0065289. Rochat, P. (2015). Self-conscious roots of human normativity. Phenomenology and the Cognitive Sciences, 14(4), 741–753.

118

G. Satne

Rödl, S. (2015). Joint action and recursive consciousness of consciousness. Phenomenology and the Cognitive Sciences, 14(4), 769–779. Rouse, J. (2015). Articulating the world: Conceptual understanding and the scientific image. Chicago: University of Chicago Press. Salice, A., & Miyazono, K. (2019). Being one of us. Group identification, joint actions, and collective intentionality. Philosophical Psychology, 33(1), 42–63. Satne, G. (2014a). What binds us together: Normativity and the second person. Philosophical Topics, 42(1), 43–61. Satne, G. (2014b). Interaction and self-correction. Frontiers in Psychology, 5, 798-1–798-11. Satne, G. (2016). Joint and various (collective) forms of intentionality. Journal of Social Ontology, 2(1), 105–116. Satne, G. (2017). Brandom and the second person. International Journal of Philosophical Studies, 25(2), 189–209. Satne, G. (2020). Understanding others by doing things together. In Synthèse. https://doi. org/10.1007/s11229-020-02692-2. Satne, G., & Salice, A. (2020). Shared intentionality and the cooperative evolutionary hypothesis. In A. Fiebich (Ed.), Minimal cooperation and shared agency, 71–92. New York: Springer. Schmid, H.-B. (2009). Plural action: Essays in philosophy and social science. Dordrecht: Springer. Schweikard, D. & Schmid, H.-B. (2013). Collective intentionality. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy (Summer 2013 edition). Searle, J. (1990). Collective intentions and actions. In P. Cohen, J. Morgan, & M. Pollack (Eds.), Intentions in communication (pp. 401–415). Cambridge, MA: MIT Press. Sellars, W. (1954). Some reflections on language games. Philosophy of Science, 21(3), 204–228. Sellars, W. (1956). Empiricism and the philosophy of mind. Minnesota Studies in the Philosophy of Science, 1, 253–329. Sterelny, K. (2012). The evolved apprentice. Cambridge, MA: MIT Press. Tennie, C., Call, J., & Tomasello, M. (2009). Ratcheting up the ratchet: On the evolution of cumulative culture. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 364(1528), 2405–2415. Tennie, C., Call, J., & Tomasello, M. (2012). Untrained chimpanzees (Pan troglodytes schweinfurthii) fail to imitate novel actions. PLoS One, 7(8), e41548. Tollefsen, D. (2005). Let’s pretend! Children and joint action. Philosophy of the Social Sciences, 35(1), 75–97. Tollefsen, D., & Dale, R. (2012). Naturalizing joint action: A process-based approach. Philosophical Psychology, 25(3), 385–407. Tomasello, M. (1999). The cultural origins of human cognition. Cambridge, MA: Harvard University Press. Tomasello, M. (2008). Origins of human communication. Cambridge, MA: MIT Press. Tomasello, M. (2014). A natural history of human thinking. Cambridge, MA: Harvard University Press. Tomasello, M. (2016). Response to commentators. Journal of Social Ontology, 2(1), 117–123. Tomasello, M., & Rakoczy, H. (2003). What makes human cognition unique? From individual to shared to collective intentionality. Mind & Language, 18(2), 121–147. Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. The Behavioral and Brain Sciences, 28(5), 675–691. Warneken, F., & Tomasello, M. (2007). Helping and cooperation at 14 months of age. Infancy, 11(3), 271–294. Wittgenstein, L. (1953). Philosophical investigations. New York: Macmillan. Zahavi, D., & Satne, G. (2015). Varieties of shared intentionality: Tomasello and classical phenomenology. In J. Bell, A. Cutrofello, & P. Livingston (Eds.), Beyond the analitic-continental divide: Pluralist philosophy in the twenty-first century (pp. 305–325). London: Routledge. Zawidzki, T. (2013). Mindshaping. A new framework for understanding human social cognition. Cambridge, MA: MIT Press.

Chapter 7

Wherein is Reasoning Social? Ladislav Koreň

Abstract  One of the main tenets of inferentialism is that reasoning is primarily a competence to play the social game of giving and asking for reasons. A rather similar idea has recently been elaborated from the naturalistic perspective by Hugo Mercier and Dan Sperber. They argue that reasoning evolved originally and primarily for the purposes of social justification and argumentation. In this chapter I compare the inferentialist account of reasoning with the approach of Mercier and Sperber. I argue that although Mercier-Sperber’s naturalistic hypothesis is promising, their account of how reasoning works faces a number of philosophical objections that inferentialism has the resources to overcome. At the same time, I show that “naturalized” inferentialists could find congenial much of what Mercier and Sperber have to say about the social origins and functions of reasoning. Finally, I consider a different naturalistic account of the social nature of reasoning due to Michael Tomasello, who submits that reasoning evolved primarily for cooperative argumentation and joint or collective decision-making. I argue that Tomasello’s account may go too far in the collectivist tradition and I conclude that a more plausible naturalistic account of the sociality of human reasoning would likely recognize a mix of different social contexts and functions of reasoning. Keywords  Reasoning · Inferentialism · Argumentative theory of reasoning · Collective intentionality

7.1  Introduction A venerable view has it that humans are animals with a unique talent for reasoning. To reason is to engage in a deliberate, conscious thinking activity purporting to work out reasons for or against accepting or doing something. Its development and

L. Koreň (*) Philosophical Faculty, University of Hradec Králové, Hradec Králové, Czechia e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_7

119

120

L. Koreň

cultivation is socially mediated, but it is primarily a capacity tailor-made to secure individual thinkers more or better knowledge and decisions. Yet this view has not gone unchallenged. An alternative account has been in the air ever since ancient times. According to it, reasoning is originally and primarily at home in communicative exchanges in which people make, challenge, defend, modify or retract claims. It comprises discursive, intersubjective performances of the production and assessment of reasons. Secondarily, it is co-opted for individual purposes, including for ratiocination in foro interno. Inferentialism offers an up-to-date elaboration of the idea that reasoning is primarily a social competence that manifests itself in the social game of giving and asking for reasons.1 In Robert Brandom’s paradigmatic version, that game consists, at a minimum, of claim-making moves and moves that request and give reasons respectively (vis-à-vis claims). To say that reasoning is a social competence to make moves in this game is to say that exercising it involves practically treating the claims (and actions more generally) of oneself (and others) as potentially standing in need of justifying reasons or as supplying reasons for or against other claims. Reasoners must also be capable of making normative discriminations manifesting their sensitivity to reasons. To play the game skillfully is to be capable of discriminating and assessing claims and reasoning moves as appropriate or not. This issue has also recently been revisited in the branch of cognitive psychology focusing on reasoning. Much research has proceeded on the assumption that reasoning is primarily a tool of individual cognition and decision-making. However, there have been controversies about how good a tool it is, under what conditions, for what kind of tasks, and measured by what standards of goodness. Meanwhile, a few dissenting voices have argued that reasoning is primarily a social competence. Arguably the best-developed approach has been proposed by Hugo Mercier and Dan Sperber (henceforth M&S). Assuming that reasoning is a trait designed by evolution that should work well when doing what it is designed to do, M&S reverse-engineer its proper function based on the evidence about its characteristic features and performance. According to them, the evidence strongly suggests two social functions intimately linked to interaction and communication. Reasoning serves a justificatory function: we articulate reasons to justify our views or decisions to one another and are able to assess and probe such reasons. It also serves an argumentative function: we articulate reasons in order to convince one another to accept our views or decisions and are able to assess and probe such arguments. M&S hypothesize that reasoning evolved to serve those functions because they had fitness-enhancing dividends for our hypersocial ancestors increasingly dependent on coordination, cooperation, reputation and communication.2 Inferentialism and M&S’s account both ground reasoning in the competences to play the social game of giving and asking for reasons. That said, their theoretical

 See Sellars (1954) and especially Brandom (1994).  Cf. Mercier and Sperber (2009, 2011, 2012, 2017, 2018). Hereafter I refer to their work as: M&S (…). 1 2

7  Wherein is Reasoning Social?

121

goals and methodological approaches are different. Engaging them in a mutual dialogue that highlights their strengths and weaknesses promises cross-fertilization that might also pave the way for a deeper and more balanced account of the social roots and nature of human reasoning. Or so I shall attempt to show in this study. Here is the plan. After introducing the main tenets of M&S’s account of reasoning, I confront it with two prima facie problems: the oversophistication problem and the spectre of  vicious regress, respectively. My diagnosis is that their common denominator is the assumption that reasoning requires representing reasons as such. I then argue that inferentialism suggests a way to avoid those problems by dispensing with that assumption. On the other hand, when it comes to the question “Why do we reason?”, I find the general tenor of M&S’s account appealing. But this issue is barely scratched in the inferentialist literature. I outline how “naturalized” inferentialists could approach it, drawing on M&S’s account as a rich source of congenial ideas. Finally, I address the worry that the two approaches being compared offer too individualistic models of the social game of giving and asking for reasons that fail to do justice to the cooperative nature of human reasoning. I focus on Michael Tomasello’s articulation of this worry.3 His own view is that reasoning evolved primarily to facilitate cooperative argumentation and collective decision-making, which require common goals, and hence collective intentionality of sorts. I argue that Tomasello’s view might go too far in the collectivist direction and conclude that a more plausible reconstruction should appreciate various interactive (including mixed-motive) contexts, uses and functions of human reasoning.

7.2  Prolegomena: Inference, Intuition and Reasoning To get a grip on M&S’s approach, we should first understand how they use three key notions: cognition, inference and reasoning. They construe the first in terms of processing or transforming of information. Inferences, then, are the means of expanding or otherwise updating information available to a cognitive system.4 They require information-carriers of sorts – representations – serving as their inputs and outputs, respectively. Depending on the kind of cognitive process, representations consumed or produced by inferences specific to it may be conceptual or non-conceptual, conscious or not. Of particular interest are intuitive inferences. On M&S’s account (M&S 2017, chapter 3), we are consciously aware of their outputs but not of the processes producing them. What qualifies their outputs as intuitions is the fact that they pop up into our conscious awareness under a specific metacognitive mode of presentation. We feel them to be right, with a positive degree of confidence. But neither our confidence nor our acceptance of them is backed up by conscious reasons. When I

 Cf. Tomasello (2014).  Cf. M&S (2017, pp. 53, 55–56).

3 4

122

L. Koreň

instantly recognize that my friend is being sad, this may be an intuitive upshot of a subpersonal processing of some data or cues (e.g. about his facial expression); but I certainly do not need to consciously represent those data as the reasons or evidence to think so. M&S submit that a large part of our cognition and decision-making recruits such unreflective processes, whose outputs (intuitive judgments or decisions) we consciously accept (with some metacognitive confidence) but not for conscious reasons. Though our conscious thoughts and acts are underpinned by such unreflective processes, we humans have a cognitive wherewithal to reflect on them. In particular, we can attend to or consider reasons for or against something we accept, are tempted to accept or merely entertain. M&S distinguish four basic cases in which this capacity – which they equate with reason – is exercised. First, reasons can be considered prospectively “as arguments in favor of new decisions or new beliefs“ (M&S 2017, p. 128). This may be either an individual process (inquisitive reasoning) or a social process whose goal is to convince others of something that one already accepts (argumentation). M&S reserve the term “reasoning (proper)” only for these cases, which they further distinguish from justification (and explanation), where reasons are considered retrospectively “to explain or to justify decisions already taken and beliefs already held” (M&S, p. 128). Inferences, broadly construed, are ubiquitous in the animal realm. But M&S are adamant that there is no evidence that non-human animals are capable of conceiving of something as apt to be accepted (rejected, retracted, revised) for conscious reasons. That seems right. Reason, at least as M&S understand it, appears to be a human-unique feat. M&S address two questions concerning this. First, what is reasoning for (the proper function(s) it is designed to fulfill)? Second, what kind of cognitive abilities does reasoning, thus construed, involve? I shall honor this order in what follows, addressing the first issue in Sect. 7.3 and the second issue in Sect. 7.4.

7.3  W  hat Is Reasoning for? Its Social Roots: Uses and Benefits For the sake of dramatic impact, let us first consider what M&S call the “classical” view of human reason.5 On that view, reasoning is first and foremost an individual thinking process that enables us to infer better beliefs and decisions based on considering or weighing evidence pro and contra. Its raison d’être is to enhance the reliability and coherence of our beliefs and decisions. Reason enables us to comprehend causal, evidential and logical relations, estimate consequences of events or actions and, based on that, select an action or action-plan according to whether it best promotes our short- or long-term goals and so is worth adopting or executing.  Cf. M&S (2017, p. 144).

5

7  Wherein is Reasoning Social?

123

As such, it has been contrasted with instincts, reflexes, stimuli-controlled perceptions, habits, and passions or intuitive seemings, its purpose being in part to supervise, pre-empt, override or correct inflexible, short-sighted and often biased responses produced by such unreflective processes. M&S think that the classical view seriously underestimates the role that unreflective processes play in our cognitive and practical economy.6 It regards them as ­troublemakers to be supervised and disciplined by reflective thinking. Yet, advances in cognitive psychology and neuroscience suggest to us a different view of the matter. On that view, we are blessed to have neurocognitive circuits carrying out unreflective processes, including affective and intuitive processes. They underlie all sorts of our skillful copying with the physical and social world surrounding us, upon which our survival, reproduction and thriving are predicated.7 In fact, conscious awareness is just the tip of a much larger iceberg, most of which consists of such processes. They are often based on evolved learning regimes and are exquisitely fine-tuned to cope with rather narrow domains of tasks. Like our perceptual process, they work reliably within their confines, yet might misfire when the conditions are somehow abnormal. Relatedly, the classical view imputes to reason much more power than is its due.8 If the purpose of reasoning is to enhance individual cognition and decision-making across the board, one would expect it to perform that function reliably enough. Yet, there is mounting evidence that people’s performance in experimental tasks designed to test individual reasoning competence is ill-calibrated vis-à-vis the cherished canons of good reasoning (logical, statistical, decision-theoretic). To drive the point home, subjects` judgments regarding the validity of arguments (of the same form) are liable to be affected by how believable they take the premises to be (belief-based reasoning) or by how believable they take the conclusions to be (belief-bias). Or, for that matter, people tend to seek concordant evidence in support of something they find believable and neglect or suppress discordant evidence (myside bias). People are also more likely to arrive at the conclusions that they wish to arrive at (motivated reasoning), tend to persist in holding their beliefs despite being presented with discordant evidence (belief perseverance), and tend to make choices they find easier to justify (reason-based choice). By M&S’s lights, the classical view fails to offer a plausible account of these phenomena. Rather, they call it into question.9 First, there is solid evidence that our unreflective cognitive processes tend to be reliable means of achieving our individual cognitive or practical goals. So when the beliefs and choices produced by them are correct or good, reasoning has nothing to improve on. Second, as a means of achieving our individual goals, reasoning is not all that reliable. And when beliefs  See especially M&S (2017, chapters 3–6).  Frith (2013) provides a good review. Damasio (1994), among others, has done much to popularize this idea. 8  For a review see M&S (2017, chapters 1–2, 11, 13) and also M&S (2011). 9  See especially M&S (2017, introduction, conclusion). A good review from a different perspective is Evans (2010). 6 7

124

L. Koreň

or decisions produced by unreflective processes are incorrect, reasoning often tends to advocate what we already find believable, rather than to check and correct them by taking into account evidence pro and contra. Yet, if its supposed raison d’etre is to reliably update our beliefs about the world, or to make sound decisions in it, we would expect it to be more impartial: i.e., sensitive not just to supporting, but also to discordant, evidence. M&S conclude that reason is unlikely to be fine-tuned to support individual cognition and decision-making. All the more so if we assume, as M&S do, that it is a faculty of mind that evolved. From this perspective, its estimated benefits are unlikely to have offset its costs: flaws, biases and the energetic needs of large brains required to implement it. After all, reflectively double-checking the reliability of one’s unreflective inferences or their coherence with other beliefs or desires is a cognitively demanding feat, vulnerable to biases and lapses of its own. M&S’s alternative hypothesis is that reasoning evolved as a specialized interactive device.10 They speculate that reasoning could have evolved to serve the function of producing and evaluating reasons in order to justify one’s claims, views or actions in the public space and thereby also justify oneself as a reasonable fellow. That was likely a significant boost to mutual interpretability and predictability  – what to expect of one another – hence to coordination, joint action and cooperation. Such a functional trait likely had adaptive effects for humans inhabiting hypersocial niches. Thanks to the development of social-communicative skills of explaining or justifying themselves in terms of verbalized reasons, ancestral humans were able to coordinate their interaction and cooperation more efficiently. This included committing effects (what others can expect from an agent and what she expects of them in turn), reputation-building effects (establishing, defending or rehabilitating one’s social profile and standing as a good cooperator or communicator), and keeping track of such effects, including indirectly via gossip.11 M&S also think that reasoning could have served another adaptive function: producing reasons as arguments aimed to persuade peers to adopt one’s beliefs, decisions, plans, etc.; and to assess the quality of others` arguments so as to be convinced by those that support beliefs and decisions likely to be good. This might have helped to solve an adaptive problem pertaining to the practice of verbal communication. Communication makes it possible to obtain cheaply a valuable commodity  – information. Being dependent on coordination and cooperation, we may assume that our ancestors had often been incentivized to share information. But this was not always the rule. People likely had divisive interests concerning scarce resources and temptation to cheat on others when they could reap the benefits of cooperation without incurring the costs. So sometimes they would have been tempted to pass on a dishonest or manipulative message.12 In such situations consumers of messages had

 M&S (2017, chapter 10).  Cf. M&S (2017, pp. 185–186). 12  This threat was amplified when non-iconic, arbitrary signals became the main currency of communication, as these do not wear the signs of their own reliability on their sleeves. 10 11

7  Wherein is Reasoning Social?

125

better not naively buy whatever message whoever tries to sell in the market place, or they would risk ending up with a non-adaptive ratio between valuable information and misinformation. Yet they should not be overly suspicious, or they would easily end up with a lesser amount of adaptive information. Rather, it pays to be epistemically vigilant,13 flexibly calibrating one’s trust according to whether the source is trustworthy or the message coherent with what the consumer knows or believes (or with other things that the producer is known or supposed to have said). If a producer of a message wishes to raise the chances that her message gets accepted, she must accordingly develop ways of convincing others, who might be vigilant, to buy it. Here, M&S submit, the argumentative function of reasoning proved valuable. If others resist, or if one expects their resistance, one might seek compelling reasons, which, if found acceptable by consumers, could make them accept the message. It then stands to reason that one should try to come up with considerations that are acceptable to others and, at the same time, support one’s intended message (claim, proposal). Symmetrically, receivers have their own tools to assist them in assessing the goodness of reasons, including sensitivity to alternatives and conflicting evidence that enable them to probe reasons offered. Initially, the function of interactive reasoning was not to secure reasoners better beliefs or decisions. At least, this holds for its productive uses. Thus, justificatory uses were designed to socially rationalize one’s already adopted doxastic or practical commitments (typically intuitive). And argumentative uses served to raise the social acceptability of a message that its producer typically already endorsed (again, intuitively). On the other hand, the consumers of reasons were more vigilant and demanding, being sensitive to the coherence of communicated message with other data, including taking account of discording evidence. Eventually, the two strategies – i.e., the consumers` interest in getting good information and the producers` interest in making their message accepted – might have converged into a virtuous equilibrium that brought about mutually adaptive dividends in terms of better beliefs and decisions. To sell their messages at the marketplace, the producers needed to adjust their strategies in expectation of vigilant evaluation on the part of real or potential consumers (including their past track record as reliable/honest or otherwise). And it might often be the best strategy to be honest and provide coherent reasons or evidence that also turns out to be reliable.14 This could explain the formation and stabilization of social norms governing such practices, including honesty, truth or norms of coherence. Reverse-engineering the adaptive function(s) of a trait is a highly speculative discipline. But M&S present their theory as accommodating much available evidence about the features of reasoning and its performance in various contexts and experimental conditions. Inter alia, research in psychology has amply documented that we excel at rationalizing our beliefs or actions, though this often involves a considerable dose of confabulation aimed to meet social-normative expectations.

13 14

 See also Sperber (2001) and Sperber et al. (2010).  Cf. M&S (2017, pp. 263–264).

126

L. Koreň

The theory further predicts better reasoning with others or in groups whose members share some common ground and goal while holding different views on the issue at hand, due to open discussion and exchange of reasons.15 Indeed, it was observed that groups fare better in Wason’s selection task (compared to solitary reasoners). The design is such that the experimenter shows you, say, the following four cards (sides up) – |A| |D| |3| |7| – telling you that each card has a capital letter on one side and a single-digit number on the other side. Then you are asked: Which cards should you turn over to test whether the following is true or false: If the card has an A on one side, then it has a 3 on the other side? Solitary reasoners usually fail (sometimes up to 90%), tending to turn over |A| and |3| (|3| is irrelevant to test the conditional) rather than to turn over, correctly, |A| and |7|. In marked contrast, when allowed to have a discussion and to argue proposed solutions, groups usually end up with the right solution.16 Finally, according to M&S, some notorious biases can be viewed as design features (rather than bugs) of reasoning if it is fine-tuned for social interaction.17 In particular, it should not surprise us that myside bias affects the producers of reasons who aim to justify their beliefs or actions or to convince others to accept or adopt them. After all, it would not sell one’s messages to point out evidence that does not favor them. But it also can be expected to engender suboptimal results in the context of solitary ratiocination (the risk of sticking to incorrect beliefs or suboptimal decisions, even reinforcing one’s confidence in them).18 However, interpersonal reasoning often involves an efficient division of epistemic labor: somewhat lazy producers of reasons tend to support their (intuitive) commitments with prima facie conforming reasons, leaving it to more demanding and objective consumers to probe them, eventually adjusting their own argumentative strategy. At the end of the day, both the producers and consumers can be expected to be better off in that they end up with better beliefs and decisions (compared to their performance when reasoning on their own).19

 Wason (1968). The effect of group polarization is explicable too: ideologically homogenous groups tend to end up even more homogenous, absent discussion of alternatives views and reasons in their support. 16  Moshman and Geil (1998). 17  Cf. M&S (2017, p. 219). 18  Cf. M&S (2017, pp. 263–264). 19  Admittedly, consumers might themselves be biased toward disconfirmation. But even then they can be expected to be non-dogmatic enough to listen to and eventually accept compelling justifications or arguments. 15

7  Wherein is Reasoning Social?

127

7.4  R  easoning as Based on Intuitive, Metarepresentational Inferences So far I have focused on how M&S address the question Why do we reason? Their answer is that reason is primarily a tool designed for justification and argumentation. They also address the question How do we reason?, focusing on the cognitive mechanisms and architecture of reasoning suitable to implement its function. Recall that M&S draw a distinction between intuitive inferences and reasoning. Only the latter involves reflection in that one is aware of a source of his inference and takes it as a reason for the conclusion. Yet, they take the two to be intimately linked. For reasoning is said to consist in a use of intuitive inferences of a special kind: inferences about the reason-conclusion relations. Let me expand on this a bit.20 First, representations of reasons are not simply given. They need to be inferred themselves from some data. It is possible in principle that the representation of P as a reason for Q is itself a reflective conclusion of some higher-order reflective inference, which involves some other reason to accept it. No doubt, this sometimes happens. But we normally do not consider reasons for reasons (etc.) beyond two or three rounds of reflective iteration. Anyway, at some point reflection about reasons for reasons clearly comes to a halt, or reasoning would never get off the ground. And in such cases we end up with an intuitive conclusion that something is a reason for something reached through an inferential process that is both opaque to us and does not involve any further considerations of reasons. Second, though, reasoning differs from intuitive inferences about particular domains of facts in that it operates at a higher level. Its task is to figure out whether representations – typically outputs of first-order intuitive inferences – are (can be) backed up by reasons. As reason is a reason for something, reasoning specializes at inferring higher-order relations between representations. This requires representing relations between representations  – a meta-cognitive, specifically meta-­ representational achievement. Finally, as committed massive modularists, M&S contend that lower-order intuitive inferences about distinct domains of facts (corresponding to distinct kinds of problems calling for different information-processing solutions) are products of a multitude of semi-autonomous modules (inferential processors).21 As reasoning seems to have a domain of its own – the reason-conclusion relations between representations – they conclude that it likely has a modular basis of its own. They call this basis a “reason module”, noting how different this is from the traditional view of reason as an all-purpose, domain-general inferential engine. Being specialized at inferring representations of the reason-conclusion relations based on relevant features of representations (such as coherence or incoherence), reason indirectly mimics a would-be domain-general inferential engine in that it can produce reflective conclusions about everything and anything. This is because representations whose  See in particular M&S (2017, chapter 6).  See M&S (2017, chapters 4, 5).

20 21

128

L. Koreň

relations it (meta-)represents can be about anything. This indirect effect is called the virtual domain-generality of reason.

7.5  Two Problems for the Metarepresentational Account I find M&S’s evolutionary perspective on the origins and function of reasoning appealing – up to a point at least. What I find a bit less convincing is their claim that its evolution was channeled via genetic routes producing dedicated cognitive predispositions specific to reasoning abilities. An alternative hypothesis could be that reasoning is an adaptive response to social pressures (including those specified by M&S), but it is rather an adaptive product of cumulative cultural evolution, ontogenetically fine-tuned via social interactions and learning.22 In my view, the evidence adduced by M&S does little to favor their hypothesis over this alternative. I would also worry that M&S’s claim that our minds are massively modular fails to carry conviction in part because reasoning, in particular, does not seem to operate modularly.23 But for the purposes of the present paper I will set these concerns aside her,24focusing just on the claim that reasoning consists in the use of metarepresentational inferences about reasons. This, I will now argue, faces a couple of difficulties discussed below. Recall that M&S maintain that reasoning essentially involves conscious attention to reasons. On their account, to attend to reasons is to represent something (P) as a reason for something else (Q). They do not say that one treats P as a reason for Q, where this might be just a matter of acting as if one consciously represented P as a reason for Q. They do not say that one just represents both P and Q and makes a transition from the former to the latter. Rather, what they say is that one must somehow represent the relation of support between P and Q. To wit: …their form is “P is a reason for Q” (for example, “That Amy has a fever is a reason to call the doctor”). (M&S 2017, p. 148)

As M&S put it, such metarepresentations concern the reason-conclusion relation, whose relata are (lower-order) representations themselves. On a natural reading of this, they have a propositional content in which something like the relational concept x is a reason for y is applied to P and Q as arguments. The first problem with metarepresentational accounts of this kind is that they threaten to over-intellectualize reasoning. Taken literally, they require that reasoners possess the quite sophisticated concept x is a reason for y (or some cognates

 Cf. Heyes (2018) and Dutilh Novaes (2018).  The basic problem is the apparent cross-domain promiscuity and context-sensitivity of reasoning that is information-hungry and sensitive to information of any kind (about any domain). 24  See the peer commentaries on M&S’s target article (2011) in Behavioral and Brain Sciences 2011, 32(2). See also the peer commentaries on M&S (2017) by Sterelny (2018), Dutilh Novaes (2018), and Chater and Oaksford (2018). Myself, I have provided a comprehensive overview in Koreň (forthcoming, Chaps. 4 and 5). 22 23

7  Wherein is Reasoning Social?

129

expressing support relations). But aren’t there smart creatures, e.g. young kids, capable of making and appreciatively responding to at least rudimentary reasoning (justifying or arguing) performances, who might not yet possess that concept?25 If so, M&S’s account seems too restrictive, as it cannot accommodate such cases.26 One might bite the bullet and say that such creatures do not reason, lacking the prerequisite reflective abilities. However, M&S do not want to say this about kids in particular (as opposed to non-human animals). In fact, they provide themselves some evidence that kids around the third year (or even earlier) manifest some nascent skills of argumentation or assessment thereof.27 They refer to studies indicating that kids that young prefer non-circular reasoning (e.g., “X went this way because I saw it go in this direction”) to circular reasoning (“X went his way because X went this direction”). Thus kids sort of appreciate that someone who says the former is a source of more helpful information than someone saying the latter. Supposing this indeed indicates nascent reasoning abilities, do kids that young also possess the concept x is a reason for y (or its cognates) apparently required to metarepresent the reason-conclusion relations? On M&S’s own account they should – or they do not reason at all. The problem is that M&S` adduce no direct evidence of such concepts and abilities in kids that young. The evidence that kids prefer non-circular over circular arguments might very well indicate some understanding of the object-level connective “because”. But it does not indicate any mastery of the concept x is a reason for y and the ability to apply it in higher-order thoughts. More generally, then, kids might have some practical competence with explanations or arguments involving other dialectical devices (“no”, “but”, “so”, etc.), which play a role in expressing attitudes of rejection, challenge or inference. In so far as this competence indicates some sensitivity to rational connections between claims, it might be taken to indicate a capacity for reasoning. Yet, there is nothing obviously metarepresentational about such dialectical moves and turns; on the face of them, they all belong to the object-­level discourse. So this early emerging reasoning competence reveals, at most, implicit sensitivity to reasons on kids’ part that falls short of a full-blown understanding of them as such.28

 Such objections have been formulated by a number of authors. Cf. Boghossian (2014), Broome (2013), and McHugh and Way (2016). Their target was, for example, Pettit’s metarepresentational account of reasoning (cf. Pettit 2007). See also Pettit’s response (2016). 26  This particular challenge applies to both reasoning (prospective) and justification (retrospective), in so far as both involve representations of something as a reason for something 27  Mercier et al. (2014). M&S (2017, chap. 16) and Castelain et al. (2018). 28  One could argue that the ability to respond to requests for reasons with ‘becausal’ claims is sufficient evidence for the ability to entertain higher-order thoughts to the effect that P counts in favour of Q. Cf. Müller (2018). But why should it be? It seems perfectly possible that young thinkers can do this without being able to express any higher-order thoughts to the effect that something counts in favour of something else. Three-year-old kids might be a case in point. If they do not have such expressive resources, what warrants the attribution of higher-order thoughts to them? In so far as they have any understanding to the effect that P counts in favour of Q, it might be a form of practical knowing how to treat something as a reason for something else. They need not represent reasons as such. I expand on this in my critical discussion of Pettit’s metarepresentational account in Koreň (forthcoming, Chap. 5). 25

130

L. Koreň

Pending any independent evidence that young thinkers possess sophisticated concepts and representational abilities, there is a more parsimonious account of their early reasoning competence, consistent with M&S’s view that it is originally a social competence, eventually internalized and co-opted for private-individual thinking. On this account, their reasoning competence co-develops hand-in-hand with their verbal-communicative competence and is primarily embodied in their dispositions to produce and respond to various dialectical moves in argumentative, explanatory or justificatory exchanges. No doubt, this already requires growing conceptual-cognitive sophistication. But it does not require specifically the ability to conceptualize reasons qua reasons. Such an ability may be a later development towards a full-blown reflective competence but not a prerequisite of the reasoning competence as such. The second problem that I want to discuss is that the metarepresentational account invites the spectre of  vicious regress. The problem specifically threatens M&S’s account of what they call reasoning proper (considering and using reasons prospectively), at least if we take it at face value.29 The way M&S officially characterize reasoning When we reason, conclusions do not just pop up in our mind as self-evident; we arrive at them by considering reasons to accept them. (M&S 2017, p. 52)

prima facie suggests that our taking account of reasons contributes to changing our view – adding beliefs, subtracting beliefs or both. So it is said to consist in prospectively considering reasons as “arguments in favor of new decisions or new beliefs” (M&S, p. 128). We are also told that it is a process of pursuing the goal of extraction of a new information from information already available by attending to reasons (M&S 2017, p. 53). And we are told that when we reason we arrive at a conclusion by considering the reasons to accept them (M&S, p. 52) and that we accept the conclusion because of higher-order thinking (or “reflection”) about them (M&S, p.  150). Indeed, this seems to apply also to cases when we assess the reasons intended by others to convince us of something – here, too, we might end up believing something new. Furthermore, M&S explicate what it takes to consider reasons in terms of entertaining higher-order representations of the reason-conclusion relations. Putting this together, we seem to be left with the following metarepresentational account of reasoning (MAR): (MAR) Reasoning involves entertaining a conscious representation R to the effect that one’s premises P support one’s conclusion Q and arriving at the conclusion Q in part because of that fact.30

 Cf. M&S (2017, p. 54).  This can be interpreted as a possible elaboration of what Boghossian (2014, p. 5) calls “the taking condition”: Reasoning (inferring) involves the thinker taking his premises to support his conclusion and drawing his conclusion because of that fact. My concerns in this section are analogous to those that Boghossian and others have articulated with respect to several possible elaborations of the taking condition. 29 30

7  Wherein is Reasoning Social?

131

The problem is that it is far from clear how R could contribute to concluding Q based on P other than via expanding the premise-set of the original inference or being involved in a further piece of higher-order reasoning. This, in a nutshell, is the gist of the problem of vicious regress: the original piece of reasoning supposedly involving the linking representation R requires another piece of reasoning that must also involve a linking representation of its own. And so on. Suppose that R supplies an additional premise, so that one actually reasons from the expanded premise-set (P, R) to Q. As this is a piece of reasoning in its own right, it would presumably require another linking metarepresentation R1 its own (to the effect that P together with R support Q). What, now, is the role of R1 in the inference from the expanded premise-set? As Lewis Carroll pointed out long ago – and many contemporaries concur – a vicious regress lurks here.31 Though we might iterate the process a few times, at some point, there must be some inference (or piece of reasoning) that no longer requires yet other representation Rn. And this should give us a pause: Why assume that concluding Q based on P requires a higher-order R in the first place? Alternatively, one may think of R’s role on the model of application of an inference RULE licensing a passage from P to Q. After all, inference rules are distinct from premises. So one might think that one consciously grasps what RULE requires (this act or state corresponds to R), recognizes that the antecedent conditions for RULE application are met, and infers (via RULE instantiation) what its consequent requires. But once again: this looks very much like a bona fide piece of reasoning; hence it requires yet another representation of a higher-order RULE1. So a version of the regress problem seems to re-enter through the backdoor.32 To be sure, there are more possible maneuvers to be considered; and perhaps some of them could overcome the regress problem. That said, the received literature on the topic reveals that this won’t be easy for metarepresentational accounts committed to something like MAR.33 As M&S characterize reasoning, the reasoner  Carroll (1895). Cf. Boghossian (2014), Wright (2014), Broome (2013), Hlobil (2014), Valaris (2014), and McHugh and Way (2016). 32  Cf. Boghossian’s elaboration (2014). It may be retorted that the process of rule-application is not a case of reasoning at all, but of something else altogether. Then the regress problem would not get off the ground. Maybe, but it is not clear why accept this, other than that it avoids representations of the reason-conclusion relations, and hence is not reasoning. For this is what is at stake. And one could argue that if we allow that such inferential processes need not involve higher-order representations of the reason-conclusion relations, we should wonder why the original inference is supposed to involve them. M&S appear to endorse an information-processing account according to which the rule-like process of transforming the representation of the premises P into the representation of the conclusion Q is subpersonal and unconscious. But then the question is what (if anything) the conscious metarepresentation R of the reason-conclusion relation contributes to this subpersonal process. I address this issue in the next section. Here I proceed on the assumption that, taken at face value, M&S’s official account of reasoning implies a substantial causal-explanatory role of R in the inferential passage from P to Q. My point is that it is far from clear that this can be explained in terms of intentional, person-level rule-application. See Koreň (2019) for a more detailed discussion. 33  See McHugh and Way (2016). 31

132

L. Koreň

c­ onsiders the reason-conclusion relations by entertaining a conscious, higher-order representation typified by R. Plus, their characterization implies that such a representation contributes to accepting the conclusion; and that this is what distinguishes reflective conclusions from intuitive ones. Any view of this sort is prima facie vulnerable to a version of the problem discussed here: it is hard to see how such a representation can contribute if not by expanding the premise-set of the original inference or being involved in yet another, higher-order piece of reasoning. Even supposing (as M&S do) that such representations are, in the last instance, intuitively formed, we still face the question of how they contribute to concluding something from something else if not via further reasoning.34 Having said this, I suspect that, at the end of the day, M&S’s would not vindicate MAR vis-à-vis the problem of vicious regress. Rather, their considered view might be that paying attention to reasons relations always or typically follows lower-order intuitive inferences; and these intuitive inferences are ultimately causally responsible for forming or reforming our beliefs.35 In other words: …we typically construct our reasons after having reached the conclusions they support. (M&S 2017, p. 138) Contrary to the commonsense view, what happens is not that we derive intuitive conclusions from reasons that we would somehow possess. What we do, rather, is derive reasons for our intuitions from these intuitions themselves by a further process of intuitive backward inference. We infer what our reasons must have been from the conclusions we intuitively arrived at. We typically construct our reasons as an after-the-fact justification. (M&S 2017, p. 142)

Now, I am not sure if the commonsense view assumes what M&S say it does. It seems to me that one who holds that some belief is intuitive typically also thinks that it is not backed up by reasons. Even so, M&S’s idea should be clear. As a reflective process, reasoning requires that one (meta-)represents P as a reason for Q. Yet, it is not because of this fact that one believes what one does. Rather, paying attention to reasons might be a form of metacognitive control accompanying and/or rationalizing something that already happened.36 Admittedly, our folk-­psychological intuitions suggest otherwise. But M&S suggest that, “to a large extent”, it is a cognitive illusion on our part: The conscious self … sees itself as making epistemic decisions to accept or reject conclusions, when in fact these decisions have been all but made at this sub-personal level. Reasoning as we consciously experience it, that is, as a series of conscious epistemic  There are attempts to avoid the problem by postulating rational intuitions of sorts – cf. Chudnoff (2014). But as these have a rather different structure, I shall not address them. An alternative could be to say that entertaining such a representation is just constitutive of reasoning and hence that there is no need for an extra contribution. Again, however, there is no hint that M&S think along these lines and so I shall not address it. 35  I expand on this in Koreň (2019). 36  Jonathan Haidt (2012) has famously propounded the same idea for the domain of moral reasoning. Richard (2020) argues that (reflective) reason has, at the very least, the right of “veto”: it can reflect on and fails to accept intuitive judgments based on subpersonal inferential processes. 34

7  Wherein is Reasoning Social?

133

assessments and decisions, may well be, to a large extent, a cognitive illusion. (M&S 2012, p. 376)

If it is not our grasp of reasons as such that makes us make up or change our views, MAR has to give (its second conjunct). MAR may perhaps explicate our person-­ level (“naive”)  view of reasoning; but then this is our cognitive illusion. But the good thing is that the problem of vicious regress evaporates. However, this radical rethinking of the role of reasons in reasoning comes at a price. First, how plausible is it to deny that we update our views by considering reasons or evidence? Even if person-level reflection is just the tip of the cognitive iceberg, it clearly does not follow that it is a mere appendix. At the very least, it seems to be well within our person-level powers to zoom in on our intuitive responses and eventually reject or resist them. In this sense, we can self-consciously stand back, control, check and self-regulate our thinking and decision-making. Our sensitivity and responsiveness to socially instituted norms of “better reason” plays a major role here. The primordial normative friction is second- or third-person pressure and appraisal. Now what matters here is what you, he or she (relevant others) think(s) of my views or actions, accepting, probing, rebuking or rejecting them, etc. And I am the one who is called to the task (responsibility) of justifying my views or actions (and hence myself) as reasonable. This is a personal level, because it is an interpersonal level. Once this is cognitively and motivationally internalized, our intrapersonal reasoning is transformed too. It then becomes possible for me to think of my own views or actions as more or less justified, accepting, probing, rebuking or rejecting them, etc. And I am the one who is supposed to do something about it. It seems to me that to talk about our capacities to make up or change our minds by reasoning is in large part to talk about such feats. Admittedly, this would not be possible without something happening in our nervous system. But to say that decisions are already taken at the sub-personal level confounds different levels of understanding. Second, note that this account of the role of reasons still invokes metarepresentations. It just does not construe them as intermediate representations implicated in extracting conclusions from premises. Thus, even if it might avoid the vicious regress problem, it still faces the problem of oversophistication. What lesson does this teach us? I think that we should distinguish two claims. (i) Reasoning requires some sensitivity and responsiveness to normative-rational relations between claims, challenges, etc. (ii) This sensitivity requires metarepresentations of rational relations. In my view, some sensitivity to reasons seems required to distinguish reasoning as a species of norm-governed, person-level inferential thinking from other kinds of thought processes not so governed. So (i) is appealing, properly elaborated. But (ii) is a much more contentious claim, which the common denominator of the difficulties discussed previously. I next argue that we can uphold a version of (i) whilst dispensing with (ii). My proposal is that inferentialism offers a useful recipe for this.

134

L. Koreň

7.6  R  easoning and the Game of Giving and Asking for Reasons Inferentialism is known as an attempt to explicate the nature of conceptual competence itself in terms of abilities to reason, themselves explicated as practical abilities to make moves in discursive practices that display a certain normative-pragmatic structure.37 It is a species of inferential role semantics, explaining the conceptual content of linguistic or mental items as conferred upon them – wholly or partly – by the role they play in a larger economy of reasoning. For my purposes here, its credentials as an account of conceptual competence will not matter. What matters is the claim that the social game of giving and asking for reasons embodies the primordial economy of reasoning. I first reconstruct the core of the inferentialist analysis of the normative-pragmatic structure of that game. Then I show what light it can shed on the issues discussed previously. Brandom argues that the game of giving and asking for reasons must incorporate performances with the significance of claims (assertions).38 For claims are basic discursive moves in a space of reasons – basic components of reasoning. They are practically treated as potentially in need of justification and as potentially justifying other claims and actions (including those of other agents). Brandom elaborates on this by saying that the claim P is treated as default justified until somebody makes a pointed challenge. To level such a challenge – e.g., by uttering something, Q, that is treated as excluding P, and hence providing reasons against P – is to prompt its author to back P up by some other claims that are treated as providing reasons for P. If the author successfully discharges this conditional task-responsibility to justify her properly challenged claim, the claim counts as justified and can be relied upon in that it licenses further theoretical or practical commitments. Should she fail in this, the claim will no longer count as justified and she would lose her initial authority on which others could rely (they are no longer licensed to use it as fodder for their theoretical or practical reasoning or defer to the author for its justification). The idea here is that only if a social-communicative practice exhibits this kind of normative-pragmatic structure can we say that it incorporates the core language game of staking claims. To be able to skillfully play this core game is to be able to reason from and to claims. At a minimum, one must be practically able to treat some public performances as having the significance of claims and assess them as correct or otherwise by adopting certain practical-normative responses. This involves treating some inferences from and to claims thereby made as good – licensing or excluding other claims, as well as being licensed or excluded by yet other claims (including being prepared to discriminate, rebuke or correct deviations from such proprieties). For all that, however, one does not need to go “meta” in the sense of making assertions or forming beliefs about the properties or relations of assertions, claims, 37 38

 Cf. Brandom (1994) and (2000). See also Peregrin (2014).  See especially Brandom (1994, chapter 3).

7  Wherein is Reasoning Social?

135

beliefs or inferences. One must only have a potential to develop, drawing on lower-­ level abilities, higher-level abilities enabling reflection. On Brandom’s account, reflection manifests itself in the mastery and competent use of expressive-logical concepts. But normative-epistemic concepts – e.g., the concept of something being a reason for/against something – are yet other means of reflection. To drive the point home, consider, for example, the following argumentative exchange: A: (P): “Fred is in the kitchen”. B: “Really?” A: (Q): “Because I saw him go there”. B: (R): “But I saw him go to the garden”. A: (S): “No, it wasn’t Fred but John”.

Here, A and B appear to enact the game of giving and asking for reasons, the issue at hand being whether Fred is in the kitchen. They apparently treat each other’s claims, queries or challenges as having the significance of giving, requesting or probing reasons. They do so, though, without explicitly representing rational connections between P, Q, R, S in judgments to this effect: that I saw him go to the kitchen is a reason to conclude that Fred is in the kitchen, etc. Still, this exchange is representative in that it features words or constructions indicating such rational connections.39 Now, A and B might be capable of making them explicit provided they already possess certain sophisticated concepts. This would require metarepresentational skills. However, a young thinker might be able to use and comprehend such constructions in perhaps simpler exchanges, without yet being able to make such explicitating claims or judgments. This being so, it seems implausible to deny such a thinker the sort of ability to track rational-inferential relations between claims, challenges, etc., though he or she might not yet have sophisticated concepts. This account, I submit, is developmentally more plausible than M&S’s metarepresentational account: to possess (rudimentary) skills of reasoning is to display some practical mastery (know how) over such discursive moves and turns in argumentative, explanatory or justificatory exchanges. Metarepresentations are not required. Hence the oversophistication problem can be dealt with. Importantly, this account implies that it is possible to make moves in the space of reasons without actually representing them as such moves. One must only be appropriately sensitive and responsive to reasons in the course of playing the game of giving and asking for reasons. At the most basic level of material inferences, to discriminate that (P) “This is a German Shepherd” is a reason for (Q) “It is a dog”, is to be disposed (in the sense of know how) to characteristic, normatively loaded responses: including rebuking, correcting or challenging one who asserts P but then  Conclusion and premise markers such as “so”, “hence”, “thus”, “because” (etc.) are tailor-made for this task; so, too, are words indicating contrast or disagreement (incompatibility) such as “but”, “no”, etc. Discursive exchanges featuring such devices indicate – without explicitly stating – what the game-players treat as correct inferential moves. See Koreň (2018) and Koreň (forthcoming, Chap. 5) for an elaboration of this thought and its ramifications for the inferentialist conception of logic. 39

136

L. Koreň

refuses to assert Q (upon being properly prompted). If one’s verbal behavior displays such a pattern, we can say that one is in the game of reasoning. In this sense, one does not have to actually entertain a thought to the effect that P is a reason for Q. Likewise for material incompatibilities or inferential moves based on them – e.g., rebuking, correcting or challenging one who utters “It is a German Shepherd”, then uttering “It is a fish”. To discriminate and honor such normative relations in one’s responses of this kind does not require that one actually represent them. In fact, I suspect that the mundane game of giving and asking for reasons proceeds to a large extent without game-players actually entertaining such higher-­order thoughts. Now the problem of vicious regress is premised precisely on the assumption that such higher-order thoughts are involved in every case of genuine reasoning. Because on the inferentialist account advanced here such thoughts are not required, the regress does not get off the ground in the first place. Inferentialists could further expand on this by pointing out that some basic inferential moves are always treated as primitively correct – or the game of reasoning would not get off the ground. Again, however, to discriminate and honor them in practice as correct – including recognizing transgressions, responding to transgressions appropriately, avoiding transgressions  – does not require that one actually entertains the thought to the effect that there is such-and-such rational relation. Having at one’s disposal the right conceptual capacities, one might express such relations  – making them explicit. That does not mean, though, that they thereby cease to be primitive. They continue to be so until properly probed. Though possible in principle, probing them is not a free lunch. To probe them one needs to reason, and the chances are that, with respect to some basic inferential moves, one will hit the bedrock. That is, one will either rely on the very inference that one attempts to probe or on other (explicit or background) assumptions or inferences that are no better off. Admittedly, in some of our more reflective moments, we can say to ourselves or to others something to this effect: P is a good reason for Q, so Q (assuming P). But, in such cases one either takes a move from P to Q to be primitively good – being disposed to make or endorse it even without reflecting upon it. This would be a case of explicitating a norm of reasoning implicitly honored all along. Or one indeed arrives at it via another piece of reasoning. In this case, however, it is arguable that the goodness of some inferential moves is not similarly up for grabs  – being taken for granted and honored in practice.

7.7  The Force of the Better Reason What picture of human reasoning vis-à-vis rationality emerges from my comparison so far? I have already expressed concern that M&S confound the different levels of understanding ourselves when they contend that our sense of control in making up or changing our minds is an illusion. I shall now suggest that the inferentialist account motivates a more balanced picture of human reason in this respect.

7  Wherein is Reasoning Social?

137

From the inferentialist perspective, there is a baseline rationality for both interpersonal and individual thinking and deliberation. In our practices of reasoning certain moves count as primitively correct (if only in a given context of inquiry), determining what counts as a reason for (against) what (if only in that context). To count as a competent game-player, one must recognize and honor them as such. This sets the baseline of rationality relative to which deviations make sense. So if one repeatedly refuses to draw an inference from P and P, then Q to Q (or fails to rebuke or correct someone who refuses to draw such inferences), we would not say that one is “in the conditional reasoning game”. Likewise for basic material-­ inferential moves (though, in their case, the context and collateral assumptions play a much larger role). One must to a considerable extent live up to such standards on pain of being treated as lacking the know-how to play the game in the first place. To posit a baseline rationality is not to deny that we are prone to errors due to laziness, lack of metacognitive self-control, self-serving biases or absence of intersubjective check. Further, it does not surprise that we may fare better in this respect when our reasoning is embedded in the social game of giving and asking for reasons, in which people aim to make their claims, commitments or actions acceptable, this being effectively cross-checked by fellow interlocutors holding different collateral information, thus being capable of adopting different perspectives that might prompt them to critique and probe their claims, etc. After all, one of the core claims of Brandomian inferentialism is that the norms governing rational thinking and acting get instituted through such interactions. From the developmental perspective, it is plausible to assume that our reasoning skills mature as we get increasingly (self-) aware that our performances take place within the practices governed by agent-­ neutral norms, including the recognition that our personal reasoning perspective is one among many and far from being privileged.40 This makes it possible to grasp or simulate different reasoning perspectives, including knowing how to switch between the roles of proponents and interlocutors in the course of discussing an issue at hand with others.41 Importantly, this also balances to some extent the yawning asymmetry in M&S’s account between the biased/lazy proponents and unbiased/demanding interlocutors. For the social norms of rationality governing our games of giving and asking for reasons hardly say that good reasons are those supporting whatever we find believable. Accordingly, to come up with obviously self-serving or poorly thought out justifications or arguments (and to wait for others to point this out, eventually adjusting one’s reasons) might not be the best strategy to pursue in the social game of giving and asking for reasons. Insofar as reason enables us to take into account and respond to “the force of the better reason”, we should be wary of neatly distinguishing its “productive” and “assessing” uses and saying that myside bias is its design  For Brandom at least, objectivity consists in the permanent possibility, present in the perspectival nature of human reasoning, of a difference between how one reasons and how one ought to reason. See Brandom (1994, p. 597). See also Tomasello (2014) for a congenial psychological perspective (in this respect at least). 41  Cf. Tomasello (2014). 40

138

L. Koreň

feature when it produces reasons, whereas objectivity is its design feature when it assesses reasons. Finally, it is worth mentioning that inferentialists may allow that intuitive processes of sorts play a role in reasoning. But they are likely to theorize the relation between reasoning and intuition in a different way than M&S do. I already noted that in reasoning we perforce take certain things for granted, whether in the form of a background, ultimate premises or basic inferential moves. Such attitudes – taking for granted, being confident, trusting – may be called intuitive. Further, we might have intuitions about reasons that reflect a socially mediated tuning of justificatory and argumentative skills. Some issues or problems are more familiar or common for us to reason about with others. Accordingly, we might form quick judgements (intuitions of sorts) about what reasons would be appropriate that facilitate our reasoning with respect (or by analogy) to them. If that is the case, their presence, absence or strength in part helps explain why we sometimes find it easy to produce or evaluate reasons, whilst other times we must try hard or are at a loss. And to the extent that such intuitions reflect social-normative constraints on good reasoning implicit in those practices, they might even contribute to counterbalancing biased responses that interfere with good reasoning. Such sensitivities might be helpful also in individual ratiocination. This, again, suggests a somewhat more balanced picture of our person-level capacity to self-regulate our thinking and decision-making in light of reasons.

7.8  W  hy the Game of Reasoning? Cross-Fertilization of Perspectives So far I have focused on the question of what it takes to reason in terms of the capacities for playing the social game of giving and asking for reasons. Inferentialism, in my view, offers in some respects a more satisfying answer to that question than M&S’s approach. But when it comes to the question Why do we reason?, inferentialists and fellow travellers  have had little to say.42 I think that this is a missed opportunity. Let me therefore outline how inferentialists could address the question. The inferentialist view implies intelligibility of a communicative practice in which participants implicitly treat claim-making utterances as moves in a space of reasons. Why and how could such a practice have evolved? Imagine ancestral critters who start to produce certain kinds of simple utterances in certain kinds of situations that present opportunities or challenges of interest to them. If such verbal responses form regular patterns, critters may come to associate them with  Inspiring attempts in this direction that I know of are made by Beisecker (2013) and Peregrin (2014). My sketchy anthropological myth bellow is indebted to their ideas. I elaborate it in a much greater depth and detail in Koreň (forthcoming). See also Peregrin (this volume) for a sketch of why and how norm-governed social practice in general and norm-governed communicative practices in particular could have emerged. 42

7  Wherein is Reasoning Social?

139

corresponding situations and learn to treat others’ responses as reliable indicators of the latter, thereby serving as each other cognitive proxies. We can further imagine that they start to shape each other verbal dispositions in the direction of conformity by producing certain kinds of negative responses to responses deviating from such beneficial patterns. By the same token, they would be under pressure to coordinate also their second-order responses to others’ responses with the preponderance of such normative attitudes in their group, because coordinated attitudes shape verbal patterns in the mutually beneficial direction. Basic norms implicitly governing their communicative exchange can result from that kind of coordination. Eventually, our critters may sophisticate their repertoire in such a way that they come to treat certain utterances as in order or not depending not only on circumstances but also on other utterances. Thus a more complex system of verbal responses could emerge structured by basic inferential and incompatibility relations. If an utterance of a speaker were assessed as conforming to an expected (sanctioned) pattern, practitioners would practically assess it as appropriate, relying on it as warranting or precluding other utterances on the part of the speaker as well as the audience. If not, they would dissent and treat it as incorrect, thereby putting the speaker under social pressure to correct himself – or else to vindicate his saying by providing reasons of sorts. And basic games of giving and asking for reasons could have been elaborated to allow speakers to coordinate their cognitive responses, as expressed in sayings, and reap even more benefits from their already useful, up-and-running practice: coordinating cooperative ventures, learning from each other, pooling knowledge, making collective plans or decisions, etc.43 Even if our mythical forebears are not yet capable of fully-fledged, self-reflective reasoning, they can discriminate and honor in practice – i.e., in what they do, rather than in what they can state or consciously conceptualize – normative constraints. It is such constraints that establish the primordial, intersubjective space of reasons, being instituted, enforced and updated by interacting agents coordinating and calibrating their performances and normative attitudes. To skillfully navigate their way through the space, agents do not need to “reflect” on its underlying proprieties, wrapping them up into claimable contents so as to make those proprieties themselves into the topic of the game of giving and asking for reasons. They exhibit a pre-reflective, practical-normative sense manifested in their normative attitudes of recognizing and treating (responding to) something as appropriate or not without yet representing it as such. As a later development, ever more self-conscious forms of discourse (and of internalized reasoning) may emerge that make it possible to make explicit, negotiate, eventually revise proprieties or norms of reasoning. This “anthropological myth” is consistent with the speculation that the impetus to the evolution of human social reasoning was a pressure to develop and maintain justificatory and argumentative practices in increasingly hypersocial ancestral niches, in which proto-communication was already in place, fulfilling informative  Non-verbal performances might be treated in the same manner, if subject to similar expectations and attitudes regarding their proprieties. They can also be default accepted if fitting an expected pattern, or else probed, criticized, explained, justified, etc. 43

140

L. Koreň

or other vital coordinating functions. Indeed, inferentialists may  find congenial what M&S say about the effects and benefits of reasoning put to its justificatory uses: By giving reasons to explain and justify yourself, you do several things. You influence the way people read your mind, judge your behavior, and speak of you. You commit yourself by implicitly acknowledging the normative force of the reasons you invoke: you encourage others to expect your future behavior to be guided by similar reasons (and to hold you accountable if it is not). You also indicate that you are likely to evaluate the behavior of others by reasons similar to those you invoke to justify yourself. Finally, you engage in a conversation where others may accept your justifications, question them, and invoke ­reasons of their own, a conversation that should help you coordinate with them and from which shared norms may actually progressively emerge. (M&S 2017, pp. 185–186).

Inferentialists are entitled to take much of this on board. That being said, there is a risk in pitting such social functions against cognitive functions of reasoning if we want to give a plausible naturalistic account of the norms of good reasoning that is consistent with its function. Consider M&S’s view that the function of reasoning (in its argumentative uses) is to persuade interlocutors to accept one’s own view. The problem is that, intuitively, reasoning may function “well” in this sense, but be “bad” in the sense that the premises fail to provide good reasons for the conclusion. Consequently, it is unclear whether and how the norm of correct reasoning can be derived from the proposed function. Further, in so far as Mercier and Sperber submit that vigilant hearers are checking arguments for their coherence or strength – thereby pressing speakers to produce “better” arguments in this sense – they seem to presuppose cognitive desiderata (reliability and coherence) as constraints on the acceptability of something as counting in favor of something else. So a problem lurks here: it seems that norms of good reasoning reflect its function only if that function already reflects certain cognitive desiderata. If a plausible naturalistic account of the norms of good reasoning as reflecting its function is called for, this militates against purely social functions of reasoning. The view sketched in this section may overcome the problem. Functionally, reasoning meets the general need of socially and cognitively interdependent creatures to coordinate and cognitively benefit from their up-an-running (if rudimentary) communicative practices (information sharing, pooling, learning, teaching) in the face of mis-coordination typified by disagreement or lack of agreement due to informational asymmetries. Given this function, and the value of social reasoning so conceived for individuals and groups, norms of good inference (including material) fit quite naturally into the picture as coordinators of such cognitive-communicative practices.

7.9  A Collectivist Anthropological Myth In their own ways, the two approaches I have compared in this study emphasize that the human capacity for reasoning is a matter of being able to play the social game of giving and asking for reasons. By way of conclusion, I will address a potential

7  Wherein is Reasoning Social?

141

challenge confronting both of them. What is interesting about this challenge is the fact that it is not meant to vindicate the traditional, individualistic perspective on reasoning. Rather, it embraces the social perspective but worries that the two approaches being compared are not social enough, or not social in the right way. It might be spelled out as follows. First, it might be claimed that human thinking in general and reasoning in particular presuppose a framework of shared goals or norms of good reasoning that presuppose a we (collective) perspective of sorts. It might then be claimed that the game of giving and asking for reasons, as the two approaches being compared portray it, does not do justice to this fact. I will focus on Michael Tomasello’s version of this challenge, which he indirectly levels by way of commenting on M&S’s approach.44 Tomasello agrees that reasoning is primarily a social activity of producing and assessing reasons. However, he worries that M&S’s account implausibly emphasizes competitive features at the expense of cooperative features. M&S’s approach, though presupposing a cooperative framework (socially living creatures increasingly depending on coordination and cooperation), assumes that the communicative practice of our ancestors featured salient non-cooperative incentives (hence recurring challenges) such as deception or manipulation by means of verbal communication. Therefore, the recipients needed to filter messages by assessing their quality, and the producers needed the art of argument to convince them. Tomasello thinks that this is not likely to have been an endemic problem, because our socially living ancestors increasingly depending on coordination and cooperation had much more incentive to be honest and cooperative than to be manipulative and uncooperative. He proposes that the game of giving and asking for reasons “occurred regularly in collaborative activities” of joint or collective decision-making, bargaining or planning in which the interests of participants are somewhat aligned, though they might have different information and views about what needs to be done, and how, to achieve an overarching goal that they share. To wit: Prototypically, we may imagine as an example collaborative partners—or even a council of elders—attempting to choose a course of action, given that they know together in common ground that multiple courses of action are possible. Given their equal power in their interdependent situation, they cannot just tell the other or others what to do; rather, they must suggest a possible course of action and back it up ... (Tomasello 2014, p. 109)

Tomasello argues that reasoning evolved to facilitate cooperative argumentation and decision-making, which in turn require a sense of shared or common goals of sorts – that is, a form of what he calls shared intentionality (which has joint and collectivist versions). It is thus a much more cooperative ability than M&S would have us assume. Reasoning being paradigmatically a cooperative venture of arriving at a reasonable belief or decision (its shared goal), a corollary is that reasoners do not aim to “win” the argument but rather to reach the best possible result vis-à-vis the shared goal. In a sense, if there is a “winner” or “beneficiary” at all, it is the group (“we”) – and through it, all the parties who contribute to it. For: 44

 Tomasello (2014).

142

L. Koreň

… neither of us wants to convince the other if we are in fact wrong … each would rather lose the argument and eat tonight than win the argument and go hungry. And so a key dimension of our cooperativeness is that we both have agreed ahead of time, implicitly, that we will go in the direction for which there are the “best” reasons. That is what being reasonable is all about. (Tomasello 2014, p. 111)

Reasonability, Tomasello explains, is set by standards shared in a cultural common ground, mutually expected to be honored and eventually enforced upon those who do not live up to them. The norms of human reasoning are thus at least implicitly agreed upon in the community, and individuals provide reasons and justifications as ways of convincing “any rational person.” (Tomasello 2014, p.112)

The idea is that originally cooperative exchanges of reasons (framed under the guise of a shared goal/good) were scaled-up into collectivist forms due to being regulated by communal normative standards (framed under the guise of a collective goal/good). Moreover, because our evolved mindset is fine-tuned for cooperation and collective action, Tomasello thinks that our reasoning is likely to display significant cooperative features. One prediction is that we can expect better reasoning in cooperative contexts structured around a shared goal (including better performance than in adversarial social contexts). Findings like those of Moshman and Geil (1998) about the better performance of groups in Wason’s selection task are pertinent in this respect, since participants were involved in a group task of figuring out the right conclusion. Tomasello’s team also conducted experiments investigating the reasoning skills of children, which indicate a more objective argumentation of children in specifically cooperative contexts (compared to competitive): … when children aim at a ‘good’ solution, rather than ‘winning’, they approach the available evidence in a less biased way and produce more arguments … and did not neglect or wilfully withhold arguments when they were after the best solution for a common goal. Thus, cooperative situations, in which children have the joint goal of winning the game together, provide a more motivating context for argument production. (Domberg et  al. 2018, p. 75)

Now, I sympathize with the idea that our socially and cognitively interdependent ancestors were not primarily motivated to exchange reasons in order to make others accept whatever view they wished, irrespectively of whether good reasons count in its favor (pace M&S). But I am concerned that Tomasello may go too far in the collectivist direction.45 Tomasello might paint too rosy a picture of ourselves – or our ancestors, for that matter – assuming that humans are more cooperatively spirited creatures than might be warranted. I think that M&S rightly take into account that humans have divisive interests and reasoning might accordingly have played an important role in their social propagation, negotiation or reconciliation. Think of sociomoral issues, where different parties might have conflicting perspectives – e.g., on how to divide valuable but scarce resources – and might bring to bear different standards (or ranking 45

 Cf. Tomasello (2014).

7  Wherein is Reasoning Social?

143

thereof) to justify their respective positions or proposals in the public arena. This is partly cooperative, to be sure – after all, they do not just enact the (primate) “right” of the stronger to take the largest share they are in a position to go away with. Still, people might not aim at some communal good in such situations, though they might aim to persuade others that it is in their (common) interest to accept their proposal. To me, it is not all that clear that such cases played a marginal role in human interaction and discourse. Likewise, Tomasello’s proposal might oversimplify the social challenges and opportunities of argumentative reasoning. In a social arena where individual interests and reputation are real factors, one might produce explanations or arguments to win over others for one’s purposes, though this might not always be in their (best) interest. They will accordingly have an incentive to critically evaluate such attempts, precisely because it pays to be alert to this possibility. This in turn gives one an incentive to produce further claims as reasons to convince others that they might place trust in him. Eventually, they may end up accepting the original claim or proposal, and it may well turn out to be to the good of all parties. Still, the dynamics of such exchanges do not resemble those of a joint activity structured around some shared goal, if only because parties might have very different goals to begin with (even if the proponent’s goal might be to win over others for some joint goal). Indeed, the very talk about the proponent staking a claim, an interlocutor probing or challenging it, and the proponent defending it suggests a dialectical or polemical element in reasoning. Further, one might produce justifications to raise or defend one’s social influence, standing, reputation, etc.46 While all these phenomena arguably make sense only in a social structure that perforce involves some cooperative framework, they need not always reflect cooperative motives within a truly joint enterprise directed at some shared goal. Perhaps a somewhat eclectic account might be appealing after all. The core tenet of all social approaches is that reasoning was primarily fine-tuned for interpersonal discourse in the public foro externo, in social situations that might vary in the extent to which they are cooperative. Much depends here on one’s preferred views, ideally disciplined by evidence (if indirect) about the likely evolution and stabilization of human cooperation in general and communication in particular. M&S and Tomasello provide two alternative reconstructions that one might want to pit against each other. A better idea, I propose, is to take them to provide complementary accounts of various social uses of reasons. My favorite view is that original forms of reasoning could have evolved with rudimentary exchanges of reasons, simply because they helped our ancestors to coordinate their verbal dispositions in a way that supported their up-and-running (if simple) communicative practice and so furthered their cognitive interests. At this fundamental level, then, reasoning is a natural elaboration of mutually beneficial communicative practices. At this level, it was in the common interest of our ancestors to coordinate or else they could not benefit. Once in place, however, this practice could have been coopted also for justification, persuasion,

46

 M&S (2017), Henriques (2003, 2011), and Desalles (2009).

144

L. Koreň

collective deliberation, etc. For some of those social exchanges of reasons to have dividends for socially and cognitively interdependent critters, they certainly did not always need to act under the guise of the common good. At any rate, we may leave it to others to speculate which of those social uses of reasoning are more likely to have been exploited in relevant evolutionary environments.

7.10  Conclusion Inferentialism offers a distinctive philosophical elaboration of the idea that reasoning consists in a competence to skillfully play the social game of giving and asking for reasons. M&S offer a distinctive naturalistic elaboration of the same idea. Yet another elaboration of the idea is given by Michael Tomasello. I have attempted to engage inferentialism in a critical but hopefully constructive dialogue with these other approaches and thereby with the cutting-edge empirical research approaching the phenomenon of reasoning from different theoretical perspectives. This is a rare sport in the received inferentialist literature. I hope to have shown that it is worth trying out. Acknowledgments  Work on this study was supported by the joint Lead-Agency research grant between the Austrian Science Foundation (FWF) and the Czech Science Foundation (GAČR), Inferentialism and Collective Intentionality, GF17-33808L. I am grateful to Jaroslav Peregrin, Ulf Hlobil, Preston Stovall and Jesus Zamora Bonilla for their helpful comments on earlier drafts of this study.

Bibliography Beisecker, D. (2013). From the grunts and groans of the cave…. presidential address. Southwest Philosophy Review, 29(1), 1–11. Boghossian, P. (2014). What is inference? Philosophical Studies, 169(1), 1–18. Brandom, R. (1994). Making it explicit: Reasoning, representing, and discursive commitment. Cambridge, MA: Harvard University Press. Brandom, R. (2000). Articulating reasons: An introduction to inferentialism. Cambridge, MA: Harvard University Press. Broome, J. (2013). Rationality through reasoning. Wiley-Blackwell. Broome, J. (2014). Comments on Boghossian. Philosophical Studies, 169(1), 19–25. Carroll, L. (1895). What the tortoise said to Achilles. Mind, 4(14), 278–280. Castelain, T., et al. (2018). Evidence that 2-year-old children are sensitive to argument strength. Infancy, 23(1), 124–135. Chater, N., & Oaksford, M. (2018). The enigma is not entirely dispelled: A review of Mercier and Sperber's The Enigma of Reason. Mind and Language, 33(5), 525–532. Chudnoff, E. (2014). The rational roles of intuitions. In A. R. Booth & D. P. Rowbottom (Eds.), Intuitions (pp. 9–35). Oxford: Oxford University Press. Damasio, A.  R. (1994). Descartes’ error: Emotion, reason and the human brain. New York: Putnam.

7  Wherein is Reasoning Social?

145

Desalles, J.-L. (2009). Why we talk: The evolutionary origins of language (2nd ed.). Oxford: Oxford University Press. Dogramaci, S. (2013). Intuitions for inferences. Philosophical Studies, 165(2), 371–399. Domberg, A., Köymen, B., & Tomasello, M. (2018). Children’s reasoning with peers in cooperative and competitive contexts. The British Journal of Developmental Psychology, 36(1), 64–77. Dutilh Novaes, C. (2018). The enduring enigma of reason. Mind and Language, 33(5), 513–524. Evans, J. S. B. (2010). Thinking twice. New York: Oxford University Press. Frith, C. (2013). Making up the mind: How the brain creates our mental world. Hoboken: Wiley. Haidt, J. (2012). The righteous mind: Why good people are divided by religion and politics. New York: Pantheon. Henriques, G. (2003). The Tree of Knowledge System and the Theoretical Unification of Psychology. Review of General Psychology 7(2), 150–182. Henriques, G. (2011). A new unified theory of psychology. New York: Springer. Heyes, C. (2018). Cognitive gadgets. Cambridge, MA: Harvard University Press. Hlobil, U. (2014). Against Boghossian, Wright and Broome on inference. Philosophical Studies, 167(2), 419–429. Koreň, L. (2018). Propositional contents and the logical space. In O. Beran, V. Kolman, & L. Koreň (Eds.), From rules to meanings: New essays on inferentialism (pp.  196–218). New  York: Routledge. Koreň, L. (2019). Have Mercier and Sperber untied the know of human reason? Inquiry. https:// doi.org/10.1080/0020174X.2019.1684988. Koreň, L. (forthcoming). Practices of reason: Fusing the inferentialist and scientific image. New York: Routledge. McHugh, C., & Way, J. (2016). Against the taking condition. Philosophical Issues, 26(1), 314–331. Mercier, H., & Sperber, D. (2009). Intuitive and reflective inferences. In J.  S. B.  Evans & K. Frankish (Eds.), In two minds (pp. 149–170). New York: Oxford University Press. Mercier, H., & Sperber, D. (2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34(2), 57–74. Mercier, H., & Sperber, D. (2012). Reasoning as a social competence. In H. Landemore & J. Elster (Eds.), Collective wisdom principles and mechanisms (pp. 368–392). New York: Cambridge University Press. Mercier, H., & Sperber, D. (2017). The enigma of reason. Cambridge, MA: Harvard University Press. Mercier, H., & Sperber, D. (2018). Why a modular approach to reason? Mind and Language, 33(5), 533–541. Mercier, H., et al. (2014). Early sensitivity to arguments: How preschoolers weight circular arguments. Journal of Experimental Child Psychology, 125, 102–109. Müller, A. (2018). Reasoning and normative beliefs: Not too sophisticated. Philosophical Explorations, 22(1), 2–15. Moshman, D., & Geil, M. (1998). Collaborative reasoning: Evidence for collective rationality. Thinking & Reasoning, 4((3), 231–248. Peregrin, J. (2014). Inferentialism: Why rules matter. Basingstoke: Palgrave Macmillan. Peregrin, J. (this volume). Normative mindshaping and the normative niche. In L.  Koreň, H.  B. Schmid, P.  Stovall, & L.  Townsend (Eds.), Groups, norms, and practices: Essays on inferentialism and collective intentionality. Cham: Springer. Pettit, P. (2007). Rationality, reasoning and group agency. Dialectica, 61(4), 495–519. Pettit, P. (2016). Broome on reasoning and rule-following. Philosophical Studies, 173(12), 3373–3384. Richard, M. (2020). Is reasoning a form of agency? In M.  Balcerak-Jackson & B.  Balcerak-­ Jackson (Eds.), Reasoning: Essays on theoretical and practical thinking. Oxford: Oxford University Press. Sellars, W. (1954) Some Reflections on Language Games. Philosophy of Science 21(3), 204–228. Sperber, D. (2001). An evolutionary perspective on testimony and argumentation. Philosophical Topics, 29(1/2), 401–413.

146

L. Koreň

Sperber, D., et al. (2010). Epistemic vigilance. Mind & Language, 25(4), 359–393. Sterelny, K. (2018). Why reason? Hugo Mercier’s and Dan Sperber’s The Enigma of Reason: A New Theory of Human Understanding. Mind & Language, 33(5), 502–512. Tomasello, M. (2014). A natural history of human thinking. Cambridge, MA: Harvard University Press. Valaris, M. (2014). Reasoning and regress. Mind, 123(489), 101–127. Wason, P.  C. (1968). Reasoning about a rule. Quarterly Journal of Experimental Psychology, 20(3), 273–281. Wright, C. (2014). Comment on Paul Boghossian, “What Is Inference”. Philosophical Studies, 169(1), 27–37.

Chapter 8

Making Sense of We-Awareness: Experiences, Affordances, and Practices Anna Moltchanova

Abstract  My goal in this paper is to support the view that a certain way of being and acting together with others, we-part engagement in paradigmatically social practices, is among the basic ways of being human. The we-part awareness of multiple first-person loci of experience amounts to the group awareness, or we-­ awareness, that is neither above individuals nor an aggregate of instances of I-awareness in which I-awareness is modified either by some special group-related content or mode. When a group member is we-part aware, they, from their respective first-person perspective, experience the action as part of the collective and not merely as an “I” that adjusts to fit the corresponding social circumstances. We-part aware group members are in an agreement on group membership. The interactions shared as we-parts is one way in which the reciprocal interpretation of one another’s behavior that inferentialism considers the basis of I-Thou sociality may be constructed. My view also challenges inferentialism, because in a non-ideal discourse that is often infused with stereotype-based exclusion and power differentials, the practiced shared perspective may belong only to the agents from the traditionally dominating majority. Looking at which group domain we-part awareness encompasses illuminates that inferentialism needs to consider how a discourse participant’s status can be tainted by implicit bias. Keywords  We-part awareness · We-awareness · Allocentric reference frame · Affordance · We-awareness and disagreement

A. Moltchanova (*) Philosophy Department, Carleton College, Northfield, MN, USA e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_8

147

148

A. Moltchanova

8.1  Introduction Those who wish to talk about collective self-awareness have some explaining to do. How can a phenomenon usually associated with an individual first-person perspective belong to a group, except metaphorically? Given the multiplicity of first-person perspectives and the differences in experience that brings, one may doubt that a group can ever be genuinely “we-aware.” Yet, most individuals participated in activities that could be reported by a participant, in the moment, as “we are [doing X],” where “we” is a subject of action/attention genuinely experienced as collective, not as a stand-in for “you and I.” Having fun with friends, working closely together on a project (like in paired computer programming), or participating in a team sport (sometimes as intense as bobsleigh) are just some examples of such activities. My goal in this paper is to support the view that a certain way of being and acting together with others, we-part engagement in paradigmatically social practices, is among the basic ways of being human. The we-part awareness of multiple first-­ person loci of experience amounts to the group awareness, or we-awareness, that is neither above individuals nor an aggregate of instances of the modified either by some special group-related content or mode I-awareness. Instead, we-awareness is comprised of individual loci of we-part awareness. When a group member is we-­ part aware, they, from their respective first-person perspective, experience the action as part of the collective and not merely as an “I” that adjusts to fit the corresponding social circumstances. We-part aware group members are also in agreement on group membership. The agreement doesn’t have to be explicit and can be based on different things, ranging from a rather strict requirement that each group member includes the same individuals in the “we” (which is necessary for a two-person we-­awareness) to a looser one that there should exist a shared understanding of the terms of membership in the group by each member. In a large group, no member may know who all of the group members are but for we-awareness to be possible, every member would need to agree, if only in very general terms, on who belongs—all employees of a corporation, for example. For the simplicity of exposition in this paper, I will concentrate mostly on two-person groups and thus will impose the strong requirement that the “we” in we-awareness has to be of the same set of individuals for each member. I do show that “the sameness” of the understanding of membership, whether it means an identical set of individual members or the same terms of inclusion, does not require that the content of experience is identical for each member. We-awareness can tolerate disagreement and dissent, like when two friends disagree on how to proceed when doing a project as “us,” or when members of a national group argue about different models of nation-building for “all of us.” My view is importantly different from the subject, mode and content accounts of collective intentionality.1 In contrast to the content and mode accounts, it highlights that when groups of individuals act, their joint actions cannot always be described as reducible to individual “I”s having special kind of intentions, different from those  For a discussion of these types of accounts see Schmitz (2018:137–139).

1

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

149

they have in a solitary action, either in terms of their content or mode. A genuine group-based practice cannot be described through modified intentions of each “I,” such as ‘I intend that we J and you intend that we J and our sub-plans appropriately mesh’ (Bratman 1993: 106; 2014: 14–15) or ‘I form a we-intention based on our having formed a joint intention’ (Tuomela 2005: 330),2 or ‘I represent what I intend to do to contribute to the group goal and represent a belief that other agents are also contributing to the goal’ (Searle 2010: 55–56). On my account, in defining a paradigmatically collective practice, we are not defining what kind of intentionality each I possesses (that is, we are not identifying whether an I has I-intentionality with the collective content or whether the I has we-intentionality). “I” as an agent is taken for granted by the content and mode theorists, but it is not obvious that the locus of first-person perspective that possesses the special intentions characteristic of a genuinely shared action is still an “I,” or that this locus represents other participants as other “I”s. The first-person locus of a group action (L) represents any beliefs concerning the group goals and the status of others in this action as part of L’s functional role in acting jointly with the rest of the group, because L as an actor and its first-person awareness are both engendered in the interaction. Thus I call that which exhibits we-intentionality from its first-person locus of activity as part of the group, a “we-part”. This may sound radical, but it doesn’t preclude that in a different context (perhaps at the same time) the locus of activity can experience as an I (for example, a rower can be we-rowing with their team while experiencing an I-itch on their wrist). Moreover, the locus of activity is not claimed to be forever locked as a we-part. It is a temporary state which a group member automatically enters in group practice, because it is how one participates in this kind of practice. My view also contrasts with the subject accounts in two ways. Although it emphasizes the phenomenal aspects of the group subject, it doesn’t consider group members as representing an ‘undifferentiated sense of us’ (Schmid 2014: 16).3 because the sense of us is differentiated into we-part loci of awareness. A group member acts as a we-part of the we-unit that is engendered in certain group practices and not as an I in social circumstances that, together with other “I”s, while intending things jointly with them, brings about ‘a we-unit irreducible to its individual members’ (Gilbert 1989: 197–8).4 I would agree that collective subjects are ontologically irreducible to individual subjects (or sets of “I”s), but, at least in cooperative group actions, the group’s we-perspective is reducible in the sense that it consists of multiple first-person loci of we-part awareness.

 “A we-intention is a special kind of aim-intention involving that the agent we-intends to … see to it jointly with the others that a certain state or event comes about” (Tuomela 2005: 330). 3  The “sense of us” is the plural equivalent of single pre-reflective self-awareness in the individual mind, which is required for the kind of mental integration that allows a self to relate to the world in an active way (Hans Schmid 2014: 16). 4  Gilbert’s view of the formation of a joint commitment is that it occurs when each member expresses their readiness to be so committed with others under conditions of common knowledge (Gilbert 1989: 197–8, 204–19, 264–273). 2

150

A. Moltchanova

In paradigmatically social practices, we-subject, we-action and we-intention are not separated. For example, individuals participating in a cooperative shared activity create common ground specific to the participants and the activity. If A and B are making camp fire, a log in the vicinity of the camp is firewood, but if B and C are arranging sitting space, the same log is something to sit on. B has different common grounds with A and with C based on their shared activities and will make socially recursive references about their intentions based on the corresponding common ground (Tomasello 2014: 57). B will also represent A’s or C’s perspectives and monitor interactions with them, including second-person self-monitoring, based on the different common grounds. Overall, I focus not only on group members’ intentionality but also their awareness in group actions and claim that we-intentionality can be accompanied by a different kind of awareness, we-part awareness. Depending on the practice they engage in, humans are attuned to their common ground with others differently. They can be we-part aware or socially I-aware. In the former case, “common ground” doesn’t apply in the traditional sense of what individuals come to share as individually self-aware, because we-awareness is not an aggregate of a number of instances of individual I-awareness. For a “we,” common ground is not what “you” and “I” share but what “we,” from our respective loci of we-part awareness, share. My view has both methodological and grounding connections to inferentialism, being in line with its pragmatism. Inferentialism identifies meaning directly with the use and holds that any meaning conferring type of treatment has to be identified with public use (Peregrin 2014: 43–4). Similarly to inferentialism’s treating normative attitudes as primitive practical attitudes that implicitly but not representationally constitute norms of assessment, my view treats the subject of experience as implicitly constructed in its engaging with what the world, in the process of human beings’ coping with it and each other, affords. First-person awareness is implicitly constructed as a we-part by some joint experiences without yet being consciously represented as such, and the reciprocal interpretation of one another’s behavior in linguistic practices may be constructed based on the behavior that is shared as we-­ parts. Inferentialism treats rule-following as a behavioral pattern (Peregrin 2014: 74), developing Wittgenstein’s insight that human shared practices are part of what it is to obey a rule and nothing outside of them serves as the basis for the rule’s correct or incorrect application. “Agreement is exhibited in rule following, but does not ground it” (Goldfarb 1985: 485).5 My view adds to this that the agents of the agreement and their status are constructed by the shared activity and they are not entering it as “I”s that ground the group practice. My focus on group membership points to one potential problem for inferentialism. Given its focus on normative attitudes that  As Wittgenstein notes, rule-following is a shared practice, “…[T]here is a way of grasping a rule which is not an interpretation, but which is exhibited in what we call “obeying the rule” and “going against it” in actual cases. … ‘[O]beying a rule’ is a practice. And to think one is obeying a rule is not to obey a rule. Hence it is not possible to obey a rule ‘privately’: otherwise thinking one was obeying a rule would be the same thing as obeying it. What it means to follow a rule correctly isn’t grounded in anything” (Wittgenstein 1953: 81e). 5

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

151

are grounded in practice of reciprocal interpretation, inferentialism applies perfectly well within the domain of agents who treat each other as parts of the same “we” in this practice. Yet, real-life discourse is often infused with stereotype-based exclusion and power differentials, and the spontaneous and implied shared perspective may belong only to the agents from the traditionally dominating majority. For example, male politicians are prone to treat the assertions of their female colleagues as not authoritative or appropriate to the same extent as the assertions of their male counterparts, even if the content is the same.6 We need to pay attention to the ­discourse participants’ implicit and pre-reflective understanding of what parts construct the “we,” and who is excluded, although formally everyone may be said to belong. Inferentialism needs to consider how implicit bias taints one’s status within, and inclusion into, the group. Below, I provide some evidence from empirical studies supporting the existence of a form of subjectivity that transcends the exclusively individual, and explain in which ways it matters for how individuals experience events together with others. I use the notion of “affordance” from ecological psychology to construct a basic model of we-awareness that explains why the fact that we-subjects comprise several first-person viewpoints doesn’t interfere with their being self-aware as a group (hereafter “we-aware”). I also consider how multiple first-person perspectives on group experience can be aligned, and how group members are able to interact with each other in the context of group practices while remaining we-part aware.

8.2  Situating We-Awareness with Respect to Empirical Studies I consider that in a small group’s face-to-face interaction, for a group to be we-­ aware, group members, from each individual first-person perspective, need to experience the subject as collective and include the same individuals in the “we.” One may claim that one’s experiencing the subject as collective seems fantastical. What appears problematic about claiming that something is experienced as a “we,” is that group bodies and minds are made up of discrete units, even if connected via certain relations. For each unit it is impossible to be aware of what other units of the “we” are experiencing in an unmediated, non-observational way. Hence, genuine we-­ awareness is impossible for a group except in some metaphorical or, perhaps, institutional sense (group members may be “aware” of having decided something (X) if there is a procedure registering that X counts as a decision). Unlike in individual self-awareness that is immediately and non-observationally available based on proprioception, kinesthetic and self-location, this kind of access is not available for

 Good discussions of stereotypes of female politicians can be found at Brown et al. (2018), Koenig et al. (2011), and Schneider and Bos (2014, 2019). 6

152

A. Moltchanova

multiple bodies in group we-awareness, because there is no single phenomenological/spatial vantage point for the evaluation of experience. These concerns may arise because of a temptation to think of we-awareness just as an individual awareness but somehow found at the level of the group (with the whole of the group being the “individual”) or as an aggregate of several instances of individual self-awareness. However, a we-subject may be genuinely irreducible to individual subjects but not because it is above individuals; while it wouldn’t possess a new level of supra-consciousness, it doesn’t have to be identical to a set of individuals experiencing as individuals either; it would be a set of individuals experiencing as “we-parts.” Hence, we-awareness would not have to be aggregated from the instances of individual phenomenal awareness as individual subjects. Moreover, the experiences that individuals are aware of do not need to be identical but undergone as a we-part while all parts of the “we,” in a two-person group, count the same members as part of the “we,” or, in a larger group, agree on the terms of membership. Below I offer some evidence that a first-person experience as a we-part is possible. I clarify what a we-part experience is by relating the findings of some empirical studies to the features of we-part awareness as I define it. Those experiences to which each member of the group may appear to have privileged access—one’s reference frame, proprioception and the mode of access to what it feels like to have an experience—may change for human individuals in the presence of others. The reference frame of individual experience is different in the presence and in the absence of other participants in an activity. For example, one unconsciously switches from processing stimuli in an egocentric reference frame to an allocentric one when one detects that the other person pays attention to the same object (often to be assessed or manipulated). When this happens, individuals start operating within a different spatial orientation in their view of the situation and they process information as if looking at the objects from the perspective that does not originate from where their own body is located. Individuals also start relying on the view of the other person or their help in manipulating objects. Hence, joint attention or a shared action can cause a first-personal shift in individual experience of the environment.7 An action performed in an allocentric reference frame can be hostile, like in boxing, or cooperative as in team rowing. The capacity for having an allocentric reference frame supports that humans possess phenomenological and neurophysiological mechanisms that enable them to experience subjectivity across the spectrum from “I” to “you and I” to “we.” That one’s experience in the presence of others may be of the shared reference frame albeit the experiencer is located at an individual vantage point undermines the claim that we-awareness is impossible due to the access to phenomenal consciousness being tied to an individual location.

 For example, Knoblich et al. observed that joint attention triggered a switch to processing the stimuli within the other’s, allocentric, reference frame (Knoblich et  al. 2011). Bockler et  al. observed that attending to objects together… makes people adopt an allocentric rather than the default egocentric reference frame (Bockler et al. 2011). 7

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

153

As far as one’s body schema (or proprioception) is concerned, the ecological content of perception that specifies one’s own embodied position in the environment doesn’t always coincide with one’s body and can extend beyond it. Ulric Neisser, for example, would include whatever moves with the body as part of this experience of one’s embodied position: a bike, one’s clothing, and so on (Neisser 1988). This view can be expanded to we-subjects. A team trying to control their hurtling down the track in a bobsleigh, or demonstrators linking elbows during a march are just two examples of groups that possess ecological we-awareness. The “we,” experienced within physical and cultural environment in relation to those not participating in the activity, are this group body, singled out by this activity. The experience of the center of [we]-gravity shifts away from the center of an individual body towards the center of the group body. The corresponding “proprioception” of we-subjects is distributed among, and comprised of, several first-person (allocentric) perspectives, in which group members are immediately perceived as parts of “our” embodied position in the environment. Hence, what the environment affords in the presence of others, given how humans are wired, provides a possibility for an immediate experience of a “we” from a first-person locus of awareness. “What it feels like” to have an experience is different in the presence of others as well, because participants in cooperative activities are able to share psychological states. For example, Michael Tomasello and Malinda Carpenter define shared intentionality as a suite of social-cognitive and social-motivational skills that allows for collaborative interactions in which participants share psychological states with one another (Tomasello and Carpenter 2007: 121). They demonstrate that humans are wired for shared intentionality. Tomasello and Carpenter note that both chimpanzees and human infants engage in gaze-following, but only human infants fully share attention with caretakers based on the context of their joint activity. For example, in a hiding-finding game, chimpanzees followed the directionality of the pointing cue, but failed to understand its meaning. They did not understand that the pointing was intended to be ‘relevant’ because they didn’t share the joint attentional frame and didn’t experience the searching as a shared activity.8 Chimpanzees and other apes gesture to manipulate others, but not to inform them of things helpfully or to simply share an experience with them. Humans, however, can share an attentional frame.9  Tomasello and Carpenter point out that human ontogeny contains two trajectories: the general primate line of development and the uniquely human line of development. To compare, while chimpanzees engage in gaze-following, manipulative communication, group action, social learning, children also engage in joint attention, cooperative communication, collaborative action, instructed learning (cultural living) (Tomasello and Carpenter 2007: 122). 9  Tomasello and Carpenter report that while chimpanzees engage in social learning, it is individual for them: being emulation learners, they merely gather information from others. Children are, in addition to emulation, able to imitate and engage in instructed learning that is the basis of cultural living (Tomasello and Carpenter highlight that instruction involves introduction to how “we” do things). Moreover, children often collaborate for the sake of collaboration and want an adult who exited their joint activity to recommit to it (although the activity can be done without the adult) (Tomasello and Carpenter 2007: 123). 8

154

A. Moltchanova

The sharing of psychological states is an important characteristic of group subjects. For example, many individuals lived through a dissolution of a close relationship when, as one of the symptoms of the breaking-up, one of the friends or partners would refuse to offer or accept sharing psychological ground (“I am not interested in your work stories,” “I don’t want to watch this show with you,” for example). The individual resists maintaining common psychological ground by deliberately withdrawing from sharing common psychological states with the other and avoiding experiences with shared attentional frame. For example, Marion Solomon and Stan Tatkin emphasize that in couple therapy one treats not individuals, but the collective, the dyadic system, and they analyze attachment patterns to help reestablish and repair relationships. They write, “[a]ttunement is the feeling of being on the same page, in alignment, or in synchrony—the feeling that launches primary attachment relationships … Misattuned moments … without repair … lead to an amplification of negative experience and a sense of threat” (Solomon and Tatkin 2011: 103–4).10 “We might not ordinarily think of the simple act of walking away or avoiding as a “separation” experience, but psychobiologically, it is equivalent” (Solomon and Tatkin, 73). When the context that enables the practice of being together is interrupted, the collective subject and corresponding we-part awareness are suspended.11 Additional support for the claim that how humans function biologically and socially allows for a spectrum (from individual to collective) of subjectivity is provided by the direct perception theory of social cognition. It defends that individuals can have primitive non-conceptual awareness of others, like in seeing, or perceiving in some other modality, that another is having an emotion or intention, etc. (Krueger and Overgaard 2013: 297). As Gallagher argues, “in seeing the action and expressive movements of the other person in the context of the surrounding world, one already sees their meaning; no inference to a hidden set of mental states (beliefs, desires, etc.) is necessary” (Gallagher 2008: 542). Before we can theorize, we can already interact, and direct perception of what others feel or intend does not require simulation or any other kind of theory.”12 The immediate knowledge of the feeling and intentions of the other may help one not only box efficiently, but also be immediately aware that a teammate is sharing one’s intention to attack the opponent’s net or is upset about losing. Each group  What is more, stonewalling, or refusal to interact, is one of the four predicting factors of divorce that the Gottman Method uses in couple therapy. The repair of relationships, on the other hand, includes, among seven things, an ongoing awareness of a partners’ world, expressing interest or acknowledgement in conversation instead of ignoring, and creating shared meaning. https://www. gottman.com/blog/an-introduction-to-the-gottman-method-of-relationship-therapy/ 11  I should note that members of a couple can have a we-part experience even when apart. Think of a couple shopping for a camping trip in different parts of town. Some of that activity, in light of the shared goal and identity is undergone as a we-part. 12  Gallagher points out that this perception is not a third-person observation, it is something that happens in the context of interaction. “…[T]he articulated neuronal processes that include activation of mirror neurons or shared representations constitute the neural correlates of a non-articulated immediate perception of the other person’s intentional actions, rather than a distinct process of simulating their intentions (Gallagher 2008: 536, 541). 10

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

155

member is we-part aware of what “we” intend or feels, which can be reported as “we want to attack the net,” “we are upset about losing the game.”13 Summing up, various phenomena support that we-awareness, comprised of multiple loci of we-part awareness, is possible. One’s ability to seamlessly come in and out of an allocentric reference frame changes group members’ perception of the world. In their shared practices, group participants may be aware of being part of a group body in a way different from being self-aware when alone, with a group member’s “proprioception” extending beyond one’s body. These group bodies are capable of sharing attentional frames. Engaging in a cooperative practice together constitutes a meaningful activity for humans in a way that differentiates them from apes: in shared intentionality, humans practice a way of being together that doesn’t make it strange for them to be aware and perceive as a we-part. Finally the possibility of immediate knowledge of the intentions and emotions of others facilitates acting as a we-part in order to coordinate with other we-parts. We-awareness, as conceptualized above, is obviously not like I-awareness, because its phenomenology with respect to a group’s location in space, the “proprioception” of the group’s “body” or a we’s access to having an intention, belief or feeling would be all distinct from I-phenomenology. That we-awareness has dispersed phenomenology doesn’t mean, however, that it would be impossible for it to be immediate and not entirely based on observation in each we-part locus of awareness. That group we-awareness can be distributed without being eliminated is the point of the rest of the paper. I will first consider how we-awareness can be comprised of several we-part perspectives and then consider why divergent content of we-part experiences does not interfere with we-awareness in a group practice.

8.3  The Parts of the Whole in We-Awareness The challenge for explaining how genuine we-awareness can exist, as we saw, comes from having to explain how several first-person perspectives somehow organize into producing one we-awareness. It is true that members of the same group, simply by position, have different first-person experiences even if they operate in the allocentric reference frame, have we-proprioception and share intention and feeling in an unmediated way: they can see colors differently, or approach the object  It appears that accepting the conclusions of the theory of direct perception may undermine the possibility of we-awareness as unmediated. However, each member’s perceptual awareness of others can be part of the internal workings of the we-self, like an individual self partially observing itself when, say, seeing one’s hand in drawing. It may also be strange to worry about this similarity here, or that we-awareness has to have the immediate aspect at all, since I mentioned its not being identical with I-awareness. Nevertheless, immediacy is an element that needs to be shared by anything we call “awareness” in the phenomenological sense. Immediacy is what makes self-awareness what it is. The we- and I-awareness phenomenology is distinct (the former consists of we-part aware loci)  but to the extent to which we want to identify something as being aware of itself, immediacy is something that seems essential. 13

156

A. Moltchanova

they are jointly manipulating from different angles, or view a puzzle in different ways (A’s allocentric view of the activity and B’s allocentric view of the activity are different). However, that collective subjects are not the same as I-subjects should not present a problem when each member is aware of the group not as constituted through iterations of individual experiences (“I am aware that you are aware that I am aware, etc.), but immediately as it happens in we-part awareness.14 Moreover, the differences of first-person experience do not need to prevent individuals from forming a “we”: so long as all members of a small-group “we” relate to the environment in the given circumstances as we-parts and include the same individuals in the “we,” the phenomenology of each first-person perspective doesn’t need to be identical. The difference in perspective does place limitations on we-awareness in the following sense. If friend A experiences puzzle-solving as part of a genuinely collective subject (“we are solving the puzzle”), but friend B tries to one-up A while pretending to work together, B and A are not we-aware of solving it. B’s experience is “A and I are competing” (a distributive collective subject) or “I am competing with A (unbeknown to A)” (an individual subject). Perhaps, concurrently, A and B may be a we-subject of some other experience, like being in a room together, if each experiences it as “we are here,” but not of the puzzle-solving together. While an individual can experience a group interaction on a spectrum of first-person experiences, from an individual (A), to a distributive collective (individual as a member of a set, A and B), to a genuine we-subject (“us,” A&B), a tipping point of a set of individuals becoming a “we”/“us” is achieved, in a small informal group, when everyone starts experiencing what is happening as part of a “we”—a possibility that the circumstances presented all along but that is now taken. While it is intuitively clear that an egoistic action of B eliminates a we-subject, what about a case when A thinks there is a we-subject but B is sincerely unaware of that? Imagine an ER waiting room in which A thinks that A and B are waiting together to hear the news about the same patient (although A and B are not acquainted). B is indeed waiting to hear the same news, but B thinks A is waiting for something else. For A, the experience is “We are waiting to hear about X,” but for B, it is “I am waiting to hear about X.” I want to clarify that it is A (who thinks they are waiting to hear about X) who is wrong about there being group we-awareness. It is not the case that B who experiences as an I-subject is wrong about the status of the group. A is feeling as if it were a “we-part.” True, there is something it is like to be “I” in “I am waiting,” and there is something it is like to be “we” (as experienced from a first-person perspective) in “We are waiting,” and A may really feel like A is experiencing the latter, but to be aware of one-self, this “one” has to be aware. The “one” of the “we” is actually two: A and B, and both A and B have to be aware of the “we”  As Tomasello (2014: 38) points out, human individuals are “attuned to the common ground they share with others, and this doesn’t always involve recursive mind reading,…if necessary they may decompose their common ground a few recursive layers deep to ask such things as what [A] thinks I think about [A’s] thinking.” 14

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

157

to be we-aware. There is no we-awareness for A&B without B’s being included as being we-part aware, B isn’t we-part aware and this “we” (A&B) is not we-aware. Hence, it doesn’t seem to be possible for a group to be we-aware when one of the participants doesn’t take the possibility for action as a “we” or doesn’t even perceive the situation as affording a we-experience. So it is A who is mistaken about their being a we-subject: that A is aware of the “we” does nothing for B who is I-aware or “I.” There is no “we” that is we-aware of the “we.” Person A may even be correct in the limited sense that it is possible they be a “we” provided the corresponding we-awareness of the “we” materializes—B will be positioned differently with respect to the environment and A, should B learn more facts about the ER waiting situation. But that won’t happen until both A and B are indeed we-part aware and hence we-aware as a group.15 The question here is not of epistemology but of ontology: a (small face-to-face) group is we-aware when it includes individuals who experience the practice they engaged in as a “we” from their first-person perspective and who agree on membership. In case of the waiting room, B neither is feeling as a “we” not is agreeing on membership. There is an ambiguity in A and B’s disagreement about the subject of awareness: suppose that A from the waiting room is trying to solve a puzzle while waiting and thinks that they (B and A) are doing it together, because B is seated in front of the same table. A can be wrong that they are solving the puzzle (there is no “we”); or that they are solving it (although “we” is still there, “we are waiting”). The existence of a we-awareness will depend on whether, when A feels that they are solving the puzzle, B considers puzzle-solving or any other action to be a group practice. Their being a we-subject of puzzle-solving should involve their both experiencing that they are doing it. Imagine this conversation: A suggests, “we are really close to solving this puzzle”; B replies, “no, we are just waiting, you are solving the puzzle; but hey, we can finish it” and B joins; or, as another option, A may concede, “never mind the puzzle, let’s just wait.” A was wrong that the “we” was doing the puzzle, and A and B adjusted (and at least two ways of doing this were possible). The tipping point at which both individuals become we-part aware of their experience is not set and settled, and in the same experience can be on a spectrum, with the group arriving at and departing from a tipping point for the genuine “we.” The switch between “I”/“you” and “I”/“we” subject can happen automatically, corresponding to the switch between affordances taken. For example, in psychological studies participants automatically switched from individual to joint lifting of passing planks that were increasing in length. The point at which the pairs (pre-­ reflectively) switched to collective action depended on the joint arm-span of the participants (Richardson et al. 2007: 845). At the point of change the experience of the subject became collective (as a “we” or coordinated “you and I”). A doubter may now be convinced that we-awareness is possible, but wonder why the existence of genuine we-awareness matters. For, if you and I are sailing together,  It is part of an agent’s acting that they are aware of the action. Doing something while intending, or thinking that one is doing, something else is a fertile soil for tragedy, precisely because the agent’s self-awareness of the action is integral to it. 15

158

A. Moltchanova

the action outcome is the same whether I sincerely participate as part of a “we” or strategically participate in “you and I” because I depend on you for something not related to sailing. My response is that genuine we-awareness matters for human agency and relationships. Human persons are normally distressed when what appeared or used to be a “we” experience is subsequently uncovered as false. Simply put, individuals do sometimes experience things while we-part aware. As agents, we are the cause of what we understand, practically speaking, and experiencing as a “we,” under certain circumstances, brings genuine we-part  awareness about. Practically and morally, it is part of most agents’ fabric of experiences. The easiest example to illustrate why we-awareness matters as well as the importance of having an account of it is the existence of heartbreak in cases of being betrayed by one’s life or business partner. When A discovers that a partner (B) with whom A felt as “one” was manipulative and strategic in those joint activities in which A was engaged wholeheartedly, as “us,” and in which A believed B to be on board as a “we,” A feels wronged. The strategic stance of the other not only makes the group unstable (the partner can defect at any moment and is not committed to the group project in the same way), but also invalidates the past experience for the deceived counterpart, for whom the genuine “we” appeared real but who now discovers that the kind of practice they engaged in was not what they understood it was. When person A discovers that A experienced the group as a genuine we-subject while person B acted for themselves, A won’t think that, despite B’s deception and manipulation, the “we” was there because A thought it to be there. Instead, A will be aggrieved that it has never been there (A was we-part aware of the imagined “we”). In general, no one is ever deceived into thinking that the “we” was or is still there merely because they continue thinking it was/is there. The fact is that it wasn’t and it won’t. Although the example of heartbreak deals with the emotional importance of shared awareness, the example’s goal is to highlight that both a “we” and we-­ awareness point to an ontological status that is constitutive of some human activities, interactions and collective states.  There is always a danger of being wrong about the subject of experience being a “we,” as other group members may not be we-part aware. There doesn’t need to be a damaged party: in a two-person group, for example, one may not be deceived but faced with a non-strategic partner who acts as an “I.” So far I have established that humans may have a capacity for we-part awareness and discussed the basic coordination of first-person awareness for we-subjects: all group members have to be “on board” in experiencing something as a “we” and including the same (in a small group) individuals in the “we,” or agreeing on the terms of membership (in a larger group). I still need to explain how multiple loci of we-awareness are (not necessarily discursively) coordinated in a group practice to support the complex composition of we-awareness. I especially need to demonstrate how discrepancies in individual experiences can be part of the same group we-awareness. A notion that is helpful for my theory is “affordance.” It doesn’t have a settled upon meaning in scholarship, but I do not discuss which view of affordances is more

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

159

convincing. Instead I elaborate on a recent version of the notion, the “form of life” view of affordances put forward by Erik Rietveld and Julian Kiverstein. I show that a form of life as it pertains to a “we” and to an “I,” even when they engage in the same kind of activity like puzzle solving, should be viewed as two distinct forms of life, because they include different abilities based on different agent-relata. I call a form of life further differentiated with respect to its agent-relatum “practice.” For example, there is a practice of collective puzzle solving and there is a different practice of puzzle solving alone because they involve, based on their corresponding agent-relatum, abilities that do not belong to the same form of life. An “affordance”based view of the phenomenal subject captures how a “we” is formed and dynamically maintained in changing circumstances. In the next section I introduce the notion of “affordance” and then clarify, using this notion, how a group’s we-­ awareness can persist through divergent actions of group members or even differences in their conceptual representations of the events in group practices.

8.4  Affordances and Awareness 8.4.1  Affordances “Affordance” is the central theoretical construct of ecological psychology coined by James Gibson (1979) as a situation-specific possibility for action. A big canoe is not “carry-able” for one person, but it is for two. As Andrea Scarantino (2003: 956) points out, affordance predicates are “time-indexed incomplete predicates, whose completer is a set of background circumstances referring to an organism at a time in a set of environmental circumstances.” Some environments can only provide affordances for groups—a busy street can be cross-able for a group but not for one person; others for either individuals or groups—a puzzle can be see-as-duck-able or see-as-rabbit-able for A alone or for A together with B. While Gibson used “affordance” for perception, we can also theorize cognitive affordances. Scarantino (2003: 960–1) gives an example of a number being divide-­ by-­two-able as a mental affordance, which requires “the language mediated ability to master concepts such as divisible number.” Even a simple affordance like a utensil being eat-with-able depends on how humans are socialized and included in sociocultural environments. Growing up, human beings are constantly exposed to what Lev Vygotsky called “scaffolding,” when their ways of interacting with the world are shaped and assisted by grown-ups, defining their “effectivities” with respect to the environment.16 For example, perceiving the intentional aspects of a bucket for a normally socialized child includes perceiving aspects that are designed for conventional use, which involves

16

 The concept of “scaffolding” is explained in Estany and Martinez (2014: 98–111, 102–103).

160

A. Moltchanova

conventional knowledge.17 Thus, the bucket’s affordance predicates include “crayonstore-­in-able,” “water-carry-in-able,” etc. Humans can also learn symbolic affordances through instruction and imitation: acquiring math skills leads to one’s being able to recognize and compute certain strings of symbols when presented with them in appropriate circumstances, thus being aware of some situations affording problem-solving. The same environment holds different affordances for different individuals based on their history, physique and abilities (including different affordances for different species). In Tomasello and Carpenter’s discussion I mentioned, it is clear that chimpanzees do not consider a bucket as hiding-finding-game-play-with-able, but the bucket offers at least this additional affordance for a human child. The human phenomenological-cognitive-behavioral niche18 as a set of affordances available to humans is shaped and sculpted by the rich variety of social practices humans engage in (Rietveld and Kiverstein 2014: 326). Rietveld and Kiverstein’s view of affordances follows Anthony Chemero’s position that affordances are not properties of the environment but relations between particular aspects of animals and particular aspects of situations. Rietveld and Kiverstein modify Chemero’s view to include a form of life in the view of affordances (Chemero 2003: 184). They define affordances as relations between aspects of a material environment and abilities available in a form of life (Rietveld and Kiverstein 2014: 335). This view allows one to consider symbolic and other such affordances from the human ecological niche not as aspects of conventional knowledge but, first and foremost, as a matter of human practice utilizing abilities that are generally acquired through training and experience in sociocultural practices (Rietveld and Kiverstein 2014: 340). I build my account upon this version of affordance-theory because the notion of “form of life” allows me to model how a group can unreflectively act as a “we” without having to combine two elements—an individual participating in a form of life as an individual while also taking up social affordances of coordination.19 Some sociocultural practices require and train abilities that are irreducibly collective, like solving a puzzle together. One can solve a puzzle alone as well, but that 17  Anika Fiebich (2014: 157), for example, identifies natural, intentional, institutional, and intersubjective affordances. Perceiving institutional affordances involves perceiving an institutional context (and being aware of oneself and others as part of the institution). 18  This is Tony Chemero’s term. According to Chemero (2009: 200–1), an animal’s phenomenological-cognitive-behavioral niche is the set of affordances available to that animal. Affordances are what we perceive, they are the content of experiences; they are also relations between what an animal can do and features of the environment; finally, the perception of affordances is also a relation: it is a relation between the animal and an affordance. He advances phenomenological realism, based on these three features and realism about affordances (realism about what is experienced as such and not just realism about the features of the environment that make up the environmental end of the relation). 19  For Gibson (1979: 128), social affordances are based upon correctly interpreting another’s intentions. “What other persons afford, comprises the whole realm of social significance for human beings.”

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

161

would be a different kind of game. Trivia teams or teams of computer programmers train together because participating in the corresponding practices together requires a different set of abilities compared to participating in them individually. I distinguish, within a “puzzle solving” or “game playing” form of life two distinct practices based on abilities that correspond to what the subject is taking itself to be. Therefore, in addition to being classified by the type of abilities a form of life offers (in human societies all of them socially rooted and trained), I introduce a further classification to the notion of a form of life, by the subject of experience. Because abilities available to one in a form of life depend crucially on whether the subject is individual or collective, we should identify distinct versions of the same form of life: solving puzzles as a group is related to, but is also importantly different from, solving puzzles alone. Thus, there are distinct practices associated with groups and individuals in what Rietveld and Kiverstein would identify as the same form of life.20 “Form of life” and “practice” will be used interchangeably in the rest of the paper provided activities are classified according to the type of ability based on the subject—individual or collective. To expand the Rietveld-Kiverstein view to a group, affordances for a group of people are relations between the abilities a set of individuals has available in a form of life and aspects of a material environment. Affordance predicates signify possibilities for experience and action based on the abilities available in a form of life to an agent (who can be collective): a large canoe is not lift-able for one (dependence on the agent); a heavily loaded weightlifting bar is snatch-able for an Olympic champion but not for an untrained person (dependence on an ability related to the way of life). A view that incorporates the notion of affordance into an explanation of perception and action can account for pre- or unreflective we-awareness in a group experience, a point to which I turn next.

8.4.2  Awareness Awareness of any kind takes place in the context of the affordance that is being realized: every individual perceives and behaves not “in the space and time of physics … but in environments” (Scarantino 2003: 950). A completed affordance is not prior to [self]-awareness (as “we,” or “I,” or “you and I”), and [self]-awareness is not prior to an affordance being realized. Awareness is part of the interaction of individuals and the environment in the process of realizing and completing one of the opportunities for experience and action. One may experience a given situation in

 As an analogy that highlights the possibility of non-overlapping pockets of “users” in a way of life (although not the difference in the type of subject) imagine different skill level leagues in soccer—players of different levels participate in the soccer-playing way of life, but at a different ability level. 20

162

A. Moltchanova

various concurrent ways—scratching an itch as an “I” while solving a puzzle as a “we.” For a continuous we-awareness, affordances in the group practice are perceived/ taken uninterrupted: each member, by continually experiencing them as such, dynamically recommits to the corresponding we-part awareness and the “we” is we-­aware of these affordances. Rietveld and Kiverstein point out that individuals have the potential to engage with affordances skillfully in a form of life and the patterns of behavior that make it up. A particular individual’s actual skilled engagement with an affordance involves unreflective behavior in manipulating their environment/engaging with it, which also has an affective aspect—getting it right (Rietveld and Kiverstein 2014: 335). This can be extended to a group. In a group with a long history (like a group of friends or a bobsleigh team), their participation in a form of life as a group [doing X as a group] has been practiced or trained for so long that group members use the abilities that constitute this practice unreflectively, in a kind of unreflective expert group action,21 at least in habitually encountered circumstances.22 Even with a change in circumstances individuals can re-group and take different affordances by exercising different abilities within the same practice as is called for by the changed circumstances while remaining a “we” without explicit reflection (at least not a reflection on their membership that is built into the exercise of the abilities from the form of life). Remaining within the same group practice in relation to the circumstances, of course, is only possible if affordances alternative to those habitually taken are available in the changed circumstances. It is worth repeating that there isn’t a form of life of being a group in addition to the practice the group is participating in. The members are a group, a “we,” only within, and with respect to, a particular practice, manipulating the environment and engaging with it as a group. For example, at a camp without running water, an environmental triggering cause of one particular group behavior may be that there is no more water in the communal tub. The tub, among other things, is everyone-together-fill-able, the affordance the group members have been realizing from time to time, habitually, all season, as a group. Under normal circumstances the group members operate unreflectively and skillfully based on common ground they established as part of the group’s unique common history of acquiring group know-how in operating the camp. If some of the campers get sick, the healthy individuals would have to regroup and fill the tub by themselves. They are realizing an affordance different from the one that has been  A good account of unreflective action is offered by Rietveld (2008).  Christian List and Philip Pettit’s account of individual motivation can shed light on how a group is represented in each participant’s first-person perspective when they act based on the collective self-awareness of the group. According to List and Pettit, individual attitudes, when they think of the group action in the first-person mode, close the identification gap if individuals are under the automatic guidance of the group: when they act as group members, they do not ask themselves whether they wish to identify with the group and then act if they have this wish. Rather, they can respond “as spontaneously as pilots do when they take their cue from the panel before them. Or at least they may do this where there are no “red lights” that suggest they should hesitate and take stock” (List and Pettit 2011: 192). 21 22

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

163

realized habitually, but still within the same practice (“running the camp together/ camping with friends”), relating their abilities from this form of life to the changed circumstances of the material environment. One may ask (since I tie the group relatum to the realization of affordances from a particular form of life) whether they are a new “we” (composed of healthy campers 1, 3 and 4 and not sick campers 2 and 5 from the initial group of 1-2-3-4-5), or the “we” is still the whole group. To answer this question, let’s consider a sport team (a group with possibly more affinity than a group of campers). Three basketball players on a break-away play do not become a different team—the ball is score-­ with-­able in the context of the whole team playing the game. Moreover, to complicate things, the ball may be, in addition, show-off-with-able for a star player. The solo playing that interferes with the practice of group playing may be judged inappropriate by the group. As Rietveld and Kiverstein (2014: 343) suggest, what determines the appropriateness of an action is whether the taking of an affordance is correct, that is, being in agreement with the communal practice. In the game context, the particular synergy of the three breakaway players, or the virtuosic throw of the star player are nested in the overall set of the abilities offered by the practice of playing basketball as a group in the corresponding environment. The exercise of abilities within this practice requires being on a particular team in a particular game. The virtuosic player or the three breakaway players are unreflectively aware and bound by the constraints on their behavior based on their awareness of membership in the group because they are participating in the corresponding practice.23 It is important to recognize that we-awareness and we-subjects rely on the members’ unreflective acceptance of membership in the same set of individuals (literally or defined as a set by certain characteristics) when they experience an event or participate in a practice as a we-part. 24 A “we” can survive various changes in the environment relatum by remaining the same “we” based on the members’ unreflective we-part awareness of the same set of members in the group as part of the same practice. Whether the campers start functioning as two groups would depend on whether 1-3-4 and 2-5 are using the abilities available in the practice of “camping as a group” as two different groups. An inferentialist framework maps in part on the notion of conceptual affordances, which offers an explanation of how the norm formation that is implicit in a discourse normally takes place. Focusing on affordances not only highlights the habitual behavioral mechanism of implicit normativity but also makes it possible to explain how stereotypes pervading a social environment may lead to implicit bias in the way the assessment of assertions is practiced. Social environment affords the inclusion of some and the exclusion of others in any discourse context; it also  The basketball example is Tony Chemero’s, as well as the notion of nested affordances. (personal conversation). 24  Returning to the discussion of who is mistaken in the ER case (A who thinks a genuine “we” is present or B who doesn’t) we can view what is happening from the point of view of affordances. A’s perceiving a we-affordance and B’s perceiving an I-affordance prevents them from jointly participating in the same version of a form of life and thus sharing the practice of waiting together. 23

164

A. Moltchanova

affords a hierarchical ranking of two propositionally identical assertions based on who produced the assertion and who is in the audience. The complex way in which implicit bias distorts reciprocal assessment problematizes the inferentialist notion of implicit normative attitudes and raises the possibility that the inferentialist view needs to be modified to account for the complexities of implicit bias that alter the domain of discourse. On my view, bias and power differentials may make it impossible for some participants, although they are formally included in a discourse, to be aware as a we-part; or, if they are we-part aware, they, based on how they are perceived due to implicit bias, may be a part not of an I-Thou but of an “I-inferior you” interaction. The discussion to which I turn next clarifies that even disagreements about group experiences can sometimes be an experience of which individuals are we-part aware and which they can settle while being we-part  aware. Below I consider how an “affordance”-based theory can explain the persistence of group we-awareness when a disagreement among group members’ conceptual representations of experience emerge. In this discussion it also becomes clear that how a group engages in a practice is unique to the make-up of the group and that the subject is not separable from a practice in which it engages and vice versa.

8.5  Accounting for Diverging Experiences How can two loci of we-part awareness cohere to form a whole especially when they have to agree upon things beyond the basic pre-reflective “us”? A study of paired programming finds that the participants align their perspectives when they jointly verbally articulate their experience based on their initial understanding of the practice as related to the shared goal. They need to make visible and audible what they can presuppose in the rest of their joint activity. Tenenberg et al. argue that coming into alignment cannot be explained by I-awareness or aggregated I-awareness alone. Programmers use socially recursive inference to establish common ground, which is not a deduction but a process of verification by acknowledging and inquiring. In this process, we-part awareness is not distinct from the experience that is being articulated (Tenenberg et al. 2015). Participants use explicit forms of communication and mutual attention to build common ground in addition to what they already share (prior experience, common culture, the setup for the interaction). An articulation of common ground in alignment belongs to both the speaker and recipient in each assessment/confirmation, offer/acceptance, greeting/response pair of statements. One step to an alignment might be A’s offering an assessment of the situation, B’s replying “bad,” A’s confirming by replying “bad” and nodding. The first “bad” implies that A was heard and offers a further assessment of the state of affairs A is pointing out. The second “bad” confirms an agreement about the assessment (“our experience is of a failure”). The articulation of common ground belongs to both A and B who co-construct (and co-­ own) a shared understanding and evaluation of the situation. Tenenberg et al. (2015:

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

165

section 4.3.1) point out that even the initial coming into alignment is a process in which partners are we-aware (what I call “we-part aware”). The additional common ground—that the situation requires attention—is not a common ground of two agreeing I’s, but of the we-part of locus 1 (A) agreeing with the we-part of locus 2 (B). That is, they agree not as “you (A) and I (B),” but as the “we” (A&B). Coming into an alignment is a dynamic process—participants are relating their abilities from the practice of paired programming to the environment they have been modifying together to increase the number of affordances for joint work. When programmers know they are in alignment, their programing work begins (often no explicit confirmation is needed, they simply arrive at a state that requires less use of recursive evidence and generates fewer demands for evidence to sustain weawareness).25 The authors of the study wanted to utilize the resulting understanding of face-to-face affordances in paired programming in the design of systems that facilitate generic long-distance paired programming. One of the conclusions of the study was that we-awareness is specific to each group, which makes generic system design difficult, because pre-programmed features of virtual environment based on the model group’s history and membership may not facilitate any realizable affordances for other groups. This demonstrates that each collective subject is inseparable from its activity and is unique within the corresponding practice/form of life and problematizes the inferentialist view of normativity in a way different from the effects of implicit bias discussed above—collective acceptance of norms in following rules may be unique to the group and while this wouldn’t preclude theorizing the general mechanism of collective acceptance, the unique group history needs some acknowledgment. Let us now consider how the discrepancy in a conceptual representation of a group practice can coexist with a continuous collective subject. Imagine that two friends who like solving puzzles together just turned a page in a picture book and are looking at the duck-rabbit picture.26 Each sees it for the first time (and hasn’t read about it). Also imagine that “we are looking at a rabbit” for one person and “we are looking at a duck” for the other is the first-person we-part awareness of the experience. Figuring out what’s in the picture is not hard, but can A and B go through this while remaining a “we”? How do A&B view the discrepancy between their assessments of the picture as a “we” and resolve it while remaining a “we”? The person who sees a duck (say, person A) experiences, allocentrically, the duck as seen by “us.” B experiences, allocentrically, the rabbit as seen by “us.” The two different first-person we-perspectives are communicated, which presents the corresponding affordance predicates for the picture, the first pointing out that the picture is together-­ rabbit-­see-able and the second that it is together-duck-see-able. Each representation is created on behalf of the “we.”

25

 The alignment is not established once and for all; it can fall apart and be reestablished.

26

 

The image is from http://www.news.com.au/technology/online/social/can-you-see-a-

duck-or-rabbit-apparently-this-optical-illusion-says-a-lot-about-your-creativity/news-story/ c9682010642308d0c4c528ca10f152e4

166

A. Moltchanova

If “us looking at a duck” that is reported by one we-part aware participant on behalf of the “we” is accepted by the other on behalf of the “we,” while being we-­ part aware, this will enlarge common ground for the “we” that they are building as we-parts. These actions can be part of their unreflective skillful exercise of abilities from the practice of puzzle solving as a group. The linguistic expressions of the experience are reflective, but dealing with them while remaining the “we” is part of the practice of solving puzzles together that is based on their unreflective skillful interaction in the practice. Kjell S.  Johannesenn points out, in discussing Wittgenstein’s notion of “practice,” that practices have a logical grammar that is at least partly expressed by inherited ways of doing things. Practice is inexpressible and the “mastery of a practice is a tacit expression of a kind of understanding … which is shown in correctly executing the sayings and doings that make up the practice” (Johannesenn 1988: 366). I would rephrase Johannesenn’s claim to add that paradigmatically collective practices also have subjects that are at least partly constructed by inherited ways of doing things and these ways are unique to group membership. Disintegrating into individual “I”s with differing views of the picture negotiated from their individual perspectives would be a fine way for the group to proceed. But given the group’s history, it is very possible that they could continue with the practice of collective puzzle solving and try resolving the disagreement as “us,” as a way of having fun together. In relating the abilities present in the practice of solving puzzles together to the environment of the picture, A and B first take two different conceptual affordances on behalf of the “we.” Then they are we-aware that that is the case (since they both own this discovery). Their coming into an alignment is an unreflective skillful use of the material environment of the puzzle by exercising the abilities presented by the practice of solving a puzzle together. The dynamics of being in alignment may involve their both taking the same affordance at the same time—seeing the duck together now and seeing the rabbit together next. Or, it may involve A’s taking the duck affordance and B’s taking the rabbit affordance at the same time on behalf of the “we.”27 Even if A and B always alternate and never see the same image at the same time, they are seeing as parts of the we, and taking these two affordances is exactly what it is, for the “we,” properly to engage with the puzzle. A&B are still we-aware and A and B are each we-part aware. This is the feature of we-part awareness—the “we” can be aware while we-parts are aware of different things as we-parts.28 A disagreement or disparity of representation may be part of a group’s skillful engagement with the environment in any collaborative process of problem solving.  The point of the puzzle is that it cannot be seen as both at a given time by a consciousness, but one of the two viable affordances can be taken at a moment in time. 28  What is more, if only one of them is at first able to make out both animals, it wouldn’t mean their we-awareness necessarily goes away. If A experiences the solution as theirs, on behalf of the “we,” through an offer of this interpretation on behalf of the “we” to B and B’s acceptance of it on behalf of the “we” they would start the process of alignment while remaining we-aware along the way. 27

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

167

Different conceptual representations of what has been happening (the members’ taking or imagining different we-affordances corresponding to the practice of group problem-solving) are valuable for the group’s coming up with a problem-solving strategy.29 Jo Angouri (2012) points out that group problem solving involves disagreement but not of the personal kind, and it is a necessary part of the process of reaching agreement. Contemporary theorists of disagreement often consider it as potentially a group-building experience.30 Even if one or several of the team members made a move in the process that leads to a setback, the group’s problem-solving may retain the stable agent-relatum because, in solving a problem collectively, correcting mistakes is behavior that is part of the process. Mistake correction happens when members adjust their use of various abilities as part of the group if their actions don’t fit the environment in light of the group goal. Collective problem solving, by its nature, includes the process of trial and error in which the affordances members take on behalf of the “we” may diverge. Not all disagreements can be settled while being we-aware. A and B’s puzzle-­ related we-awareness may disintegrate should one of them (say person A) become competitive and perceive seeing hidden shapes as the opportunity to one-up the other. Then A would exit the practice of collective puzzle solving and won’t be participating in problem-solving as a “we-part.” If A plays the game as competitive and not collaborative and realizes the affordance the puzzle presents as alone-seen-as-­ duck-rabbit-able, their cooperative puzzle solving is gone, and they can no longer partake in the practice of solving a puzzle as a group (although A and B may continue to operate within their respective allocentric reference frames for processing stimuli throughout the competition). Both I-affordance and we-affordance predicates that respectively correspond to competitive and cooperative puzzle-solving practices have been present as time-indexed and incomplete. Which affordance is realized depends on which practice A and B find themselves engaged in under the circumstances. Just to clarify, solving the same puzzle alone would be different from either competing or solving it together. At the phenomenological-­ neurophysiological level, the reference frame and proprioception would be egocentric, and in terms of abilities that pertain to the practice of individual puzzle-solving, there would be no enhancement of one’s thinking provided by cooperative suggestions shared through a socially recursive inference, and so on. Summing up, group members may resolve a disagreement or perceived discrepancy of experience on the basis of working out what “we” should be doing or experiencing in the practice of doing the activity together. They achieve this through iterations leading to an alignment, while continuously confirming through  D. Schiffrin argued that disagreement among friends signals intimacy and sociability and may not destroy but strengthen interlocutors’ relationships. Discussed in Sifanou (2012: 1560). 30  Locher and Watts (2005) argue that disagreement can be a valued activity associated with group creativity and better group decisions. Sifanou (2012) points out that different cultures have different views on disagreement and disagreement may indicate interest through involvement in interaction rather than indifference (agreement or silence). 29

168

A. Moltchanova

interaction that they remain, and engendering in this interaction that they are, a wesubject in the practice of finding the solution. Thus, group we-awareness is possible in both disagreement and its resolution.

8.6  Conclusion My goal in this paper was to support that a certain way of being and acting together with others, we-part engagement in paradigmatically social practices, is among the basic ways of being human. I defended that we-awareness is not reducible to individual awareness of individual subjects. It is grounded in the we-part awareness of each locus of first-person perspective of the group. An obvious challenge for someone talking about group we-awareness this way is to explain how it exists while being a seemingly dispersed conscious experience. I answered this challenge in several steps. I first discussed how empirical research supporting the existence of an allocentric reference frame, group proprioception as well as a shared attentional frame and intentionality, substantiates the claim that it is possible for an individual, from this individual’s first-person perspective, to experience an event as a we-part. I next gave an account of group we-awareness and defended that in a small face-­ to-­face group that is we-aware, group members, from each first-person perspective, experience the subject as collective and include the same individuals in the “we.” Finally, I explained how we-awareness can be distributed without being eliminated. I clarified in which way we-awareness can be comprised of several we-part perspectives and explained why divergent content of we-part experiences does not interfere with group we-awareness in a cooperative practice. In my argument, I used the notion of “affordance” and extended a recent version of the notion, the “form of life” view of affordances, to group subjects. I defended that a form of life experienced by an “I” and by a “we” should be viewed as two distinct practices, because they include different abilities based on different agent-relata. I concluded that while a disagreement about who is the subject of experience (like in the misidentification of the subject by one of the group members) eliminates we-awareness, a disagreement on the conceptual representation of an event or the disparity of experience among group members can be part of the group’s normal dynamics in a particular group practice and even result in the group’s enlarging common ground for we-awareness.

Bibliography Angouri, J. (2012). Managing disagreement in problem solving meeting talk. Journal of Pragmatics, 44, 1565–1579.

8  Making Sense of We-Awareness: Experiences, Affordances, and Practices

169

Bockler, A., Knoblich, G., & Sebanz, N. (2011). Giving a helping hand: Effects of joint attention on mental rotation of body parts. Experimental Brain Research, 211, 531–545. Bratman, M. (1993). Shared intention. Ethics, 104(1), 97–113. Bratman, M. (2014). Shared agency: A planning theory of acting together. New  York: Oxford University Press. Brown, E.  R., Phills, C.  E., Mercurio, D.  G., IV, Olah, M., & Veilleux, C.  J. (2018). Ain’t she a woman? How warmth and competence stereotypes about women and female politicians ­contribute to the warmth and competence traits ascribed to individual female politicians. Analyses of Social Issues and Public Policy, 18(1), 105–125. Chemero, A. (2003). An outline of a theory of affordances. Ecological Psychology, 15(2), 181–195. Chemero, A. (2009). Radical embodied cognitive science. Cambridge, MA: The MIT Press. Estany, A., & Martinez, S. (2014). ‘Scaffolding’ and ‘affordance’ as integrative concepts in cognitive sciences. Philosophical Psychology, 27(1), 98–111. Fiebich, A. (2014). Perceiving affordances and social cognition. In M. Gallotti & J. Michael (Eds.), Perspectives on social ontology and social cognition (Studies in the philosophy of sociality 4) (pp. 149–166). Dordrecht: Springer. Gallagher, S. (2008). Direct perception in the intersubjective context. Consciousness and Cognition, 17, 535–543. Gibson, J. J. (1979). The ecological approach to visual perception. Hillsdale: Erlbaum Associates. Gilbert, M. (1989). On social facts. Princeton: Princeton University Press. Goldfarb, W. (1985). Kripke on Wittgenstein on rules. The Journal of Philosophy, 82(9), 471–488. Johannesenn, K. S. (1988). The concept of practice in Wittgenstein’s later philosophy. Inquiry, 31, 357–369. Knoblich, G., Butterfill, S., & Sebanz, N. (2011). Psychological research on joint action: Theory and data. In B. Ross (Ed.), The psychology of learning and motivation (Vol. 54, pp. 59–101). Burlington: Academic. Koenig, A. M., Mitchell, A. A., Eagly, A. H., & Ristikar, T. (2011). Are leader stereotypes masculine? A meta-analysis of three research paradigms. Psychological Bulletin, 137(4), 616–642. Krueger, J., & Overgaard, S. (2013). Seeing subjectivity: Defending a perceptual account of other minds. In S.  Miguens & G.  Preyer (Eds.), Philosophische analyse/philosophical analysis: Consciousness and subjectivity (pp. 297–319). Munchen: Walter de Gruyter. List, C., & Pettit, P. (2011). Group agency: The possibility, design, and status of corporate agents. Oxford: Oxford University Press. Locher, M. A., & Watts, R. J. (2005). Politeness theory and relational work. Journal of Politeness Research, 1, 9–33. Neisser, U. (1988). Five kinds of self knowledge. Philosophical Psychology, 1(1), 35–59. Peregrin, J. (2014). Inferentialism: Why rules matter. Basingstoke: Palgrave Macmillan. Richardson, M. J., Marsh, K. L., Baron, M., & R. M. (2007). Judging and actualizing intrapersonal and interpersonal affordances. Journal of Experimental Psychology, 33(4), 845–859. Rietveld, E. (2008). Situated normativity: The normative aspect of embodied cognition in unreflective action. Mind, 117, 973–1001. Rietveld, E., & Kiverstein, J. (2014). A rich landscape of affordances. Ecological Psychology, 26(4), 325–352. https://doi.org/10.1080/10407413.2014.958035. Scarantino, A. (2003). Affordances explained. Philosophy of Science, 70, 949–961. Schmid, H.  B. (2014). Plural self-awareness. Phenomenology and the Cognitive Sciences, 13(1), 7–24. Schmitz, M. (2018). Co-subjective consciousness constitutes collectives. Journal of Social Philosophy, 49(1), 137–160. Schneider, M.  C., & Bos, A.  L. (2014). Measuring stereotypes of female politicians. Political Psychology, 35(2), 245–266. Schneider, M. C., & Bos, A. L. (2019). The application of social role theory to the study of gender in politics. Advances in Political Psychology, 40(Suppl. 1). https://doi.org/10.1111/pops.12573.

170

A. Moltchanova

Searle, J. R. (2010). Making the social world: The structure of human civilization. Oxford: Oxford University Press. Sifanou, M. (2012). Disagreements, face and politeness. Journal of Pragmatics, 44, 1554–1564. Solomon, M., & Tatkin, S. (2011). Love and war in intimate relationships: Connection, disconnection, and mutual regulation in couple therapy (Norton series on interpersonal neurobiology). New York: W. W. Norton and Company. Tenenberg, J., Roth, W.-M., & Socha, D. (2015). From I-awareness to we-awareness in CSCW. Computer Supported Cooperative Work (CSCW). https://doi.org/10.1007/ s10606-014-9215-0. Tomasello, M. (2014). A natural history of human thinking. Cambridge, MA: Harvard University Press. Tomasello, M., & Carpenter, M. (2007). Shared intentionality. Developmental Science, 10(1), 121–125. Tuomela, R. (2005). We-intentions revisited. Philosophical Studies, 125, 327–369. Wittgenstein, L. (1953). Philosophical investigations. Oxford: Basil Blackwell.

Part III

Themes from Wilfrid Sellars

Chapter 9

Belief Attribution as Indirect Communication Christopher Gauker

Abstract  This paper disputes the widespread assumption that beliefs and desires may be attributed as theoretical entities in the service of the explanation and prediction of human behavior. The literature contains no clear account of how beliefs and desires might generate actions, and there is good reason to deny that principles of rationality generate a choice on the basis of an agent’s beliefs and desires. An alternative conception of beliefs and desires is here introduced, according to which an attribution of belief is, in a certain sense, an assertion on the believer’s behalf and an attribution of desire is, in a certain sense, a command on the desirer’s behalf. In terms of this conception of the attribution of beliefs and desires, we can begin to understand how attributions of beliefs and desires can be explanatory, although we still cannot expect the attribution of beliefs and desires to generate predictions based on an assumption of rationality. Finally, the reality of beliefs and desires is likened to the reality of money; it is the reality of things governed by objective norms. Keywords  Belief · Assertion · Communication · Folk psychology · Objective norms

9.1  Introduction What are we doing when we attribute beliefs and desires? What purpose does it serve to do that? Among analytic philosophers there prevails what I will call the widespread dogma: In attributing beliefs and desires, we are positing unobservable entities in a person for purposes of explaining and predicting his or her behavior. The widespread dogma is supposed to reflect a kind of realism about beliefs and desires. Beliefs and desires are real things. So we should expect to identify them by

C. Gauker (*) Fachbereich Philosophie in der Kultur- und Gesellschaftswissenschaftlichen Fakultät, Universität Salzburg, Salzburg, Austria e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_9

173

174

C. Gauker

locating them in a sequence of causes and effects. That is what we are doing when we say that someone believes something or desires something. It might be protested that in calling this a widespread dogma I am ignoring important developments in recent philosophy of folk psychology. Yes, it might be said, some years ago, in the heyday of Daniel Dennett (1971) and Jerry Fodor (1981), philosophers may have exaggerated the place of attributions of belief and desires in folk psychology, saying that we explain and predict human behavior primarily by attributing mental states such as beliefs and desires. But more recently Heidi Maibom (2007) has emphasized that we can also understand other people in light of the rules and conventions that govern society. To this list, Kristin Andrews (2012) adds stereotypes and character traits. Tadeusz Zawidzki (2013) has observed that the attribution of mental states plays an important role in enculturation. What these authors say may be true, but what they say does not contradict the widespread dogma. We can maintain that the central function of attributing beliefs and desires is the explanation and prediction of behavior, while allowing that we have other means of explaining and predicting behavior as well. One can maintain that the attribution of beliefs and desires serves auxiliary functions, such as enculturation, while maintaining that the practice is structured by the function of providing explanations and predictions of behavior. And indeed if we look closely at the writings of these authors, we find that they do, with various qualifications, accept the widespread dogma.1 I will argue that the widespread dogma is flat-out false. There is some truth in the idea that we can explain people’s behavior in terms of beliefs and desires. I doubt there is much truth in the idea that we can predict people’s behavior in these terms. In any case, the raison d’être for attributions of beliefs and desires lies elsewhere. The attribution of beliefs and desires belongs to a communicative practice that facilitates interpersonal cooperation. In particular, in attributing a belief we, in effect, make an assertion on another person’s behalf. In attributing a desire we, in effect, make a command on another person’s behalf. What exactly I mean by that remains to be explained. Insofar as we can explain behavior in terms of beliefs and desires, we do so only by exploiting these more fundamental practices. First, I will explain why I think the widespread dogma is false. The kind of connection between belief, desire and action that people have supposed there to be  For example, Maibom, defines the “classical model” as holding that “a desire for something combines with a belief concerning how to bring about the desired state of affairs” and thereby “causes” action (2007, pp. 567–568). She then mentions some variants on this model and writes, “Models such as these are extremely powerful although their deployment is limited” (2007, p.  568). Zawidzki nowhere in his book doubts that what he calls “sophisticated mindreading” (2012, e.g., pp. 2, 66) is possible and enables prediction. His thesis is just that it is possible only where a more primitive kind of mindreading has shaped the minds of the members of a society to the extent that sophisticated mindreading becomes possible (pp. 2–3). Of these three authors, probably the most skeptical toward the widespread dogma is Andrews (2012), but even she writes that “Traditional folk psychology certainly has a place in our world” (p. 109). The place she thinks it has is the explanation and prediction of the behavior of a person who is in an “anomalous” situation (pp. 111, 220). 1

9  Belief Attribution as Indirect Communication

175

really just does not exist. Second, I will explain my own conception of the raison d’être for attributions of beliefs and desires, namely, as a kind of indirect communication. Third, I will try to explain how a practice of explanation in terms of beliefs and desires might be built on top of the practice of attributing beliefs and desires for purposes of indirect communication. Finally, I will return to the realism issue and consider in what ways my conception of beliefs and desires can qualify as realistic. The relation between my thesis and Wilfrid Sellars’s philosophy of mind, as presented in Empiricism and the Philosophy of Mind (1956), is complex. On the one hand, it is clear that Sellars conceived of thoughts as theoretical entities that we posit in order to explain and predict behavior, which is precisely the conception that I will criticize. On the other hand, it is also clear in his allegorical story of the genius Jones, who is the first person to offer explanations of people’s behavior in terms of thoughts, that Sellars models thoughts on overt speech. In this essay, I am not concerned with the nature of thoughts as occurrent events (for more on that, see my 2018), but rather with mental states such a belief. But I can claim Sellars as my inspiration to the extent that he conceived of the practice of attributing intentional states as a development in our linguistic practice, in which intentional states generally are conceived as moves in an on-going conversation that thinkers engage in with themselves and others.

9.2  The Yawning Gap If you ask a philosopher how beliefs and desires lead to action, almost all of them will answer thus: People do what they believe will satisfy their desires. Call this the belief-desire law. As formulated here, the belief-desire law is obviously a simplification. This formulation carries the false presupposition that people always have just one desire at a time (or that all of their desires are fulfilled by the same action). I will argue that hardly any improvement is possible.2 A first attempt at a reformulation would be this: People do what they believe will satisfy their strongest desire. But of course, that’s not right either. People do not always do what they believe will satisfy their strongest desire, because they believe they can more probably satisfy some weaker desire. You might prefer a career as a wildlife photographer, but you settle for dentist, because you can see how to make that happen. In place of “the strongest desire,” one could try substituting “the most proximal” desire. What is that? If it’s just the strongest desire, then this is the same proposal as before. If it’s the desire that the person who has it does act on, then the principle that people do what they believe will satisfy their desires becomes useless. We need to see what action a person performs before we can know what the most proximal desire in this sense is. So the principle, so construed, is useless for prediction. But it

 The core of this critique was set out in my 2005.

2

176

C. Gauker

is also useless for explanation. The principle collapses into the principle that, whatever a person takes him- or herself to be doing, he or she wanted to do precisely that. Perhaps that principle distinguishes a human being from a tree or a stone, but beyond that it yields no understanding of why people act as they do. At this point, it should be apparent that if we want to find any truth in the belief-­ desire law, we will have to reformulate it rather radically. Talk of belief and desire, it may be said, is just loose talk. If we want to formulate in a precise way a principle linking beliefs and desires to action, then we have to substitute subjective probability for belief and utility for desire. A vague gesture toward formal decision theory is such a common response to the doubts that I have raised here about the belief-desire law that I think I have to explain with some care why I think it leads nowhere. Suppose there is a certain mutually exclusive set of outcomes, such that an agent must choose between actions that may result in one or another of the members of this set. Call these the basic outcomes for a given decision problem. We can define a lottery over basic outcomes as an assignment of a probability to each of the basic outcomes, such that the probabilities sum to 1. In choosing a course of action, an agent can be thought of as thinking of that action as assigning a probability to each of the basic outcomes. So an action can represented as, in effect, a lottery in this sense and a decision is the selection of such a lottery. Suppose also that we have a measure of the utility of each of the basic outcomes. Then the expected utility of an action (i.e. a lottery) may be calculated as follows: Multiply the utility of each basic outcome by the probability of that outcome. Take the sum of all these products. The result is the expected utility of the action. In these terms, we can formulate a principle linking belief-like states and desire-like states to action thus: People maximize their expected utility. Call this the expected utility formulation of the belief-­ desire law. There may be several problems with this principle, but the one I want to focus on is this: What is utility, really? How are we supposed to assign numbers to basic outcomes in a way that represents their value to us? For some purposes, the amount of money we would pay may be a serviceable representation of value, but that of course will not be serviceable in all cases. The only good definition of utility that has ever been proposed is one that infers utility from choices. Suppose also that the agent ranks all lotteries over basic outcomes, that is, for any two of the lotteries decides whether he or she prefers the first over the second, the second over the first, or is indifferent between them. Moreover, the ranking satisfies a number of rationality constraints. For instance, one of these says that if any two lotteries A and B are alike, except that A has basic outcome M where B has basic outcome N and M is preferred to N, then A is preferred to B. (For a complete list, look for any presentation of the Von Neumann-Morgenstern utility theorem, such as that in Resnik 1987.) If all of the constraints are satisfied, then, it turns out, we can assign numbers to basic outcomes in such a way that one lottery is preferred to another if and only if the first has the higher expected utility. In other words, the calculation of expected utility represents the ranking. In light of this account of utility, we can see that the expected utility formulation of the belief-desire law falls short of expectation. First, each utility scale is defined

9  Belief Attribution as Indirect Communication

177

only for a given decision problem, with its unique array of basic outcomes. Utility does not describe values that can be applied to multiple decision problems over the course of time. Second, utility is defined for a decision problem only on the assumption that the agent’s preferences conform to the rationality conditions. Usually they will not, because we are not completely rational and because we have no settled disposition with regard to the ranking of lotteries. Even if they do, all of the work of making a decision goes into producing the ranking, without which there is no utility scale. So all we are saying in saying that a person maximizes expected utility is that he or she acts on the choice of the top-ranked lottery, of those that represent actions that the agent can perform. The agent does not suffer from an attack of akrasia; his or her motor movements conform to his or her choice. I know of no serious attempts in the literature on intentional attribution to formulate in a satisfactory way a belief-desire law.3 Having glanced at some false starts, I now want to argue that we can be sure in advance that there will be no satisfactory formulation. The assumption underlying the search for such a formulation is that somehow a person’s actions can be rationalized in terms of his or her beliefs and desires, or probabilities and preferences. There may be some sense in which that is so, and I will try below to explain the sense in which it is so. But there can be no satisfactory formulation of the belief-desire law, because it is not the case that a description of a person’s decisions (or of the actions that follow upon the decisions) can be derived from a description of his or her beliefs and desires, not even if we assume that the person conforms to some general principles of rationality. The reason is that we always have several pertinent, but incomparable scales of value that disagree in their ranking of our options. Option A is ranked above option B on scale of value 1 and option B is ranked above option A on scale of value 2, and there is no common currency into which both values 1 and 2 can be translated, that we can use to compare the overall value of A versus B. For example, suppose that there are two paths I can follow in walking from my home to my office on campus. One is a straight route along a street filled with traffic. It’s not a pleasant walk, but it gets me to my office in a minimum amount of time. The other is a meandering route through a park. It is pleasant to walk among the trees and to hear the birds sing, but it takes more time. In choosing between the routes I value time saved. The sooner I get to my office, the more work I can accomplish. But in choosing between the routes I also value pleasure experienced. The straight route is faster and in that way preferable. The route through the park is more pleasant and in that way preferable. But there is no common currency – no happiness units  – such that I can measure both options on both scales, translate both

 On the contrary, the problems I have pointed to are regularly papered over. For example, when Nichols and Stich (2003) develop their hybrid of the theory-theory and the simulation theory, they present an elaborate boxological account of the mental processes that the interpreter is supposed to simulate (p. 40, figure 2.3), which includes a box for beliefs and a box for desire and arrows from these to a box for decision-making, but they explicitly dodge the questions of what the arrows mean and what is supposed to happen in the decision-making box (pp. 121–122). 3

178

C. Gauker

measures into happiness units, take the sum for each option, and then choose that which has the higher sum. There is nothing unusual about this case. It hardly ever happens that we face a choice between two options and cannot find some respect in which the first is better than the second and some respect in which the second is better than the first. The availability of competing scales of values need not produce indecision. We can find these competing scales of value even when it is quite obvious to us which option we prefer. Precisely because the scales of value are incomparable, we cannot say that our choice is dictated by our values. This means that we cannot expect to have a general principle that tells us what we should choose as a consequence of our attitudes. So we cannot expect to be able to explain people’s behavior on the assumption that they conform to such a principle. There is a yawning gap between our actions and our standing attitudes. This has occasionally been denied (e.g., in Chang 2004), but I know of no halfway rigorous methodology for bridging the gap. Naturally, this observation does not rule out methods having a more limited purview, such as methods that dictate our choice on the assumption that our only objective is to make as much money as possible. Here I have been assuming that if there is a way to predict behavior on the basis of attributions of beliefs and desires, then it must be a principle of rationality that connects the beliefs and the desires to the action. One might reply that the principle connecting beliefs and desires to actions is not a principle of rationality but a merely empirical generalization, discoverable either through reflection on our everyday practice or through psychological research. Fine, but then we should at least be able to formulate some hypotheses that we cannot immediately recognize as inadequate or false. I am unaware of any such hypotheses.

9.3  The Communicative Conception I might be waiting for a plane at the gate in the airport and hear a garbled message in a foreign language come over the public address. I could not make out what was being said, but I have the impression it concerns my flight. I observe that the man who has been sitting across from me gets up, goes up to the gate attendant and exchanges words with her. When he comes back to his seat, I lean over and ask, “Was that announcement about the flight to Heathrow?” He replies, “Yes, the plane is delayed.” I ask, “How long will the delay be”. He replies, “Well, I can’t say, but the gate attendant said the plane is not yet on the ground; she thinks it will be on the ground within half an hour.” The story illustrates what I take to be a common occasion for attributing a thought, or a belief to another. One person wishes to assert on behalf of another what he or she is not prepared to assert for him- or herself. In this case, my neighbor in the waiting area is not prepared to take a stand on when the plane will arrive, but he is prepared to tell me what the gate attendant thinks. He knows that that is not a direct answer to my question, but he leaves it to me to decide for myself whether I

9  Belief Attribution as Indirect Communication

179

want to accept what the gate attendant thinks. The objective of my neighbor is not to explain or predict the behavior of the gate attendant. It is to answer my question as well as he can given that he himself is not in a position to make any assertions that would count as answering my question. He does this by, so to speak, asserting on behalf of the gate attendant. He asserts on her behalf, not in the sense that he makes assertions that are accounted as hers, but in the sense that he asserts in her place without asserting the same himself. He does this by attributing to her a thought, the thought that the plane will arrive within half an hour. Attributions of desire are similar. I am in the Milan Cathedral. I observe a commotion near the altar. A stream of tourists passes me by, heading toward the exit. I flag one of them down and ask what is going on. He says, “The guards want us to leave the cathedral right away”. The tourist attributes to the guards a want, a desire. They want us to leave the cathedral. In saying this, the tourist’s objective is not to explain or predict the behavior of the guards. The tourist is in no position himself to tell me to leave. Perhaps he does not know why we are supposed to leave. What the tourist is doing, in effect, is passing on to me, the command, or request, that he received from the guards. The tourist commands on behalf of the guards. The generalization that these examples inspire are these: The communicative conception of attributions of belief: To attribute a belief is to make an assertion on another person’s behalf. The communicative conception of attributions of desire: To attribute a desire is to make a command on another person’s behalf. In the rest of this section, I will leave the case of desire behind and focus on the case of belief. In order to put the communicative conception of belief into perspective, I will begin with a survey of the functions of assertion. These will be functions in the sense of useful effects that we expect assertions to have, such that their regularly (but not always) having those effects is what keeps the practice of making assertions alive in our species. In light of this I can elaborate on the functions of attributions of belief, conceived as assertions on another person’s behalf.4 The primary function of assertion is, of course, to inform. Traditionally, informative communication has been understood as a matter of the speaker’s expressing a thought, that thought’s having a certain content, and that content’s being grasped by the hearer. But we do not need to commit ourselves to this conception of the process in order to accept that by means of assertions we inform. The case can be described in a less theoretically loaded way as follows: The speaker speaks. As a direct result, the hearer becomes disposed to say the same thing, and this has an effect on the hearer’s nonverbal dispositions as well. A second function of assertion is to corroborate the assertions of others. Someone says something. I agree, but I observe that others do not agree with what that person

 Matthew Van Cleave and I (Van Cleave and Gauker 2010) have used this communicative conception of belief and desire to characterize the cognitive development that allows children to master false-belief tasks. 4

180

C. Gauker

has said. (They do not become disposed to say the same thing.) So I add my word to the word of the speaker hoping that the corroboration will persuade them. There is at least one more function of assertions. By making assertions, we classify ourselves. We sometimes make assertions to let people know which camp we belong to. In asserting that p we classify ourselves with the group that asserts that p. President Trump says global warming is a Chinese hoax, and in that way he classifies himself with the donors he wants to appeal to. This is not the primary function of assertion, but some people seem to make assertions for no other reason at all. In light of this taxonomy of assertions, we may taxonomize assertions on behalf of others. For each of the functions of assertions, there is a corresponding function for assertions on another person’s behalf. Assertions on behalf of another might, in the simplest case, take the form of indirect discourse sentences of the form, “S said that p”. I will start by explaining the functions of these. Then I will come back to the case of sentences of the form, “S believes that p”. Thus, one of the functions of asserting on someone’s behalf is to inform on that person’s behalf. To conceive of this, let us contrast two cases. In one case, Black asserts that p to Brown. Brown accepts Black’s assertion and asserts the same to Green. But in another case, Brown is not prepared to simply accept Black’s assertion and so will not make the same assertion to Green. But Brown considers that Green might be prepared to accept Black’s assertion. So Brown stands in for Black in a conversation with Green, not by making the same assertion as Black, and not in the sense of endorsing Black’s assertion, but by making Black’s assertion on Black’s behalf, while making clear that that is what she is doing. Brown makes clear that she is standing in for Black by not simply repeating Black’s assertion but by saying, instead, “Black says that...” followed by the words that Black asserts. (These words may have to be modified to accommodate the fact that they are spoken by Brown at a different time and place. So if Black’s words were “I will be in Chicago tomorrow”, the words that Brown speaks might be, “Black said that he would be in Chicago today”.) Just as we might assert in order to corroborate the assertion of another, we might assert on behalf of another in order to corroborate on behalf of another. Suppose that Black’s assertion that p is not sufficient to persuade Brown. But if Green were to assert that p as well, then Brown would be persuaded. In that case, even if Brown does not hear Green assert that p, Brown might be persuaded if Black not only asserts that p but also asserts that Green says that p. Brown may not be prepared to accept that p simply on the basis of Black’s say-so, but Brown may have no reason to question Black’s assertion that Green says that p. So the combination of Black’s testimony and Black’s assertion on behalf of Green may be sufficient to persuade Brown. Finally, just as an assertion may serve to classify the speaker (as belonging to the group that asserts that p), so too an assertion on behalf of someone else may serve to classify that person. Just as President Trump classifies himself as a climate-­ change denier in asserting that global warming is a Chinese hoax, so too I can classify him as a climate-change denier in asserting on his behalf that global warming is a Chinese hoax, by saying, “Trump says that global warming is a Chinese hoax”.

9  Belief Attribution as Indirect Communication

181

It remains to explain why an assertion on someone’s behalf may take the form “S believes that p” or “S thinks that p” rather than the form “S says that p”. To some extent, this may be just a stylistic variation. We say “S thinks that p” when we could just as well have said “S says that p” (when it’s clear that we had no other reason to put it that way). But in addition, there is the case of gisting. Sometimes we do not want to say, “S says that p”, because the words in “p” are too far in meaning from any words that S actually spoke, although even for “says” we do not require perfect fidelity. For instance, Viktor Orbán might give a speech, in which he implies, without ever saying it, that Hungary would be better off outside of the European Union. In that case, we do not want to say, “Viktor Orbán says that Hungary would be better off outside the EU”. But it might be fair to convey the gist of what he said by saying, “Viktor Orbán thinks that Hungary would be better off outside of the EU”. Another kind of case in which we might prefer “believes” or “thinks” to “says” is that in which someone behaves very much like someone who might say that p, although he or she does not actually say it. For instance, we observe that someone deliberately avoids policemen. We might say, “He thinks he’s going to get arrested.” Call these cases as-if cases. Insofar as the pertinent behavior includes verbal behavior, the as-if cases might shade off into cases of gisting. Finally, there may be cases in which we are justified in forming the hypothesis that someone has said something in “inner speech” that he or she has not said out loud. For instance, while hiking in the mountains in early spring, we observe someone start to climb an icy path on a narrow ridge but then turn back and come down. We say, “He thinks that path is too dangerous”. We can call this the case of hypothetical acts of inner speech. The act of inner speech postulated in such cases is not itself a belief; it is, rather, an occurrent thought. However, the attribution of an occurrent thought may in turn be the occasion for an attribution of a belief.

9.4  T  he Explanatory Function of Attributions of Beliefs and Desires In light of the communicative conception of attributions of beliefs and desires, we can better understand the way in which attributions of beliefs and desires can serve to explain behavior. But before I explain how, I want to dampen expectations. We should not expect that this conception of attributions of beliefs and desires will help us to understand our capacity to predict people’s behavior in terms of beliefs and desires. In fact, I do not believe that we have any such capacity. If you think that we use attributions of belief and desire to predict people’s behavior, then I put to you the following challenge: Tell me about an episode from real life in which you actually succeeded in predicting what somebody would do and in which it seems clear that you succeeded precisely by means of attributing beliefs and desires to them. I have put this challenge to many people. It is remarkable how freely they violate the stipulation that it must be a case from real life. They find it irresistible to exercise their imaginations and dream up cases in which we

182

C. Gauker

surely could have predicted people’s behavior in terms of beliefs and desires. The trouble with made-up cases seems to be that it is all too easy to imagine that you know that nothing but the beliefs and desires you attribute could be relevant to the behavioral outcome. But in fact you never know that. So I require a case from actual history. If the practice of predicting in terms of beliefs and desires is as commonplace as philosophers suppose, then it should be no problem to come up with a real-­ life example. My expectation is that it will almost always turn out that the success of your prediction can equally well be explained in some other way and you really have no good reason to believe that you succeeded on the basis of attributions of beliefs and desires. There are in fact many ways to predict what people will do. One way is straight induction. In the United States people almost always look both ways before crossing the street. I can easily rationalize this behavior in terms of beliefs and desires, but that rationalization is not the basis for my prediction. In the US children are taught to do this and they do it. This is a purely inductive generalization. Although people in Austria have the same reasons as people in the United States to look both ways before crossing the street, they do so only rarely and I have come to no longer expect it. Sometimes we can make predictions on the assumption that certain norms are operative in society, even if we do not have much inductive evidence that they are respected. In Salzburg, most of the people who get on a bus do not buy a ticket in advance, probably because they have a long-term pass. But if you are in town with me as a tourist, then I expect that you will get a ticket before you get on the bus (from the Automat, from the driver, or from me). Possibly there is some kind of inductive generalization underlying this expectation. But equally I expect that friends of mine will conform to basic social rules. Another important basis for forming expectations about human behavior is understanding the consequences of possessing certain skills. If I know that you are a skilled chess player, and if (somehow, given that I am not one) I can see that the best thing for you to do is to take my rook with your bishop, then I can expect that you will do just that. Your skill in playing chess is demonstrated by success in playing chess, but that does not make my prediction a straight induction; it is mediated as well by some strategic thinking on my part. Another important skill is the skill of speaking a language. Many of our predictions of human behavior are grounded in our understanding of the consequences of possessing the skill of speaking a language. If you tell me that you are looking for the St. Peterskirche, and I tell you that you have to turn left at the next corner and then go through the first arch on the right that you come to, then I expect to see you do that. If you are a stranger to me, I have no other basis for my expectation other than the assumption that you speak English and understood what I said. In a lot cases in which we can predict what people will do on some basis other than attributions of beliefs and desires, we can also provide a rationalization for what they do in terms of beliefs and desires (as I will attempt to explain below). So it is easy for philosophers to suppose that the basis for the prediction must be the attributions of beliefs and desires that provide the rationalization. In light of the yawning gap identified in Sect. 9.2 between our actions and our standing attitudes,

9  Belief Attribution as Indirect Communication

183

it should be clear that the attributions of beliefs and desires cannot provide a basis for prediction by way of rationalization. Nonetheless, the fact that we can often rationalize in terms of beliefs and desires behavior that we predicted on some other basis may be one source for the conviction among philosophers that we regularly predict behavior on the basis of attributions of beliefs an desires. Occasionally, attributions of belief and desire may contribute something to a prediction, not by providing a rationalization but by evaluating a variable that figures into an inductive generalization. Beliefs and desires are real states (as I will affirm in the next section); so like anything else that is real, their presence might sometimes serve as a signal for what will happen next. For example, if we know that Matthias attends every lecture in the philosophy department’s lecture series, provided he knows where and when it takes place, then on the basis of this generalization and the fact that he thinks the lecture will take place in room 301 at 3 o’clock, we can predict that Matthias will be in room 301 at 3 o’clock. Likewise, to the extent that societies are governed by norms defined in terms of beliefs and desires, our prediction that a person will act in accordance with a norm may depend on our attributions of beliefs and desires (but I have not been able to think of any clear examples). In any case, these uses of attributions of belief and desire in prediction are not the kinds of predictions that philosophers have supposed to be possible, namely, predictions grounded in rules or generalizations about decision-making based on beliefs and desires. On the one hand, I have suggested that some predictions might be grounded in an attribution of linguistic competence. On the other hand, I have claimed that a central function of attributions of belief is a kind of indirect communication. Should we not find that an assertion or command on another person’s behalf may be the basis for a prediction, just as the fact that a person has said something may be the basis for a prediction? (I thank Ronald Loeffler for raising the question.) I do not want to rule this out a priori. But what more immediately falls out of the combination of these theses is that we might on occasion be able to predict what a person does by positing a covert act of inner speech. For example, if there are two drawers and A says to B, indicating one of the two, “It’s not in that one”, we might imagine that B will say to herself, “It must be in this one” and on that basis predict that she will open the other one. But while, as I have said, the attribution of an act of inner speech may be the occasion for an attribution of a belief, a prediction based on the former is not a prediction based on the latter. Still, I do not want to deny that attributions of beliefs and desires can be explanatory in a way that exceeds their power to facilitate predictions. Our attributions of beliefs and desires are explanatory in a way that rests on our sense of appropriateness in conversation. Our capacity to use a language entails a capacity to distinguish between discourses and conversations that make sense and discourses and conversations that do not make sense. This is a capacity that merits analysis and explanation, but I will not attempt any such theoretical account of it in this paper. It is a capacity that speakers seldom fail to show when they are going about their daily business, but which easily breaks down when people are called upon to address political questions or philosophical questions. It is a capacity that students have to learn to extend to a new intellectual environment when they are asked to write essays in philosophy

184

C. Gauker

courses. We readily tell our students “You’re not making sense”, but we find it difficult to tell them in a positive and general way what making sense consists in. For example, here is a monologue that makes sense (though it’s not interesting): A sensible monologue. It was raining. I had an umbrella. I didn’t get wet. Here is a monologue that does not make sense: A nonsensical monologue. It was raining. I was an attorney. I didn’t get wet. It is tempting to try to say what the difference is between these monologues, but I will not make the attempt, because it would not generalize. For present purposes, the important point is this: We recognize this distinction. We can all draw it more or less reliably, and we can draw it more reliably when talking about some topics rather than others. The topics on which one person may draw it reliably may differ from the topics on which another person may draw it reliably. We can exercise our sense of appropriateness in at least three ways. We exercise it in speaking – by making sensible contributions to conversations and by engaging in sensible monologues. (The monologues may be conceived as internalized conversation. See Gauker 2018.) Second, we exercise it in judging the discourse of others as sensible. We do this both en passant, while engaged in conversation with others, and we do it in a more studied way when we evaluate other people’s speeches and writings. But there is a third possibility. We can exercise our sense of appropriateness in speaking on behalf of others in attributing beliefs and desires. For example, I might carry on the sensible monologue exhibited above, but on behalf of another, if say the following: “He thought it would rain. He didn’t want to get wet. So he took an umbrella.” This is explanatory, but how? Not because some principle of rational thought takes him from his belief that it will rain and his desire not to get wet to the act of taking an umbrella. (Maybe he does not enjoy getting wet, but thinks it’s healthy or wants to impress his friends with his machismo.) Rather, it makes sense – is appropriate – in the manner in which the sensible monologue makes sense. The action of taking the umbrella is explained by being represented in the context of a sensible monologue carried on onbehalf of the agent. Generalizing, an action may be explained in terms of beliefs and desires by being represented in the context of a sensible monologue carried on on behalf the agent by means of attributions of beliefs and desires. Granted, this account of the manner in which attributions of beliefs and desires may be explanatory is promissory to the extent that it employs two crucial elements that I have not put into any kind of theoretical context. The first of the two is the concept of appropriateness, or making sense. The second element is the idea that appropriate, sensible dialogues and monologues can be rewritten as attributions of beliefs and desires.

9  Belief Attribution as Indirect Communication

185

9.5  The Question of Realism I now wish to return to the question of realism. The widespread dogma is motivated by a stance of realism with respect to beliefs and desires. By treating them as theoretical entities that we postulate for the sake of explanation and prediction, we are supposed to take the healthy attitude of realism toward beliefs and desires. Can the communicative conception of belief attributions likewise acknowledge that beliefs are real things? I should emphasize again that my subject is states of belief, not occurrent thoughts. The communicative conception concerns only the attribution of belief states. The communicative conception does not encourage reductive realism about beliefs, that is, the supposition that beliefs are states of the brain, although it does not quite rule that out either. The grounds for attribution I have been discussing just do not seem adequately sensitive to what might be going on in a person’s brain for us to hope for a typewise (or species- or individual-relativized) reduction of belief states to brain states. But the communicative conception of belief attributions raises no doubt against the thesis that occurrent thoughts are token-wise identical with events in the brain. So even reductive realism about occurrent thoughts remains on the table. Beliefs are real things. They are real things, like promises, votes and money. Like all of these things, they have the reality of things governed by objective norms. Objective norms are rules that one is required to follow but which come to one from the outside. One might choose to impose them on oneself, but quite apart from any such choice, they are imposed on one from the outside, by the community to which one belongs, and one really is obligated to conform to them. These rules sometimes speak of entities, entities that would not exist apart from some such body of norms. These entities have the reality of things governed by objective norms. For instance, a fundamental rule governing promising is: One ought to keep one’s promises. Where no such rule is observed (for instance, among the clams in the sea), there are no promises. A possible rule governing voting is: One should not vote twice in the same election. There could be systems of voting that allow some people to vote twice on some issue that especially affects them. But apart from some such rules governing voting, there are no such things as votes. A simple rule governing money (there are many complex ones) is: If something costs five euros, one should not pay less than five euros for it. How two things can both count as promises, or both count as votes, or both count as money, though somewhat different rules pertain to them is a difficult philosophical question, but not one that I will take up here. The case of money is especially interesting. What does it mean to say that I have a certain amount of money, say, 10,000 euros? Money does have a solid form, in paper and coins. But to say that I have 10,000 euros does not mean that there is a stack of bills with my name on it in a vault somewhere. It does not even mean that there is an arrangement of magnetic particles on a server somewhere that records my possession of 10,000 euros. Even if the bank’s records were destroyed, I could

186

C. Gauker

maintain that my money still exists, although the bank cannot locate it. I might even be able to persuade a judge that the bank has lost my money (perhaps on the basis of some documents I received from the bank). My 10,000 euros exist, but the bank cannot find them. Money is objectively real. My 10,000 euros might exist, though everyone is wrong about how much I have. It might exist even if I too lose track of it. It exists in the case where both my bank and I have lost track of it, but my nephew comes to me on my deathbed and asks me to sign off on the transfer to him of the 10,000 euros I did not know I had. So it exists also in the case in which my nephew never tracks down my money. But there is no such thing as money apart from the social acceptance of the norms governing attributions of quantities of money to people and institutions. Similarly, beliefs have the reality of things governed by objective norms. We might agree to this while disagreeing over the content of the objective norms that govern them. My own account derives from the communicative conception of the attribution of beliefs that I have been adumbrating in this essay. The central objective norm governing the attribution of belief might be formulated thus: One may attribute to S the belief that p whenever doing so serves either as a) indirect information, or b) indirect corroboration, or c) categorization of S, and either 1) doing so is a stylistic variation on an attribution of saying that p, or 2) the proposition that p is the gist of other things that S has said, or 3) S acts in the manner of one who says that p, or 4) the positing of an act of inner speech of saying that p is warranted by S’s behavior. This is a norm of permission, detailing the circumstances under which one may attribute a belief. The first half (alternatives a–c) describes the motives that might move a person to attribute a belief. The supposition is not that speakers are free to decide when an attribution serves as indirect information, etc. Whether attributing to S the belief that p serves as indirect information depends on what S has said and done. The second half of the norm (alternatives 1–4) sets out some constraints. If none of them is satisfied, then it might be better to attribute a saying to S rather than a belief. Condition 1 among the constraints, concerning stylistic variation, might be easily satisfied, but it is not trivial either. It is satisfied only where it is clear that “believe” is not preferred to “says” for some other reason. Condition 4 presupposes that the positing of an act of inner speech can be warranted, and thereby presupposes that acts of inner speech do occur. The reality of acts of inner speech is not that of things governed by objective norms. They might have as well the reality of physical events. But the beliefs governed by the central objective norm governing beliefs are not themselves identifiable with acts of inner speech. In discussing the case of money, I emphasized that things possessing the reality of objective norms may exist though no one knows about them. Does the reality of belief extend as far as that, on my account? Yes, I think it does. There will be circumstances under which a belief might very well have been attributed, in conformity to

9  Belief Attribution as Indirect Communication

187

the central norm governing the attribution of belief, though none is actually attributed, and afterwards there may be no occasion for anyone to think of it again.

9.6  Conclusion In sum, I have tried to render plausible the following claims: Beliefs are not theoretical entities that we posit in order to explain and predict people’s behavior. Rather, to attribute a belief is to make an assertion on another person’s behalf. Although it is not the primary function of attributions of belief to provide explanations of behavior, they can do that too. But we hardly, if ever, predict what people will do on the basis of their beliefs. Nonetheless, beliefs are real things. Like money, they have the reality of things governed by objective norms. Acknowledgements  This publication puts into print the content of a lecture I have been giving occasionally for several years, first in 2012 at the University of Parma, the University of Salzburg, and the University of Johannesburg, then again in 2015 at the University of Graz, and finally in 2017 at the workshop at the University of Vienna that was the impetus for this volume. I thank Ronald Loeffler, Ladislav Koreň and Preston Stovall for their comments and questions on a draft of the written version.

Bibliography Andrews, K. (2012). Do apes read minds? Toward a new folk psychology. Cambridge, MA: MIT Press. Chang, R. (2004). All things considered. In J. Hawthorne (Ed.), Philosophical perspectives, Vol. 18: Ethics (pp. 1–22). Oxford: Blackwell. Davidson, D. (1974). Belief and the basis of meaning. Synthese, 27, 309–323. Dennett, D. (1971). The intentional stance. Journal of Philosophy, 68, 87–106. Dennett, D. (1981). True believers: The intentional strategy and why it works. In A. F. Heath (Ed.), Scientific explanation (pp. 53–75). Oxford: Oxford University Press. Fodor, J.  A. (1981). Three cheers for propositional attitudes. In Representations: philosophical essays on the foundations of cognitive science (pp. 100–123). Cambridge, MA: MIT Press. Gauker, C. (2005). The belief-desire law. Facta Philosophica, 7, 121–144. Gauker, C. (2018). Inner speech as the internalization of outer speech. In P. Langland-Hassan & A. Vicente (Eds.), Inner speech: New voices (pp. 53–77). Oxford: Oxford University Press. Maibom, H. (2007). Social systems. Philosophical Psychology, 20, 557–578. Nichols, S., & Stich, S. (2003). Mindreading. Oxford: Oxford University Press. Resnik, M.  D. (1987). Choices: An introduction to decision theory. Minneapolis: University of Minnesota Press. Sellars, W. (1956). Empiricism and the philosophy of mind. In H.  Feigl & M.  Scriven (Eds.), Minnesota studies in the philosophy of science 1: The foundations of science and the concepts of psychology and psychoanalysis (pp. 254–329). Minneapolis: University of Minnesota Press. Van Cleave, M., & Gauker, C. (2010). Linguistic practice and false-belief tasks. Mind and Language, 25, 290–328. Zawidzki, T.  W. (2013). Mindshaping: A new framework for understanding human cognition. Cambridge, MA: MIT Press.

Chapter 10

Sellars on Rational Agency as Presupposing Collective Attitudes Jeremy Randel Koons

Abstract  Sellars inherits the Kantian idea that “to say that man is a rational animal, is to say that man is a creature not of habits, but of rules” (“Language, Rules and Behavior,” p. 298). For familiar Wittgensteinian reasons, conceptual activity cannot be rule-following activity. Sellars therefore articulates a via media between rule-­ following activity and mere rule-conforming behavior—something he calls pattern-­ governed behavior. The task is to explain how pattern-governed behavior is related to rules without being rule-following—and without assimilating it to mere rule-­ conforming behavior. The rules in question are those making explicit the language transitions that constitute the meaning of linguistic tokens. A key thing distinguishing pattern-governed behavior from mere rule-conforming behavior is a linguistic ‘superstructure’ containing, among other things, a metalanguage comprising these rules—material rules of inference making explicit normative proprieties implicit in pattern-governed behavior. Further, these norms implicit in pattern-governed behavior presuppose the existence of collective attitudes. It thus emerges that pattern-governed behavior must be understood as behavior in the we-mode in the strong sense (to use Tuomela’s language). Finally, the material rules of inference in the metalanguage are expressed using we-intentions; and so the very thing that separates pattern-governed from mere rule-conforming behavior—a metalanguage in which these rules are explicitly formulated—presupposes collective attitudes. Thus, on several different levels, pattern-­governed behavior (and therefore conceptual activity—perhaps even sapience itself) presupposes collective attitudes. I conclude with some reflections on how these considerations point us toward an argument for the reality of the Sellarsian moral community. Keywords  Sellars, Wilfrid · Rules · Collective attitudes · Pattern-governed behavior

J. R. Koons (*) Philosophy Department, Georgetown University in Qatar, Doha, Qatar e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_10

189

190

J. R. Koons

10.1  Introduction Sellars inherits the Kantian idea that “the conceptual is the rule-governed” (Lance 2000, p. 118). Sellars writes, “When God created Adam, he whispered in his ear, ‘In all contexts of action you will recognize rules, if only the rule to grope for rules to recognize. When you cease to recognize rules, you will walk on four feet’” (1980d, p. 298).1 For familiar Wittgensteinian reasons, conceptual activity cannot be rule-­ following activity—at least not ‘all the way down.’ Thus, Sellars must find a via media between rule-following activity and mere rule-conforming behavior (as is displayed by a thermometer, a parrot, or even by iron rusting in the presence of oxygen). Sellars (1963d) articulates precisely such a middle way—something he calls pattern-governed behavior. But what, precisely, is the relation between pattern-­governed behavior and rules? I hope to accomplish several tasks. First, I will rehearse Sellars’s argument that the rules in question are those making explicit the language transitions that constitute the meaning of linguistic tokens—the heart of Sellars’s inferentialism. Second, I will argue that a key thing distinguishing pattern-governed behavior from mere rule-conforming behavior is a linguistic ‘superstructure’ containing, among other things, a metalanguage comprising these aforementioned rules—the rules that make explicit the normative proprieties implicit in the pattern-governed behavior. Third, I will argue that these norms which are implicit in pattern-governed behavior presuppose the existence of collective attitudes, and that pattern-governed behavior must therefore be understood as behavior in the we-mode in the strong sense (to use Tuomela’s language, to be elucidated in the course of the essay). Fourth, I will argue that the material rules of inference in the metalanguage are expressed using we-­ intentions; and so the very thing that separates pattern-governed from mere rule-­ conforming behavior—a metalanguage in which these rules are explicitly formulated—presupposes collective attitudes. Finally, we will see how Sellarsian ‘trainers’—which are another key part of his solution to the problem of pattern-­ governed behavior—help cement the pattern-governed behavior/rules/collective attitude nexus. Thus, on several different levels, pattern-governed behavior (and therefore conceptual activity—perhaps even sapience itself) presupposes collective attitudes. I conclude with some reflections on how these considerations point us toward an argument for the reality of the Sellarsian moral community.

 For Sellars, of course, these are defeasible rules—a position that allows him to evade many of the criticisms of rule-based philosophical systems of the last few decades (such as the particularist challenge in ethics). I develop these themes at greater length in my (2019). 1

10  Sellars on Rational Agency as Presupposing Collective Attitudes

191

10.2  The Conceptual as the Rule-Governed Sellars knew of the difficulties created by making rules so central to his philosophical picture, and rehearses something like Wittgenstein’s regress argument at the beginning of his (1963d, §§1–2/p. 300). Using the word ‘red’ cannot be a matter of following a rule (e.g., “Say ‘red’ in response to red objects”), because this rule deploys the word ‘red’ in a way that presupposes mastery of the concept. If one has not already mastered the concept red, then one will need a second rule, telling one how to apply the concept red as it makes an appearance in the first rule. And our regress of rules is off and running. Sellars wrote a number of papers (e.g., 1963d, 1980d) attempting to resolve this dilemma: How can conceptual activity be rule-­ governed, if understanding such activity in terms of rule-following leads to a vicious regress? The answer cannot be that to be bound by norms is merely to conform to the relevant set of rules. A thermostat conforms to a rule, as does iron rusting in the presence of oxygen. But these items are not bound by norms (at least not in the same sense that rational agents are), and they certainly are not sapient (i.e., reasoning) beings. Fortunately, these are not our only two options. Sellars writes, We have tacitly accepted a dichotomy between (a) merely conforming to rules: doing A in C, A′ in C′, etc., where these doings “just happen” to contribute to the realization of a complex pattern; (b) obeying rules: doing A in C, A′ in C′, etc., with the intention of fulfilling the demands of an envisaged system of rules. But surely this is a false dichotomy! (1963d, §12/p. 303)

There is a third option: that of pattern-governed behavior. Such behavior occurs because of a rule (and so not merely in conformity with a rule), but not because the agent in question is explicitly following a rule. The key is to outline this via media between rule-conforming and rule-following behavior.2 Pattern-governed behavior is, firstly, conditioned behavior: “Certainly, we learn habits of response to our environment in a way which is essentially identical with that in which the dog learns to sit up when I snap my fingers” (1980d, p. 137). Thus, our ability to respond to red objects with the judgment, “Lo, a red object!” is not explained in terms of following a rule, but rather in terms of conditioning, which piggybacks on top of a pre-existing ability to respond differentially to elements in our environment—an ability we share with non-sapient animals, and indeed with some non-intelligent devices. One sees a parallel with Wittgenstein’s response to his own regress argument: “there is a way of grasping a rule which is not an interpretation” (Wittgenstein 1958, §201/p. 81). Wittgenstein offers an analysis of what Sellars would call ‘tied behavior’ that is parallel to Sellars’s, writing that when we follow a sign-post, we do not interpret it; we merely react as we have been trained  For a detailed discussion of what, for Sellars, distinguishes pattern-governed behavior from mere rule-conforming behavior, see Wolf and Koons (2016, pp. 242–50). 2

192

J. R. Koons

to do: “I have been trained to react to this sign in a particular way, and now I do so react to it” (1958, §198/p. 80).3 If this were the whole story, then the distinction between pattern-governed and merely rule-conforming behavior would collapse. To reinstate this distinction, we must answer the following questions: What are these rules that are so central to Sellars’s picture? How are these rules related to pattern-governed behavior in a way that differentiates it from mere rule-conforming behavior? And what is the relation between these rules, this pattern-governed behavior, and collective attitudes? One terminological issue needs to be resolved before I begin. Sellars lumps together under the heading of ‘pattern-governed behavior’ two very different kinds of behavior. On the one hand, he includes behavior by non-agents that has been selectively reinforced by evolutionary pressures, such as the dance of the bees (1963d §§13–14/pp. 325–6). On the other hand, he also includes behavior that while in itself is conditioned (and, as we will see, not actions in the full sense) is in the space of reasons and (although not rule-obeying) happens—in a sense to be explained—because of rules. We might call the former pattern-governed1 behavior, and the latter—the kind displayed by rational agents—pattern-governed2 behavior. This suggests that a more perspicuous rendering of the different types of behavior to be explained has us distinguish between pattern-governed1 behavior and pattern-­ governed2 behavior, with rule-obeying behavior as a subset of the latter, and all three types of behavior as subsets of rule-conforming behavior4:

 Of course, all of this conditioned pattern-governed behavior rests on—and is integrated into— pattern-governed behaviors which exist at the purely physiological level. Our ‘second nature’ (to borrow McDowell’s phrase) needs to be seen as connected to and arising in connection with our ‘first nature’—our non-normative, physical ability to respond differentially to elements in our environment, an ability that has an evolutionary rather than a normative explanation. The melding of our first and second natures, and the corresponding integration of the normative agent into its physical environment, is crucial for Sellars. As Preston Stovall writes, “Simple neurophysiological systems evolved so as to be disposed toward various reflex actions tracking states of the world and tending toward particular responses to it. In the individual organism these tendencies to take up habits are shaped by the environment over the course of its life, just as the neurological capacities of species are shaped by evolution. In our species the advent of the (initially implicit) rules of criticism that form the foundation of a community’s practices of linguistic education allow our reflexive dispositions to be appropriated by a training program that causes them to have the conceptual content they do in virtue of the roles they come to play within the sociolinguistic institutions of our communities.” (Stovall 2016, p. 628). It is this integration of the normative into the causal-genealogical story that allows Sellars to explain the relation between the normative and the natural. 4  Preston Stovall suggested that Sellars’s terminology here needed clarifying, and suggested that the different types of behaviors could be seen as forming nested categories as follows. 3

10  Sellars on Rational Agency as Presupposing Collective Attitudes

193

1 . Rule-conforming behavior: e.g., iron rusting in the presence of oxygen 2. Pattern-governed1 behavior: e.g., the dance of the bees 3. Pattern-governed2 behavior: e.g., intelligent behavior (such as perceptual takings and inference) that are not instances of rule-obeying behavior, but bear some relation to rules (to be determined in the course of this paper) 4. Rule-obeying behavior: Behavior that is consciously rule-following (i.e., behavior performed with conscious reference to the rules dictating the behavior in question) Sellars’s articulation of pattern-governed2 behavior is key to resolving the problem of this paper. Thus, throughout the paper I will simply refer to ‘pattern-­governed behavior’ with the understanding that what is at issue is pattern-governed2 behavior.

10.3  Conceptual Content and Material Rules of Inference Crucial to the following argument is an understanding of Sellars’s famous taxonomy of ‘language transitions.’ There are language-entry transitions, as when photons strike my retina and I form the belief that there is an orange pencil in front of me. These take one from a “situation which is not a position in the [language] game [to] a situation which is a position in the game” (1963d, §22/p. 329). Second, there are language-language transitions—chiefly, inferences—as when I transition from the belief that the pencil is orange to the belief that the pencil is colored. Such transitions take one from one position in the language game to another position in the language game. Finally, there are language-departure transitions, as when I say to myself, “I shall raise my hand now,” and then raise my hand. These transitions take

194

J. R. Koons

one from a position in the language game to a behavior which is not a position in the game. Let us briefly confine our discussion to descriptive terms. Sellars holds that language-­language moves—paradigmatically, inferences—are central to meaning.5 Sellars’s argument for why material inference rules are essential to language (in that they determine the meaning of a language’s descriptive terms) defies short summary, although it may be familiar to most readers of this article. Sellars writes, “Everyone would admit that the notion of a language which enables one to state matters of fact but does not permit argument, explanation, in short reason-giving, in accordance with the principles of formal logic, is a chimera. It is essential to the understanding of scientific reasoning to realize that the notion of a language which enables one to state empirical matters of fact but contains no material moves is equally chimerical” (1963d, §81/p. 355). One of Sellars’s earliest commitments is that a concept is defined by the inferential proprieties governing its use. Thus, if you claim or believe something is copper, you are committed to its being metal. If you believe something (a tone) is A above middle C, you are committed to its not being red (synesthetes excluded, I suppose). As Sellars argues (1980c), if these ‘material invariances’ didn’t hold, then these universals (‘red,’ ‘copper,’ etc.) simply would be different universals. The laws that hold with respect to descriptive terms make them the descriptive terms they are. We use material rules of inference—“If x is A above middle C, then x is not colored”—to make explicit the norms that a person is committed to in deploying a descriptive concept. For Sellars, these norms—expressed by material rules of inference—constitute (at least in part) the meaning of descriptive terms.6 To resume the thread of our argument: Our question was, “What distinguishes pattern-governed from mere rule-conforming behavior?” The answer is given by reference to the material rules that express the content of the descriptive (and other) terms in the language. But what, precisely, is the relation between pattern-governed behavior and the material rules expressing the norms implicit in this behavior? To avoid the Wittgensteinian regress outlined above, Sellars insists that the fundamental account we must give of the above sorts of language transitions cannot see them as rule-following, and indeed cannot even regard them as actions. Thus, the material rules of inference express the rules we would be following, if we were following rules—but pattern-governed behavior still cannot be seen as rule-following. So while we have outlined the kind of rules that Sellars has claimed are central to agency, we haven’t yet uncovered the relation between these rules and

 There is actually dispute in the Sellarsian community as to whether Sellars held that only language-language transitions constitute meaning, or whether meaning is constituted by all three kinds of transitions. It is not necessary to settle that dispute here. 6  It will be noted that for Sellars, meaning is hardly a priori, but consists of inferential connections that are, more often than not, discovered empirically. Also, for Sellars, the meaning-constituting inferential proprieties are roughly those that are the most counterfactually-robust. Thus, “If p is a cat, then p is a mammal” is central to the meaning of ‘cat’; arguably, “If p is a cat, then p is afraid of cucumbers” is not. 5

10  Sellars on Rational Agency as Presupposing Collective Attitudes

195

pattern-governed behavior. It is to this task that we now turn. Having done so, we can then finally turn to perhaps the central task of the paper: demonstrating how these rules and the pattern-governed behavior they make possible both, in different and overlapping ways, constitutively depend on collective attitudes.

10.4  Trainers, Metalanguage and Pattern-Governed Behavior 10.4.1  Trainers and Trainees As noted above, Sellars avoids the regress of rules argument by describing all three kinds of transitions—language-entry, language-language, and language-­departure— as conditioned: “To learn pattern-governed behavior is to become conditioned to arrange perceptible elements into patterns and to form these, in turn, into more complex patterns and sequences of patterns” (1963d, §17/p. 327). The first step in delineating a special category of pattern-governed behavior is to explain how pattern-­governed behavior, while not rule-following, nevertheless happens because of a rule and is explained by a rule. To do this, Sellars first introduces a distinction between two types of rules: He distinguishes ought-to-do’s, or rules of action; and ought-to-be’s, or rules of criticism. Rules of action are self-explanatory—they are rules about what an agent ought to do—rules of the form, “Do A, when in C.” Rules of criticism are rules about how things ought to be, and while they can apply to agents, they can also apply to objects, or states of the world, or whatever. Ought-to-be’s (unlike ought-to-do’s) often concern things that cannot be directly willed. Thus, to offer one of Sellars’s examples, “It ought to be the case that people feel sympathy for the bereaved.” Now, feeling sympathy is not something that people can will. But ought-to-be’s imply ought-to-do’s: that is, they imply actions that will bring about the state of affairs that is the object of the ought-to-be. Thus, we can try to (for example) raise children with the appropriate emotional sensitivity so that they feel sympathy toward the bereaved. This distinction between types of rules is related to a second distinction, between trainers and trainees. It ought to be the case that little Timmy applies the word ‘red’ to red objects. However, obviously, (1) little Timmy cannot be following a rule in applying the word ‘red’ to red objects (for all the reasons outlined above); and (2) in most cases the non-inferential judgment little Timmy forms in response to stimuli is non-voluntary, and so not a matter of his choosing. Therefore, the people whose job it is to teach little Timmy to use the English language (‘trainers’, Sellars calls them; see 1980a, 4.III.24–30/pp. 85–7) train or condition Timmy to respond to red objects with the word ‘red.’ Thus, trainers (who are able to follow ought-to-do’s) are able to condition a language learner’s behavior so that he conforms to ought-to-be’s.

196

J. R. Koons

These rules that govern Timmy’s actions are embodied in the intentions of the trainers. Corresponding to the two types of rules outlined above are two types of intentions. One is an intention to do something, expressed as a ‘shall-do,’ e.g., “Shall (I do A)”. Second, there is an intention that some state of affairs obtain, expressed as a ‘shall-be’, e.g., “Shall-be (The world is at peace)”. Timmy’s trainers recognize the intention: Shall be [Language-learners respond to red objects with the word ‘red’]

Again, Sellars thinks that ought-to-be’s can only be realized if connected to ought-­to-­do’s; and so Timmy’s trainers can reason to the following ought-to-do: Shall [I correct Timmy if he responds to a red object with the word ‘blue’]

Eventually, little Timmy will also be able to follow ought-to-do’s himself. Crucially, even though little Timmy’s behavior is not rule-following (and, Sellars would say, consists of acts rather than actions), it is connected to rules in a number of ways. As Sellars writes, “Pattern governed behavior…is the concept of behavior which exhibits a pattern, not because it is brought about by the [agent’s] intention that it exhibit this pattern, but because the propensity to emit behavior of the pattern has been selectively reinforced, and the propensity to emit behavior which does not conform to this pattern selectively extinguished” (1974c, p. 423).7 And the trainers selectively reinforce and extinguish behavior in trainees because they are following ought-to-do’s. This is a first and fundamental connection between pattern-governed behavior and rules, for Sellars. As Sellars writes, The basic point to bear in mind is that a piece of pattern-governed behavior is as such not an action…Yet it is covered by a rule and, indeed, a rule which is involved in the explanation of its occurrence. The rule which directly covers it is, however, an ought-to-be, and it is involved in the explanation by virtue of the fact that it was envisaged by the trainers who assisted the speaker in acquiring his linguistic ability. Trainees conform to ought-to-be’s because trainers obey corresponding ought-to-do’s. (1974c, p. 423; 1980a, 4.III.29–30/p. 87).

We will have more to say about these rules, the intentions of trainers, and the connection of both to collective attitudes. But first, we must turn our attention to another, critical connection between pattern-governed behavior and rules.

 As described, pattern-governed behavior also encompasses (for example) evolutionarily-selected patterns and behaviors. And indeed, Sellars uses the outcomes of evolution to illustrate patterngoverned behavior; see (1963d, §§13–14/pp. 325–6). This, of course, makes it urgent to find what distinguishes sapient pattern-governed behavior from non-sapient (or even non-sentient) patterngoverned behavior. 7

10  Sellars on Rational Agency as Presupposing Collective Attitudes

197

10.4.2  Pattern-Governed Behavior and the Metalanguage Eliding various complexities (which Sellars works through in [1953]), material rules of inference convey that one sentence is inferable from another. Thus, if I say, “If it were to rain, then the streets would be wet,” I am conveying that the sentence “The streets will be wet” is inferable from “It will rain.” In Sellars’s rather loose sense, then, such talk is metalinguistic, in that it allows us to talk about linguistic expressions and hence make explicit the norms governing relations among such expressions, norms that are generally implicit in our practice. Such talk is often covertly metalinguistic, for Sellars, in that although it occurs in the ‘material mode’ (i.e., we seem to be talking about objects, e.g., “If it rains, the streets will be wet”), what we are actually doing is metalinguistic, in that we are issuing a license to infer one claim from another claim. Sellars’s idea that such apparently object-language talk is really metalinguistic (in his particular sense) should be kept in mind throughout the following. And although I have been talking about descriptive terms, the view outlined here applies to all sentences asserting material rules of inference, whether they concern descriptive terms, or whether they express normative inferential proprieties concerning non-descriptive language. These tools allow us to prise further apart pattern-governed behavior and mere rule-conforming behavior. Pattern-governed behavior requires that the assessors— those following the ought-to-do’s—possess the ability to access the metalanguage, where they can explicitly formulate as explicit rules the norms implicit in their practice. Let me explain: In (1980d), Sellars wrote that “[A]bove the foundation of man’s learned responses to environmental stimuli—let us call this his tied behavior—there towers a superstructure of more or less developed systems of rule-­ regulated symbol activity which constitutes man’s intellectual vision…Such symbol activity may well be characterized as free—by which, of course, I do not mean uncaused—in contrast to the behavior that is learned as a dog learns to sit up, or a white rat to run a maze” (1980d, pp. 137–9). It is this ‘superstructure’ that I have been characterizing above. For Sellars, ‘free’ doesn’t mean uncaused—for, as we saw, Sellars thinks that all language transitions and moves are conditioned. ‘Free’ is better understood as revisable in response to reasons. That is, although our practice consists of various conditioned responses which could be represented in a metalanguage by “formation and transformation rules” which express what the agent would be following if she were obeying rules (1963d, §17/p. 327), the crucial point is that using our metalanguage, we can represent these behaviors and subject them to rational scrutiny and (if necessary) revision. That is, we are not simply stuck with our practices as is—we can revise them, if and when necessary, in response to reasons. As McDowell notes, “‘Responsiveness to reasons’ is a good gloss on one notion of freedom” (1994, p. xxiii).8

 For more on how the ability to represent rules allows for autonomous self-governance by individuals and revision of norms by communities, see Stovall (forthcoming). 8

198

J. R. Koons

For us to be free and autonomous in this way—for us to be able to submit the commitments implicit in our practice to rational scrutiny—we need to be able to formulate these commitments in a way that allows for their discussion. This cannot be done without at least a minimal, Sellarsian metalanguage in which we make explicit, in the form of rules, the norms governing our language moves and transitions. If we want to be able to engage in the revision of our linguistic practices (revision which is necessary to the rationality of such practice9) we need to be able to formulate such claims as, “She is clearly a witch, because she has a wart on her nose”, a claim which conveys commitment to a material rule of inference—a commitment which can now be defended or challenged. As Sellars argues, although we speak in the ‘material mode’ of objects or events being related in law-like ways (“If she has a wart on her nose, then she is a witch”), such talk is covertly metalinguistic in that it asserts an inference license—it makes explicit that one claim is inferable from another. Thus, in engaging in the game of giving and asking for reasons, we make evidential claims which state explicitly inferential relations between claims. You can disagree with someone without moving to the metalanguage. You can assert that she is a witch, and I can assert that she is not. But we cannot argue about who is right and settle on the truth without moving to the metalanguage. For as soon as you say, “She is a witch because she has a wart on her nose,” you have expressed a commitment to the following (defeasible) material rule of inference: “S has a wart on her nose” implies “S is a witch”

And for me to argue that you are wrong (“She can’t be a witch, because she’s the vicar’s wife”), I must express a different chain of reasoning, which will take the form of a different (defeasible) rule of inference: “S is the vicar’s wife” implies “S is not a witch”

Thus, there is no rational interpersonal resolution of disputes—no communal revision of norms based on reasons and evidence—without a Sellarsian metalanguage. More fundamentally, though, to the extent that we understand agency as being in the space of reasons, we cannot treat as an agent in the full Sellarsian sense someone who cannot access the metalanguage.10 This is related to the point just discussed about resolution of disputes. To be sure, Sellars’s picture of language presupposes that we can condition individuals to proper linguistic behavior. Thus, we can condition Smith to respond to A’s by uttering ‘A’, B’s by uttering ‘B’, etc. Surely, Smith needn’t be able to access a metalanguage to count as doing this, correct? But let us complicate the picture: Suppose Smith is somehow exhibiting an improper differential response. He is responding to A’s with linguistic utterances of ‘B’, rather than

 Sellars famously claims that revisability is central to rationality; see (1963b, §38/p. 170). For a detailed argument that revisability (rather than alternate, Cartesian notions like certainty, infallibility, or possessing foundations) is central to rationality, see Wolf and Koons (2016, pp. 16–21). 10  Jaroslav Peregrin pushed back on my claim that pattern-governed behavior requires that agents possess a metalanguage in which they can make explicit the transitions implicit in their patterngoverned behavior; the following paragraphs are meant to address his objection. 9

10  Sellars on Rational Agency as Presupposing Collective Attitudes

199

‘A’. Can we correct Smith without accessing the metalanguage? It would certainly be easier to correct Smith, if we had access to a metalanguage: We could tell Smith to respond to A’s by uttering ‘A’. But this doesn’t show that we need a metalanguage to correct Smith. Let us grant (for the sake of argument) that we don’t need the metalanguage as yet: Suppose we can simply engage in more conditioning to elicit the proper response from Smith, much as you would condition a desired differential response in an animal. But let us complicate the picture by one more degree—and demonstrate where the metalanguage is in fact necessary. Suppose we wish to correct Smith—but he balks, on the ground that he disagrees with us, and believes that his antecedent usage is correct. Can we resolve this disagreement without accessing a Sellarsian metalanguage? If we want to engage in rational resolution of this dispute, we need to make claims that are metalinguistic, in Sellars’s sense. Thus, if Smith wants to explain why we are wrong, he will need to say something like, “This is a donkey (and not a mule) because of its long ears and dorsal stripe”—thereby expressing an inference license between two claims. This is therefore a metalinguistic utterance. Thus, there can be no rational resolution of disagreements, no rational settling of disputes about usage, without the ability to make metalinguistic utterances. (Indeed, I think Sellars’s view here might be more radical. I suspect that on Sellars’s account, for Smith even to express his disagreement—to say, for example, “No, these are donkeys [i.e., and not mules]”—is to express a rule that [in Smith’s view] ought to govern language-entry transitions. But I won’t insist on this more radical claim.) Thus, the superstructure (both the language-language transitions, and the metalinguistic expressions which convey commitment to various material rules of inference) make possible a number of elements which are central to agency. First, they make possible conceptual content, which (in Sellars’s view) is not possible for creatures who are capable only of language entry-transitions.11 Also, by (second) making possible the explicit formulation of the material rules of inference which are implicit in our linguistic practice, they make possible (third) the game of giving and asking for reasons, which in turn makes possible (fourth) the revisability of our linguistic, scientific and moral practices—a revisability which is central to rationality. What is the relation between material rules of inference and collective attitudes? We risk getting ahead of ourselves. Let us first introduce the notion of collective attitudes. We will then be able to see how these are intrinsically tied both to pattern-­ governed behavior and material rules of inference (and, via these, to the very notion of sapience).

11

 See, for example, Sellars (1957, §108/pp. 306–7).

200

J. R. Koons

10.5  We-Attitudes, Strong and Weak So we see one kind of relation between practices and rules. What is the relation between these items and collective attitudes? Let us begin by noting that natural language contains many instances of non-distributive plural predication, where we predicate something of a group of objects that cannot be predicated of any member of that group. For example, one can assert: We surrounded the building. The rocks rained down. The toys were scattered across the floor.

Other natural-language examples seem to attribute actions to groups—actions not obviously attributable to individual members of the group (e.g., “The Senate passed the measure,” “The team won the game,” etc.). As Frederick Stoutland notes, it is characteristic of such groups that they can perform actions that individuals cannot perform: “The senate does things individual senators cannot do, like pass laws or issue a resolution. The university appoints a new president, which no member can do” (Stoutland 1997, p. 47). Given that we do attribute actions to groups; and given that these are often actions that, prima facie, we cannot attribute to any individual member of the group; it remains to understand what it is for a group to act together— and what it is to act as a member of a group. For Gilbert, the notion of acting together—acting as a group—is built around the idea of joint commitment. Joint commitment is a type of group attitude, and contrasts with personal commitment, as when a person acting alone makes a personal decision. Gilbert writes, “two or more people are acting together if they are jointly committed to espousing as a body a certain goal, and each one is acting in a way appropriate to the achievement of that goal, where each one is doing this in light of the fact that he or she is subject to a joint commitment to espouse the goal in question as a body” (Gilbert 2014, p. 34). Joint commitments imply individual commitments. However, these are not personal commitments; they are what Gilbert calls ‘dependent individual commitments,’ and they differ from personal commitments in that the former derive from joint commitments. A personal commitment is one that, for example, might arise from an individual decision—say, Andrea’s decision to have lunch at the Nutmeg Restaurant today (Gilbert 2014, p. 31). A joint commitment, on the other hand, is a commitment of two or more people. Thus, if Heinrich and Andrea “jointly commit to espouse a certain goal as a body…[then their] joint commitment is not composed of a personal commitment of Heinrich’s and a personal commitment of Andrea’s. Rather, it is their commitment, the commitment of the two of them” (Gilbert 2014, p. 31). Tuomela concurs, writing that “the intention to perform one’s part has ‘holistic’ content as it involves joint action. This personal (but still ‘nonprivate’) intention is based on the joint intention in question” (Tuomela 2002, p.  20). When Tuomela describes this derivative intention as ‘personal but nonprivate’, this should be read

10  Sellars on Rational Agency as Presupposing Collective Attitudes

201

as equivalent to Gilbert’s ‘individual but not personal’. For reasons of clarity, I will use Gilbert’s terminology. Gilbert (2014, p. 263) gives an example that can help us understand the relation between collective attitudes, dependent individual attitudes, and personal attitudes. In Gilbert’s example, it is the view of the Joneses that Johnnie must be home by midnight. Now this is not the personal view of either Mr. Jones (who doesn’t see why Johnnie shouldn’t come home whenever he pleases), nor of Mrs. Jones (who thinks Johnnie should be home by ten). But, as Gilbert writes, “the Joneses have discussed the matter and compromised. They have arrived at what they may properly characterize as ‘our view’—not his, nor hers, but ‘ours,’ and in this case, ours only. Each understands that it is now incumbent upon each one to express ‘their’ view (that is, the compromise view) in front of Johnnie” (2014, p. 263). Ergo, this is the collective attitude of the Joneses. Further, Mr. and Mrs. Jones each have a dependent individual attitude (which is realized in their individual interactions with Johnnie)—but this attitude is parasitic upon, and cannot be understood apart from, their collective attitude. Finally, their dependent individual attitudes cannot be understood as mere personal attitudes, since (again) the collective view of Mr. and Mrs. Jones does not reflect the personal view of either. Thus, collective attitudes are not reducible to an agglomeration of personal attitudes. As Olen and Turner write, there is in Sellars a categorical difference between “the genuinely collective [versus] the merely enumerative fact of various individuals having parallel intentions or beliefs” (2015, p. 964, emphasis added). Let us turn to characteristics of we-intentions, or we-attitudes more broadly. Tuomela has a helpful botanization of different kinds of intention. First, one may have a strictly personal (I-mode) intention, as when I intend to go for a walk today. Second, one can act as a group member in a weak (I-mode) sense. Central to the notion of any group is an ethos, which Tuomela defines as “the set of constitutive goals, values, beliefs, standards, norms, practices, and/or traditions that give the group motivating reasons for action” (2007, p. 16). Tuomela writes that acting as a group member in the weak sense “is based on the group members’ sharing the ethos but being only privately (and not collectively) committed to it” (2007, p. 29), and gives the following example: “Thus there may be a group of persons who grow flowers in the village commons and intend to make their small village look beautiful in this and perhaps other ways. Each of them is only privately committed to making the village beautiful, and they mutually know or believe that the others are similarly committed” (2007, p.29). Finally, one can act as a group member in a strong (we-­ mode) sense. One does this, roughly, when one acts because of a group reason. Tuomela gives the following general definition of a we-attitude (in the strong sense): A we-attitude in its strongest, full-blown sense is defined as follows for a person, say x: (a) x has ATT(p) and (b) he believes also that the others in the group, g, have ATT(p) and he also (c) believes (or at least is disposed to believe) that it is mutually believed…that the members have ATT(p). (2002, p. 23)

202

J. R. Koons

I note two things about Tuomela’s account. First, mutual belief is essential to having a group attitude. For ATT(p) to be a truly collective attitude, it is not enough that each of us individually hold ATT(p); rather, each of us must believe that the others hold it, too. Second, group action requires acting for a group reason. For Tuomela, this roughly means acting on an attitude with content p, on the belief that others are acting on this attitude with this same content, and the belief that it is mutually believed that members have this attitude with this content and are disposed to act on it. To illustrate acting for a group reason, Tuomela gives the following example: We can have p = “Our club house is beautifully decorated” and ATT = want. When some persons act for (because of) this we-want to have the house beautifully decorated, each person has a composite full reason because of the collective end in question. This full reason “internalistically” described consists of his wanting to have the house beautifully decorated and his believing that also the others want so and his also believing that it is mutually believed that the members want to have the house beautifully decorated. (Tuomela 2002, p. 23)

Tuomela thinks that acting in full-blown we-mode conceptually requires acting for a group reason—acting on a we-attitude—but argues that this group reason need not be a psychological or motivational reality. Rather, the social reason can “function as a kind of presuppositional and unreflected ground serving to characterize the very activity in question” (2002, p.  88). Thus, for example, when discussing the social constitution of money, Tuomela notes that when using coins to pay you for something, I must recognize coins as money, and I must believe that you recognize coins as money, and I must believe that you believe that I recognize coins as money. But these are conceptual presuppositions of the social institution of money, and need not be consciously entertained by me during the transaction. Gilbert’s account of collective action and collective belief is to my mind less ‘intellectualist’ than Tuomela’s in ways that will emerge as the essay progresses. Gilbert doesn’t rely on the notion of mutual belief, nor on the notion of acting on a group reason. As noted above, Gilbert defines group action such that it involves joint commitment. She doesn’t explicitly parse ‘commitment’ in terms of belief, which allows for a more expansive view of what commitment entails. Nor does she quite say that joint action involves acting for a group reason (rather than merely acting on a joint commitment). What, precisely, is joint commitment? Gilbert writes, A given joint commitment can always be described in a sentence of the following form: the parties are jointly committed to X as a body. Acceptable substitutions for “X” are psychological verbs such as “believe,” “intend,” and so on. What is it to be jointly committed to X as a body? As I understand it, this joint commitment has two related aspects. First, the parties are jointly committed to bring it about, as far as is possible, that the parties emulate a single body that Xs. Second, it is understood that their doing so is to be a function of the joint commitment in question. (Gilbert 2014, p. 175)

We will have reason, as we move along, to prefer something closer to Gilbert’s account rather than Tuomela’s account—or at least to modify Tuomela’s account more in the direction of Gilbert’s. But we are not yet in a position to motivate these

10  Sellars on Rational Agency as Presupposing Collective Attitudes

203

modifications. So let us turn our attention to the next question: What is the relation between practices, rules, and collective attitudes?

10.6  Collective Attitudes and Pattern-Governed Behavior We have seen that there are complex and important relations between practices, pattern-governed behavior, rules, and rational agency. What is the relation between these items and we-attitudes? There are multiple relations. Social practices are constituted by behaviors which must be represented as pattern-governed behavior in the we-mode.12 Using the tools offered to us by Tuomela and Gilbert, we can construct an argument that norms and social institutions rely on collective attitudes, and that acting on such norms involves acting in the we-mode in the strong sense (in Tuomela’s terminology). As noted above, there are differences between Tuomela’s and Gilbert’s accounts, and so one might be more suitable than the other. Let us begin by looking at Tuomela’s argument that, for example, social institutions presuppose we-attitudes. Tuomela illustrates this with reference to institutions, but the applicability of this argument is much wider. Recall Tuomela’s requirement that having a group attitude and acting on it involves the element of mutual belief, and acting on this mutual belief. Social institutions constitutively depend on this mutual belief and acting on this mutual belief—if not psychologically, then conceptually or presuppositionally. To return to the example of money, Tuomela notes, “Suppose x lacks the belief that y believes that x believes that Coins are correct exchange money for the group in question. Here x thus believes that it is not (fully) rational for y to engage in the exchange, because y, in accordance to what has been assumed, might not believe that x has a belief to the effect that…Coins are correct exchange money” (Tuomela 2002, p. 177). The same applies for norms generally, and not just for institutions. If, for example, I want to speak to you using the word ‘bird’ to mean bird, then not only must (a) I believe that ‘bird’s are •bird•s,13 but also (b) I must believe that ‘bird’s are •bird•s for you, and (c) I must believe that you believe that ‘bird’s are •bird•s for me. Again, none of this should be supposed to be a psychological reality; only in ‘breakdown’ situations will we consciously entertain any of these beliefs. (In the famous words of Inigo Montoya from The Princess Bride, “You keep using that word. I do not think it means what you think it means.”) But these commitments are conceptual presuppositions of language use, and indeed of the use of norms generally. Thus, institutions and norms presuppose we-attitudes in Tuomela’s sense, and are thus in the basic case instances of behavior in the we-mode in the strong sense.  Tuomela (2002) analyzes social practices in terms of collective pattern-governed behaviors.  I am here using Sellars’s helpful dot-quote notation. Dot quotes form a common noun that serves to classify words according to their functional role. Thus, “‘Rouge’ (in French) means red” becomes, for Sellars, “‘Rouge’s are •red•s”, meaning that ‘rouge’s function roughly the same in French as ‘red’s do in English. 12 13

204

J. R. Koons

However, Tuomela’s account of group action as involving mutual belief is too strong as it stands. For many cases, this requirement of mutual belief is plausible— particularly for the examples that populate the literature on group action (e.g., two people carrying a table together). But when we scale up our account to involve large social collectives (like users of a particular language), the requirement is less plausible. Something like mutual belief is at play, but (for example) I am not sure it is plausible to attribute to me a belief that you believe that I use the word ‘bird’ to mean bird when I speak. I wouldn’t use the word ‘bird’ in speaking to you if I didn’t in some sense think that there was a shared understanding of what the word meant. The key is fleshing out this ‘in some sense’. With what should we replace the mutual belief requirement, if it is too strong? Brandom argues that on the pragmatist way of thinking, one should “prefer an order of explanation that begins with what is implicit in practice (what people do) and proceeds to an account of what they explicitly believe or say, over one taking the opposite tack” (Brandom 1994, pp. 101). Let me note that any social practice—even one in which significant disagreement exists—presupposes massive agreement. Indeed, the very possibility of disagreement presupposes massive agreement. The ability to use language presupposes an agreement in what Wittgenstein would call ‘forms of life.’ These forms of life are not, in the first instance, beliefs or the like. Rather, they are habits of behavior (including linguistic behavior). Consider an analogy. Imagine someone walking a meandering path that threads its way through a series of ornamental ponds. The person may never entertain the belief that the path will support her weight, or that the water in the pond will not, or that if she attempts to walk across the pond she will get wet. But implicit in her behavior is a set of expectations about the way the world behaves. If the path instead meandered through a series of grassy patches, the person might cut across the grass, her behavior demonstrating her differing expectations about the behavior of the world. ‘Expectation’ is a misleading word here—it implies an occurrent mental state where there probably is none. Again, let us remind ourselves of Brandom’s comment that what we are doing here is making explicit commitments that are merely implicit in behavior, rather than attributing to the agent occurrent states to explain the behavior in question. The case is similar with agents who share a form of life. When I utter an English sentence to a fellow English-user, implicit in my behavior is the expectation that this will have a certain effect on the person in question. It might simply result in an updating of the other agent’s ‘scorecard.’ Or it might result in a certain corresponding behavior on the part of the other agent (if, for example, I have asked her for the time). But implicit in our real-time interactions with sharers of our form of life is the expectation that we do share this form of life—that we share a language, or a set of norms for greeting (e.g., shaking hands), or for conducting transactions in the checkout line at the grocery store, or rowing a coxed eight, or whatever the particular shared activity or set of norms in question is. Indeed, this element of mutual expectation—that each of us understands our role, that we will fulfill our part, and so on—is a fundamental and ineliminable part of the form of life. (What would be

10  Sellars on Rational Agency as Presupposing Collective Attitudes

205

the point of bringing groceries to the checkout counter if I didn’t expect the cashier to sell them to me? What would be the point of my ringing up your purchase if I didn’t expect you to pay? What would be the point of my rowing if I didn’t expect you to do so also? What would be the point of using language if I didn’t expect the appropriate updating or behavior on your part?) I would contend that for the Brandomian, pragmatist reasons gestured at above, we should extend the mutual belief requirement to include also this element of mutual expectation implicit in practice. Of course, some cases of group action do involve mutual belief. But most cases of group action will not involve mutual belief, and this is too strong of a requirement to impose. So from now on, let us understand the mutual belief requirement as involving mutual belief or expectation (in this implicit, Brandomian sense). Thus, to re-cast our example in terms of expectation rather than belief, suppose I want to speak to you using the word ‘bird’ to mean bird. Implicit in my behavior is (a) a commitment on my part that ‘bird’s are •bird•s, but also (b) an expectation on my part that ‘bird’s are •bird•s for you, and (c) an expectation on my part that you have an equivalent expectation, i.e., an expectation that ‘bird’s are •bird•s for me.14 Thus, my commitment regarding the meaning of bird is a commitment in the we-­ mode in the strong sense, as the mutual expectation of this commitment underlies the possibility of linguistic communication. That is, this element of mutual expectation is partially constitutive of linguistic norms, as the ability to communicate using linguistic tokens presupposes this element. Thus, the linguistic norms governing the use of ‘bird’—that is, that constitute •bird•—must be expressed in the strong we-­ mode. (I will have more to say on this shortly, when I talk about the role of collective attitudes in the expression of material rules of inference.) An individual’s norm-following behavior will generally be parasitic upon a set of group attitudes. This is most clear in the case of norms that are intended to solve coordination problems. Consider the norm of driving on the right side of the road. In following this norm, there is implicit in my behavior (a) a commitment to driving on the right side of the road, (b) an expectation that you will drive on the right side of the road, and (c) an expectation that you expect I will drive on the right side of  I am not using ‘commitment’ or ‘expectation’ with any particular technical definition in mind. One might give ‘commitment’ a Brandomian gloss, in which the term has both psychological and normative dimensions. Thus, to attribute a commitment to someone is to convey certain behavioral expectations (say, that the person will not infer ‘is the largest land mammal’ from ‘is a bird’). But it is also to convey certain normative expectations, expectations which are not conveyed merely by a behavioral prediction. (If I predict that it will rain tomorrow, or that a lion will stalk a particular zebra, and my prediction fails to come true, in neither case has a norm been violated. But if I attribute a commitment to you, and you fail to comply with this commitment, you have violated a norm—or at least I take you to have done so.) It has been suggested to me (by Preston Stovall) that perhaps I should understand the distinction as follows: Commitment as self-directed—to acknowledge a commitment is to hold myself to certain normative proprieties; to attribute a commitment to Smith is to hold Smith accountable to these proprieties. Expectation, on the other hand, is otherdirected—my expectations are of another’s behavior; the expectations I attribute to Smith are directed toward the behavior of agents other than Smith. In any case, as I said, I don’t have a formal view on this question, but am relying on a more-or-less intuitive set of concepts. 14

206

J. R. Koons

the road. These implicit commitments are manifested in my behavior: I drive on the right side of the road, I don’t swerve onto the shoulder (as I might if I expected you to suddenly deviate from the norm and veer into my lane), and so on. Given the above element of mutual expectation, this norm, too, is underlain by a we-­ commitment—an attitude in the strong we-mode—which the norm presupposes for its very ability to function as the norm that it is. Again, these commitments are implicit in our behavior—we generally don’t have an explicit attitude toward them. These beliefs only come to the psychological forefront in what Heideggerians call breakdown situations—when the normal smooth functioning of things is interrupted (as when someone is driving on the wrong side of the road). And since a shared expectation is a conceptual presupposition for my behavior, that means that my behavior is we-mode. Again, we see that social practices are constituted by behavior that must be described in terms of shared attitudes, and therefore as we-mode behavior. Thus, we should prefer something closer to Gilbert’s account rather than Tuomela’s account—or at least to modify Tuomela’s account more in the direction of Gilbert’s, where the joint belief condition is weakened to something like joint expectation, joint commitment, or something less implicative of an occurrent attitude on the part of the agent. There is another way Tuomela’s account is too strong. The element of expectation I propose above is implicit and, in an important sense, presuppositional for group action. Here is another analogy: Crucial to Sellars’s anti-foundationalist picture of knowledge is that a bit of language A can be conceptually and epistemically dependent on some theoretical commitments T, without A being inferred from T.  Nevertheless, this dependence makes T epistemically prior to A in a way that keeps A from being foundational; the dependence is crucial to both the epistemic and conceptual status of A. This is the role, for example, played by knowledge of standard viewing conditions in Sellars’s account of perceptual knowledge. Knowing that one is in standard viewing conditions is a condition on an observation belief being justified, but is not a premise from which the observation belief is inferred. Ergo, for Sellars, there is epistemic priority that is not necessarily inferential dependence. This is also how we should understand the role of the above implicit expectation condition in group attitudes, such as in the deployment of linguistic norms, or the utilization of social institutions like money. This view is also, I think, more consistent with Gilbert’s view than with Tuomela’s: Tuomela insists that acting in the we-mode in the strong sense requires acting for a group reason. But as I argued above, action in the we-mode in the strong sense can instead involve mutual expectation, where these expectations are presuppositions of the action rather than reasons for the action. (In the language of defeasible inferences, these expectations are ‘enablers’; they should not be seen as ‘premises’ that would be appealed to if we were to reconstruct the agent’s reasons for action.) Thus, acting in the we-mode in the strong sense does not seem to require either mutual belief, or acting on a group reason. Rather, it seems to require the more modest requirement of mutual expectation or joint commitment, which is a less intellectualist, more Gilbertian—and indeed more pragmatist—view of collective action.

10  Sellars on Rational Agency as Presupposing Collective Attitudes

207

One final comment: I have been writing about linguistic and other sorts of norm-­ governed behavior as instances of group action. This is something of a misnomer, for Sellars. For as I have emphasized, our pattern-governed behavior is not rule-­ following, and consists of acts, not actions. Nevertheless, this behavior has the complex relation to mutual expectation outlined above. Therefore, as I noted at the beginning of Sect. 10.6, we must understand this norm-governed behavior as pattern-­governed behavior in the we-mode, in the strong sense.

10.7  Material Rules of Inference and Collective Attitudes I argued in Sect. 10.4 that part of what distinguishes pattern-governed behavior from mere rule-conforming behavior is the availability of a metalanguage, in which the normative proprieties implicit in practice can be made explicit in the form of material rules of inference. What is the relation between collective attitudes and material rules of inference? When Sellars talks about collective attitudes, he generally only talks about collective intentions. Further, Sellars tends to talk about collective intentions only in the context of language exit transitions (primarily in connection with moral intentions), leading one to think that collective intentions (or intentions generally) only figure in practical reasoning leading to overt action. However, Sellars indicates that he understands practical reasoning much more broadly. In (1988), Sellars writes, that “accepting a proposition is, in a broad sense, a doing. It is not a physical doing, but rather a mental doing, but it is a doing, nonetheless” (1988, p. 302). Also, in (1969), Sellars chips away at the practical/theoretical distinction, writing, It is, as Kant pointed out, one and the same reason which is in some of its activities ‘theoretical,’ and in some of its activities ‘practical.’ Of course, if one gives to ‘practical’ the specific meaning ethical then a fairly sharp separation of these activities can be maintained. But if one means by ‘practical’ pertaining to norms, then so-called theoretical reason is as larded with the practical as is practical reasoning itself. (1969, p. 511)

So if we are to understand material rules of inference as essentially practical, what form will they take? There are a couple of different options, corresponding to the different types of intention in Sellars’s philosophy. As noted in Sect. 10.4, there are two types of rules in Sellars’s philosophy: ought-to-do’s (rules of action) and ought-­to-­be’s (rules of criticism). Corresponding to these two types of rules are two types of intentions. One is an intention to do something, expressed as a ‘shall-do,’ e.g., “Shall (I do A)”. Second, there is an intention that some state of affairs obtain, expressed as a ‘shall-be’, e.g., “Shall-be (The world is at peace)”. Should we understand material rules of inference as ought-to-do’s or ought-to-­ be’s? And correspondingly, are these rules expressed by shall-do intentions or shall­be intentions? Space prevents a detailed discussion of this issue, but it seems better to understand such expressions as expressions of shall-be intentions. While we may understand an assertion of a material rule of inference as an expression of an

208

J. R. Koons

intention that the underlying normative propriety be realized, construing this in terms of a shall-do intention seems problematic: Shall (I infer ‘Fido is a mammal’ from ‘Fido is a dog’)

This seems to blur the lines between practical and theoretical reasoning a bit too far. Sellars notes that “every intention must have at least one ‘up-to-me’ constituent” (1980e, §80/p. 89). An intention must have at least one element that is (at least as far as we can discern) under our control. But it seems that inferring, insofar as it is connected with belief, is in the standard case not under our control. However, Sellars is far less strict in applying this condition with shall-be’s. As I have already noted, shall-be’s and ought-to-be’s often concern states of affairs that cannot be directly willed (e.g., “It ought to be the case that people feel sympathy for the bereaved”). But such shall-be’s and ought-to-be’s are allowable, for Sellars, as long as the state of affairs in question can in principle be brought about by an allowable series of shall-do’s (such as those acted upon by trainers). So it is perhaps better if we understand assertion of a material rule of inference as follows: Shall-be (I infer ‘Fido is a mammal’ from ‘Fido is a dog’)

However, this still cannot be correct. Material rules of inference cannot be expressed by this type of intention. In explaining why, I will follow the reasons Sellars lays out in his discussion of moral judgment. Suppose Smith and Jones intend as follows (1980e, p. 96): Jones: Shall be [the nations remain at peace] Smith: Shall be [the nations remain at peace]

As Sellars says, since Jones and Smith are expressing personal intentions, although the descriptive content in the intentions is (in a sense) the same, the expressions of intention are not identical—they are at most parallel. Furthermore, it is not obvious that Smith and Jones can have a genuine moral disagreement, rather than a mere disagreement in attitude. Similarly, if Smith and Jones intend as follows: Jones: Shall be [I infer ‘Fido is a mammal’ from ‘Fido is a dog’] Smith: Shall be [I infer ‘Fido is a reptile’ from ‘Fido is a dog’]

it is not obvious they hold contradictory attitudes—any more than if they held the following pair of intentions: Jones: Shall be [I buy strawberries, if I am at the store] Smith: Shall be [I buy cherries, if I am at the store]

Just as in the moral case, the key (for Sellars) is that we endorse material inferences (like moral inferences) as a member of the community of rational agents. Thus, as Sellars writes, Probability statements are intersubjective. If Jones says to Smith, ‘It is probable that p’ and Smith agrees, they are agreeing about the same thing. This suggests that It is probable that p has something like the sense of

10  Sellars on Rational Agency as Presupposing Collective Attitudes

209

It is [epistemically] reasonable (for us, now) to accept ‘p’. (1974b, p. 437)

Thus, we must represent expression of material rules of inference as having the following form: Shall be (We infer ‘Fido is a mammal’ from ‘Fido is a dog’)

Now, if Smith and Jones express the following intentions, then they are genuinely disagreeing: Jones: Shall be [We infer ‘Fido is a mammal’ from ‘Fido is a dog’] Smith: Shall be [We infer ‘Fido is a reptile’ from ‘Fido is a dog’]

This fits nicely, I think, with Sellars’s discussion of ought-to-be’s and ought-todo’s (and trainers and trainees) in (1974c), (1980a), and elsewhere. Recall our discussion from Sect. 10.4 of the education of little Timmy: Trainers (who are able to follow ought-to-do’s) are able to condition a language learner’s behavior so that he conforms to ought-to-be’s. Thus, Timmy’s trainers act to realize a shall-be intention that embodies a rule regarding language-transitions (in this case, language-entry transitions): Shallwe be [Language-learners respond to red objects with the word ‘red’]

Of course, the trainers also act to bring about (in their trainees) shall-be intentions that embody language-language and language-departure transitions. Also, Sellars thinks that ought-to-be’s can only be realized if connected to ought-to-do’s; and so Timmy’s trainers can reason to the following ought-to-do: Shallwe [I correct Timmy if he responds to a red object with the word ‘blue’]

Note where our reasoning has led us: The intentions of trainers—which take as their object these ought-to-be rules—must be understood as we-intentions. Thus, one of the fundamental connections between pattern-governed behavior and rules— the connection between the trainers following ought-to-do rules in order to inculcate the proper ought-to-be’s in the trainees—is realized via the collective intentions of the trainers. Thus, we see (again) a fundamental connection between pattern-­ governed behavior, rules, and collective attitudes. There is one final point to clear up. As noted above, Sellars argues that intentions must have one element that is ‘up to me’. But what element is up to me in an intention about inferential proprieties—especially if, as I am willing to concede, doxastic voluntarism is false? Perhaps Sellars is best read as meaning that shall-do’s must have an element that is up-to-me. The whole point of ought-to-be’s (when they concern the behavior of persons) and the shall-be’s with which we express them is that they target behavior that, in general, cannot be directly willed, but must instead be conditioned by trainers. I do not choose to infer “Fido is a mammal” from “Fido is a dog,” any more than a person can choose to feel sympathy for the bereaved. But these transitions are ones that can be taught and conditioned by trainers. Thus, they have in the sense appropriate to shall-be’s an element that is under the control of the community of rational agents. Presumably, a state of affairs that cannot (physically or logically) be realized or brought about is not a suitable object of a shall-be. This

210

J. R. Koons

is the only sense in which we can require a shall-be (and the ought-to-be which it expresses) to have an element which is under our voluntary control: that it expresses a state of affairs which it is possible to bring about indirectly by acting on a shall-do, or a series of shall-do’s. Thus, to round up this section, the distinction between pattern-governed and merely rule-conforming behavior lies partly in our ability to formulate, as material rules of inference, the norms implicit in such pattern-governed behavior. Further, our expression of these rules in the metalanguage will take the form of expressions of collective intentions—intentions that we adopt a certain inferential pattern. The relation between pattern-governed behavior and rules is further explained via the intentions of trainers, who condition trainees to follow ought-to-be’s. We also saw that the trainer’s intentions must be seen as we-intentions, thus revealing an intersection between rules, pattern-governed behavior, and collective attitudes. Thus, in distinguishing pattern-governed from rule-conforming behavior—that is, in explaining what it is to be sapient—we must invoke collective attitudes. These are essentially presupposed by the kind of sapience possessed by rational agents such as us.15

10.8  Concluding Remarks We are in a position to draw some important conclusions from the above: • Meaning is constituted (largely—perhaps solely) by norms governing language-­ language transitions. • A crucial element in pattern-governed behavior (as opposed to mere rule-­ conforming behavior) is the ability to formulate explicitly these norms in the metalanguage, in the form of material inferential rules. Among other things, we saw that an agent cannot count as a rational agent—cannot engage in the game of giving and asking for reasons, or express or attempt to resolve reasonable disagreement over norms implicit in practice—without being able to access the metalanguage. Nor can these norms be rationally revised without the metalanguage. • Material rules of inference are expressed as we-intentions, and hence the distinction between pattern-governed and merely rule-conforming behavior presupposes the existence of a metalanguage in which we can formulate, in terms of collective attitudes, the normative proprieties implicit in our practices (and, for  This isn’t to say that a non-linguistic being (who a fortiori couldn’t go metalinguistic) couldn’t think. As deVries points out, Sellars comes around to the belief that non-linguistic beings can have thoughts*, where these are understood on analogy with the full-blown thoughts had by linguistic creatures: deVries argues that through an “analogical extension of the concepts of the framework of conceptual thinking to subconceptual thought…there will be ample reason to attribute thoughts* to animals, but thoughts* will not be identical to thoughts, just as, for example, the spin property of electrons is not identical to the spin property of tops, even though in some sense it is modelled on it” (deVries 2005, p. 187). 15

10  Sellars on Rational Agency as Presupposing Collective Attitudes

211

example, subject them to rational scrutiny and possibly revision). Thus, for Sellars, part of what distinguishes a rational agent from a mere rule-conforming object or entity is membership in a collective, a we.16 • While pattern-governed behavior is not rule-following, it conforms with ought-­ to-­be’s. These are inculcated by trainers acting on ought-to-do’s, which are embodied in the trainers’ collective intentions. Thus, there is another crucial connection between pattern-governed behavior, rules, and collective attitudes. • In acting on the norms implicit in our practice, we are acting in the we-mode in the strong sense. For example, my commitments regarding the meaning of a word in the language presupposes the element of mutual expectation. Indeed, the element of mutual expectation is partially constitutive of linguistic norms, as the ability to communicate using linguistic tokens presupposes this element. Thus, we must understand this norm-governed behavior as pattern-governed behavior in the we-mode, in the strong sense. • Therefore, at the very least our ability to act on linguistic norms—and therefore quite possibly our sapience and rational agency—presupposes both collective attitudes and a metalanguage in which we can formulate rules expressing the norms implicit in our practice. Plausibly, this is true of a wide range of other norms as well—particularly, as noted above, norms that solve collective action problems. One of Sellars’s overarching commitments—a project in which one of his major works, Science and Metaphysics, arguably culminates—is the attempt to prove the reality of the moral community. In a number of places (e.g., 1967, 7.XX.144/pp. 225–6), Sellars lays out the strategy for how the reality of this community would be established. However, Sellars himself seems ambivalent as to whether he can succeed in establishing the reality of this community, and concludes that “the argument for the reality of an ethical community consisting of all rational beings…remains incomplete” (1967, 7.XX.145/p. 226). Nevertheless, perhaps the argument of this paper points in the direction of how such an argument would begin. If the argument here is correct, then the individual thinker or norm follower cannot be prior in the conceptual order to the community of norm followers. The ‘I’ conceptually presupposes the ‘we’—and it presupposes it in a very particular way. The individual presupposes the community not in the sense that individuals require communities for material support (though this is generally true), or for flourishing (though this is also true). Rather, as noted, the norms that lie at the foundation of agency presuppose— as a conceptual matter—a range of we-attitudes, and hence the existence of the community. Thus, one’s answerability to a system of norms is not separable from the community’s adherence to (broadly) the same set of norms. Thus, it seems as though it follows from the above line of reasoning that to think of oneself as bound by norms is necessarily to think of oneself as a member of a community bound by similar norms. Indeed, we can put the point in an even stronger way17: To have the 16 17

 See, for example, Sellars (1963b, VII/pp. 38–40).  Preston Stovall suggested this stronger formulation of the point.

212

J. R. Koons

ability to speak a natural language, and so have the rich mental life that is possible for us only via a language, one must be a member of a community within which one sees oneself as, in a sense, participating in a common project that everyone (in the linguistic community) takes part in. If this isn’t quite the Universal Kingdom of Ends, insofar as there are translation schemes between languages it is at least a Universal Kingdom of Communication. Thus, the reality of the Sellarsian community seems to follow from the above considerations. This, of course, does not pretend to be a complete derivation of Sellars’s formal end of morality, or of the requirement of moral concern for all rational beings. That is a project for another occasion. But, I hope, we can at least see how such a project would begin. Acknowledgments  This essay draws on material from Koons (2019). I am grateful to Stefanie Dach, Willem deVries, Bhaskarjit Neog, Michael Wolf, Bill Wringe, and participants in a book manuscript workshop sponsored by the Center for International and Regional Studies (CIRS) at the Georgetown University School of Foreign Service in Qatar for providing me with extensive feedback on a draft of that work. Preston Stovall and Jaroslav Peregrin also read the present essay, and provided many helpful comments and suggestions.

Bibliography Brandom, R. B. (1994). Making it explicit: Reasoning, representing, and discursive commitment. Cambridge, MA: Harvard University Press. deVries, W. A. (2005). Wilfrid Sellars. Montreal: McGill-Queen’s University Press. Gilbert, M. (2014). Joint commitment: How we make the social world. Oxford: Oxford University Press. Koons, J. (2019). The ethics of Wilfrid Sellars. New York: Routledge. Lance, M. (2000). The word made flesh: Toward a neo-Sellarsian view of concepts, analysis, and understanding. Acta Analytica, 15(25), 117–135. McDowell, J. (1994). Mind and world. Cambridge, MA: Harvard University Press. Sellars, W. (1953). Inference and meaning. Mind, 64(247), 313–338. Sellars, W. (1957). Counterfactuals, dispositions, and the causal modalities. In H. Feigl, M. Scriven, & G. Maxwell (Eds.), Minnesota studies in the philosophy of science, Vol. II (pp. 225–308). Minneapolis: University of Minnesota Press. Sellars, W. (1963a). Science, perception and reality. Atascadero: Ridgeview Publishing Company. Sellars, W. (1963b). Philosophy and the scientific image of man. In Sellars (1963a), 1–40. Sellars, W. (1963c). Empiricism and the philosophy of mind. In Sellars (1963a), 127–196. Sellars, W. (1963d). Some reflections on language games. In Sellars (1963a), 321–358. Sellars, W. (1967). Science and metaphysics: Variations on Kantian themes. Atascadero: Ridgeview Publishing Company. Sellars, W. (1969). Language as thought and communication. Philosophy and Phenomenological Research, 29(4), 506–527. Sellars, W. (1974a). Essays in philosophy and its history. Dordrecht: D.  Reidel Publishing Company. Sellars, W. (1974b). Are there non-deductive logics? In Sellars (1974a), 417–438. Sellars, W. (1974c). Meaning as functional classification. Synthese, 27(3/4), 417–437. Sellars, W. (1980a). Naturalism and ontology. Atascadero: Ridgeview Publishing Company.

10  Sellars on Rational Agency as Presupposing Collective Attitudes

213

Sellars, W. (1980b). In J. Sicha (Ed.), Pure pragmatics and possible worlds: The early essays of Wilfrid Sellars. Atascadero: Ridgeview Publishing Company. Sellars, W. (1980c). Concepts as involving laws and inconceivable without them. In Sellars (1980b), 95–123. Sellars, W. (1980d). Language, rules, and behavior. In Sellars (1980b), 129–155. Sellars, W. (1980e). On reasoning about values. American Philosophical Quarterly, 17(2), 81–101. Sellars, W. (1988). On accepting first principles. In J. Tomberlin (Ed.), Philosophical perspectives, volume 2, epistemology (pp. 301–314). Atascadero: Ridgeview Publishing Company. Olen, P., & Turner, S. (2015). Durkheim, Sellars, and the origins of collective intentionality. British Journal for the History of Philosophy, 23(5), 954–975. Stoutland, F. (1997). Why are philosophers of action so anti-social? In L. Alanen, S. Heinamaa, & T. Wallgren (Eds.), Commonality and particularity in ethics (pp. 45–74). London: Palgrave Macmillan. Stovall, P. (2016). Nature, purpose, and norm: A program in American philosophy. Journal of the American Philosophical Association, 2(4), 617–636. Stovall, P. (forthcoming). Rationality, autonomy, and obedience to linguistic norms. Synthese https://doi.org/10.1007/s11229-020-02609-z. Tuomela, R. (2002). The philosophy of social practices: A collective acceptance view. Cambridge: Cambridge University Press. Tuomela, R. (2007). The philosophy of sociality: The shared point of view. Oxford: Oxford University Press. Wittgenstein, L. (1958). Philosophical investigations (3rd ed.) (G.  E. M.  Anscombe, Trans.). New York: Macmillan Publishing Company. Wolf, M. P., & Koons, J. (2016). The normative and the natural. London: Palgrave Macmillan.

Chapter 11

Understanding What We Ought and Shall Do: A Hyperstate Semantics for Descriptive, Prescriptive, and Intentional Sentences Preston Stovall

Abstract  This essay is part of a larger project aimed at making sense of rational thought and agency as part of the natural world. It provides a semantic framework for thinking about the contents of: (1) descriptive thoughts and sentences having a representational or mind-to-world direction of fit, and which manifest our capacity for theoretical rationality; and (2) prescriptive and intentional sentences having an expressive or world-to-mind direction of fit, and which manifest our capacity for practical rationality. I use a modified version of Allan Gibbard’s hyperstate semantics, employing both maximally determinate possible worlds and maximally determinate plans of action, as a basis for providing a unified understanding of moral judgments and expressions of individual and collective intentionality – they one and all give voice to our ability as rational agents to adopt the perspectives of various individuals within a community and consider how we would behave were we in their positions. In the course of spelling out the view I draw on and criticize ideas that Wilfrid Sellars advanced in the middle of the twentieth century, while employing the tools of contemporary modal logic and model-theoretic semantics to give perspicuous formulation to his thought that the moral judgment ‘one ought to do A’ should be understood in terms of the collective intention ‘we shall do A’, where the pronoun denotes the unrestricted class of rational agents and the ‘ought’ is in some sense unconditionally binding. Keywords  Collective intentionality · Deontic modality · Expressivism · Planning semantics · Sellarsian ethics

P. Stovall (*) Faculty of Philosophy, University of Hradec Králové, Hradec Králové, Czechia e-mail: [email protected] © Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9_11

215

216

P. Stovall

11.1  Introduction: The Moral ‘Ought’ and the Intentional ‘Shall’ Though he is not often discussed alongside the ‘big four’ of collective intentionality (Bratman, Gilbert, Searle, and Tuomela),1 Wilfrid Sellars did much to stimulate interest in the subject in the middle of the twentieth century. Over the course of a series of works addressing the nature of moral reasoning, Sellars developed an account of collective intentionality on the basis of which to understand moral judgment (e.g. 1951, 1963, 1966a, b, 1967a, chapter 7 of 1968/1992, 1976, 1980). In doing so he outlined (here one could not say developed) a theory of the semantic functions of modal terms like ‘shall’, ‘ought’, and ‘could’. His idea is that the assertion of a sentence like ‘bankers ought to dress professionally’ gives voice to a collective intention concerning what anyone in the community of bankers shall do. Moral judgments are or involve a universalization of this collective intentionality across all rational beings: use of the moral ‘ought’ is a way of giving voice to what any of us shall do, where the ‘us’ is unrestricted and where ‘shall’ connotes an intention-in-expression (as opposed to a claim that such an intention exists). This idea relates judgments about what is right and wrong concerning exercises of instrumental practical rationality – which concern what anyone ought to do to achieve a particular end in a given situation – to judgments about what is right and wrong concerning morality. And while the instrumental ‘ought’ concerns action that is directed at some end, and so its universality is circumscribed by that end as a condition on the force of the ‘ought’, the moral ‘ought’ is one whose force is in some sense unconditionally binding. In this regard Sellars sees himself spelling out a conception of human cognition consonant with Kant’s.2 Looking back, we can also understand this as an expressivist or non-cognitivist analysis of moral judgment.3 While I will argue below that we have good reason to dispute some of what Sellars concludes about moral reasoning, collective intentionality, and the logic of the associated modals, I believe that in some areas he saw farther and more clearly than is commonly appreciated. The last chapter of Science and Metaphysics contains a discussion of the different roles  For the reference see the Introduction to Chant, Hindriks, and Preyer (2014). For exceptions see Koons (this volume), Olen and Turner (2015), Peregrin (this volume), and Schweikard and Schmid (2013, §2.3). 2  From the closing chapter of Science and Metaphysics (1968), a book whose subtitle is “Variations on Kantian Themes” (p. 176): 1

…in my opinion the metaphysics of morals is but a fragment of a broader critique of practical reason. 3  Though the term ‘expressivism’ was not then in common currency, Sellars occasionally discusses non-cognitivism (usually in connection with emotivism) – cf. (1967b). See Klemick (2018) for a quasi-realist defense of Sellars’ non-cognitivism about ethics, and Koons (2019, chapter 5), for a general discussion of Sellars’ expressivism.

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

217

played by the contents of collective intentions as against their mode as collective, and of the two complementary directions of fit that characterize descriptive and intentional states of mind, that resonates today (its echoes of a classical pragmatist use of the scientific categories employed to understand neural activity as a basis for thinking about theoretical and practical rationality, however, have gone almost totally silent in the intervening years). From pp. 188–9: 33….I can intend that someone else do A (intend him to do A), and he can intend to do A.…There is, nevertheless, an asymmetry. For, obviously, only his intention can grow directly into his volition and his action. My intention with respect to his action can grow into a volition only if practical reasoning draws a conclusion from it concerning some influence I can bring to bear. 34. This latter consideration highlights the obvious fact that even where the descriptive content of two people’s intentions is in the strongest sense the same not only are the intendings numerically different, which is true even in the case of belief, but the total content of the intendings involves a special mode of egocentricity (expressed by the word ‘shall’) which is, in practical discourse, the counterpart of the egocentricity of demonstratives. The latter is the egocentricity involved in the impact of the world on discourse, the former is the egocentricity involved in the impact of discourse on the world in volition.

With the benefit of hindsight, however, and looking across decades of work on model theory as a basis for semantically evaluating different vocabularies, we can see not only that Sellars made some conceptual moves that are not forced upon us, but also that his survey of the terrain needs to be re-examined. For Sellars left at best an outline for a semantics of moral and intentional discourse, and while Koons (2019) offers a systematic elaboration and defense of Sellars’ ethics, and relates Sellars’ ethics to the work of Bratman, Gilbert, Tomasello, and Tuomela on collective intentionality, we still lack a semantic framework sufficient to unify the prescriptive and intentional modalities under a common conception of practical rationality (I use ‘deontic’ and ‘prescriptive’ interchangeably in this essay, concerning what agents ought, may, and are forbidden from doing). There are also problems with the sketch of an account Sellars left. By simply identifying deontic judgments with expressions of collective intentions (e.g. 1963, part 12; 1968, p. 219; Koons, Chap. 5, also identifies ‘ought’ judgments with ‘shall’ judgments), it would appear that ‘we ought to A and we shall not A’ is rendered as the contradiction ‘we shall A and we shall not A’. But though an assertion of the first sentence may involve some sort of practical irrationality, there is a non-empty mental state associated with it, as we’ve recognized for millennia  – judgments such as these led the prophets to rend their clothes at what their communities were doing. Where negation is a complementarity operator, however, an assertion of the second sentence has no associated mental state (I am glossing over issues of internal and external negation that are discussed in detail in Sects. 11.3.2 and 11.3.3). At the same time, Sellars thinks that a judgment that one shall do something commits one to the claim that one will (e.g. at section 4 of 1963; though rejected by Ludwig 2016, p. 115, this view is not uncommon in the literature on collective intentionality – cf. Bratman 2014, p.  76, Mele 1989, p.  19 and the references in note 1, and

218

P. Stovall

Tuomela and Miller 1988, p. 375).4 If we allow this implication to hold, and we identify ‘shall’ with ‘ought’, we need to make sure that commitment to the claim that one ought to do something does not also commit one to the claim that one will – for just as surely as ‘we ought to do A and we shall not’ is coherent, so is ‘we ought to A and we will not’. This again suggests distinguishing ‘ought’ from ‘shall’. Additionally, if deontic judgments simply express collective intentions it is not clear how to understand the claim that something is permitted but not obliged (cf. Koons 2019, pp. 132ff.). And what sort of intentional state of mind is expressed when one says that the performance of some general act is obligatory, two of its specific instances are separately permitted, while one intends to do one or the other – e.g. ‘I ought to go to the lecture, I am permitted to sit in the front, I am permitted to sit in the back, and I shall sit in the front’? This essay grounds a unified conception of moral reasoning, individual and collective intentionality, and practical rationality on a solid semantic footing by using the resources of model theory and modal logic to supply an interpretation for a language that includes descriptive vocabulary, prescriptive vocabulary, and intentional vocabulary. On this semantics representational interpretations are provided for descriptive vocabulary and expressivist interpretations are provided for prescriptive and intentional vocabulary. Using a modified version of the plan-theoretic semantics presented in Gibbard (2003), my goal is to explicate the idea that both moral reasoning and collective intentionality can be understood as the exercise of practical rationality across perspectives of person, place, and time. The result is a view on which descriptive sentences represent the space of possible worlds, while prescriptive and intentional sentences express commitment to plans of action that involve putting oneself into the shoes of others within a community. The view diverges from Gibbard and Sellars in various ways (some of which I detail below) but the upshot is a semantics able to establish a number of implication relations that the descriptive ‘will’, the prescriptive ‘ought’, and the intentional ‘shall’ ought to conform to. I believe this framework will prove useful in sorting out some of the debate in the literature on collective intentionality, but my aim in this essay is to present the semantics, explain the resources its models employ, discuss Sellars’ account alongside my own, and motivate the conclusion that there is more to be learned from studying views of this sort. With the exception of a brief discussion of Bratman’s views on planning mental states as a basis of collective intentionality in Sect. 11.3.8, I defer to future work a comparison of this semantics with the views of the big four and the surrounding literature.

 Bratman (2014) notes that this condition is too strong to be a necessary condition on intention, Mele (1989) defends the weaker claim that if one intends to do something one at least does not believe that one will not, Tuomela and Miller (1988) hedge with a parenthetical claim that one at least believes that one probably will do what one intends, and in footnote 23 of (1963) Sellars includes a ceteris paribus clause. These weaker commitments could be represented here instead of the stronger one, but as all are optional I stick to the simpler formulation so as to illustrate how to accommodate the view for those so inclined. 4

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

219

Two final programmatic notes. First, my focus here lies solely on intentions for action, and I leave questions about collective attention, thought, and other mental states undiscussed. Second, as my interest lies in the modal reasoning and judgment that attends practical rationality I do not discuss modalities for non-agentive normative evaluation, e.g. concerning how organisms and artifacts ought to be or behave.

11.2  A Deontic Hyperstate Semantics 11.2.1  Background and Framework In other work (Manuscript) I show that by taking a class of atomic descriptive sentences interpreted via possible worlds, a class of prescriptive sentences interpreted via plans of action of the sort discussed below, and the standard set-theoretic definitions for the Boolean operators, the content of any mixed or pure descriptive and prescriptive sentence can be recursively modelled, where the mind-to-world descriptive function and the world-to-mind prescriptive function is kept distinct. In this part of the essay I present an abbreviated formulation of that semantics, and in Sect. 11.3 I discuss its extension to account for intentional sentences. The central semantic concept is the notion of a hyperstate, introduced in Gibbard (2003) and understood as an ordered pair of a maximally determinate possible world and a maximally determinate plan of action (a hyperplan). Think of a hyperplan as a specification of what a person would do at every possible point of choice, including situations where one is other people – and so a hyperplan specifies what to do if one were Caesar faced with crossing the Rubicon, or a colleague faced with the choice of whether or not to attend a lecture. Just as descriptive mental states, having a mind-to-world direction of fit, can be understood as sets of possible worlds (the worlds at which the descriptions are true), so can prescriptive mental states, having a world-to-mind direction of fit, be understood as sets of plans of action (the plans where the agent, or the members of a class of agents, chooses to do the thing prescribed). Because a hyperstate includes both a world and a plan, it is possible to distinguish various features of the contributions that descriptive and prescriptive or deontic thoughts make to our cognitive economy. While the semantic values of descriptive and prescriptive sentences can be provided on this basis, a straightforward extension of this framework would use these hyperstates to model the contents of intentional sentences as well. This leads to the problem that the sentence ‘we ought to A and we shall not A’ becomes the contradiction ‘we shall A and we shall not A’, however. And so in Sect. 11.3 I distinguish deontic hyperplans from intentional hyperplans, and I define a deontic-intentional hyperstate as an ordered triple of a possible world, a deontic hyperplan, and an intentional hyperplan. But I begin here because the ordered pairs – and the descriptive/prescriptive fragment of the language – are simpler to work with.

220

P. Stovall

11.2.2  Single-Mindedness and Indifference The deontic hyperplans that model prescriptive mental states require that we distinguish two different ways of choosing – single-mindedly and indifferently. To forecast, one chooses an action single-mindedly when one regards it as obliged and one chooses an action indifferently when one regards it as merely permitted  – thus, belief that one ought to attend the lecture and that one is permitted to sit in either the front or the back of the room will be understood as a plan to choose single-mindedly to attend the lecture while choosing indifferently whether to sit in the front or the back. The distinction between these two choice attitudes has gone neglected in the literature on planning expressivism. While it is discussed in Gibbard (2003), that discussion has been seldom mentioned by his critics. Silk (2015, p.53, footnote 8) notes that by passing over this distinction many of Gibbard’s critics have not done service to his view. Dreier (2006, 2009) advances a similar distinction, using the same terminology, but without noting its occurrence in Gibbard. Interestingly, Sellars introduced the same distinction in his own work (e.g. 1966b, pp. 113–115; 1976, p.57), using the very same terms that both Gibbard and Dreier use, though to my knowledge he never discusses it in any detail (after introducing the distinction in the latter essay Sellars writes “but I am concerned to sketch rather than map”). As neither Gibbard nor Dreier note that Sellars discusses this distinction, it appears that the notions of single-minded and indifferent choice attitudes have been independently developed at least three times in the last half century.5 These notions are absolutely central to a planning semantics for the deontic modalities, however. For in order to distinguish the strong and weak deontic modalities of obligation/forbiddance and permission, each point of choice on a deontic hyperplan must be specified according to whether the decisions undertaken are done so either single-mindedly or indifferently. The hyperplans that model intentional mental states do not need to mark this distinction, however (why this is so will become clear below). That is, each point of choice on an intentional hyperplan (introduced in Sect. 11.3) specifies only which actions are taken and which are not, without regard for the attitudes under which the choices are made. This indicates, I will suggest, that the intentional state of mind, whether individual or collective, is a less cognitively demanding and conceptually sophisticated sort of mental state than the state we express when we issue prescriptions under the deontic modalities (this also motivates keeping two sets of books for deontic and intentional plans).6 The notion of a single-minded choice is to be understood through the attitude of rejection, and the indifferent choice can be introduced on this basis (Gibbard introduces talk of rejection, but the definition of indifference in terms of rejection and  Sellars, Gibbard, and Dreier each speak of preferential and indifferent choices, but because one might suppose that any choice exhibits one’s preferences I choose to contrast indifference with single-mindedness. And as I provide my own definition of indifference in terms of single-mindedness I will use this terminology. 6  I examine relations of priority among deontic, individually intentional, and collectively intentional mental states in Stovall (2021), and more extensively in Stovall (Forthcoming). 5

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

221

single-mindedness is my own). Just as one can believe that some state of affairs obtains, so can one reject some course of action. An agent chooses an action A single-­mindedly when she rejects every action B that is incompatible with A (rejection is a term of art in this sense). An agent chooses an action A indifferently just in case she is able to choose single-mindedly and there is some action B that is (1) incompatible with A and (2) is such that she could have picked B without changing any of her single-minded choices. When an agent makes a choice indifferently she has rejected rejecting each of the choices that she could take without changing any of her single-minded choices. Thus, it is of the nature of both single-mindedness and indifference that one who chooses from these frames of mind is at the same time attitudinally related to choices she does not undertake. The intentional choice may, however, be specified solely in terms of the choice itself; this again suggests that the deontic frame of mind involves a sort of cognitive sophistication that the merely intentional does not. It is clear that we can be in these planning states of mind, and the mental states associated with the assertions of sentences concerning what is forbidden, what is obliged, and what is permitted can be specified in terms of them. For instance, suppose I have decided that I ought not go home early this afternoon, that I ought to go to the department colloquium, and that it is both permissible that I sit in the front of the room and permissible that I sit in the back of the room. In this circumstance my mental state includes rejecting going home and rejecting not going to the talk, modelled by the set of hyperplans where I single-mindedly choose to go to the talk by rejecting every choice that is incompatible with going to the talk – including, among other things, the choice to go home. If I do not change my mind when it comes time to attend the talk, then I actualize this planning mental state by choosing single-­ mindedly to go to the talk. Finally, believing that both sitting in the front of the room and sitting in the back of the room are permitted is a mental state modelled by the set of plans where I variously choose indifferently to sit in the front of the room and indifferently to sit in the back of the room. When it comes time to act on one plan rather than the other I actualize the plan in question by choosing indifferently between those choices, which is to say I reject rejecting the alternative (the option I could have undertaken without changing any of my single-minded choices). And so should I choose to sit in the front of the room that choice is both an indifferent choice to sit in the front of the room and an indifferent choice not to sit in the back of the room.

11.2.3  A  Hyperstate Semantics for Descriptive and Prescriptive Sentences In order to model sentences concerning what restricted classes of people ought to do in specific circumstances, it will occasionally be useful to talk about types of circumstances (weeknights, weeknights where there is a colloquium talk, etc.), kinds of actions (walking to a colloquium talk, walking quickly to a colloquium talk, etc.),

222

P. Stovall

and classes of agents (all of the people in my department, all philosophers, etc.). Tradeoffs can be made in terms of where we locate complexity, and for the semantics of prescriptive sentences I will assume that circumstances include information about the classes of agent(s) who make choices there, while keeping kinds of action separate from the circumstances in which they occur.7 E.g., deontic hyperplans concerning how everyone in my department should dress for work will be framed in terms of what the unrestricted class of agents would do were they to find themselves in a circumstance where they are members of my department faced with the choice of how to dress for work. A language that included kind-restricted quantifiers could be supplied a semantics that distinguishes the non-universal classes of agents to which prescriptions apply from their circumstances of action. The introduction of quantificational devices would also facilitate a treatment of de re and de dicto modality, but doing so would require relativizing semantic values to points of evaluation and such a discussion would take us too far afield (see the Appendix for a brief overview of these issues). In future work I hope to examine the interaction between quantification and modality on a hyperstate semantics in more detail, but here I treat prescriptions as applying universally and build their restricted applications into the circumstance condition. Additionally, all deontic modality over sentences concerning groups of people should be read de dicto (though again, see the Appendix for a discussion of de re and de dicto modality). Consider a language L consisting of atomic descriptive sentences, atomic sentences falling under one of the deontic modalities of obligation, permission, and forbiddance, and the sentences composed from these together with the Boolean operators. Hyperstates can be used to provide a semantic evaluation on which the descriptive sentences represent the space of possible worlds and the prescriptive sentences express plans of action. Let a possible world w be defined as a maximally consistent set of true sentences. That is, for every descriptive sentence d ∈ L and every w, either d is true at w or ~d is true at w. The content of a descriptive sentence can then be understood as the set of worlds at which it is true. In this regard descriptive sentences can be understood as representations of the space of possible worlds. Plans of action are defined in terms of the choices made on them, where these plans are maximal in the sense that they specify what anyone would do at any every of choice. More precisely, let a deontic hyperplan hD be defined as a maximally consistent plan of action such that, for every circumstance C, every agent α able to make a choice at C, and every action A that α is able to choose to perform at C, either (exclusively): 1. 2. 3. 4.

α single-mindedly chooses to A at C on hD; or α single-mindedly chooses not to A at C on hD; or α indifferently chooses to A at C on hD; or α indifferently chooses not to A at C on hD.

 Sellars (1968, p. 195) treats hypothetical imperatives in much the same way, holding that they apply to all rational beings insofar as they find themselves in certain circumstances (see the discussion at Koons 2019, pp. 72ff.). 7

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

223

Officially α is a metalinguistic variable but I abuse notation and occasionally treat it as an instance in the object language (e.g. in Sect. 11.3.8). Various requirements that a semantics must meet in order to model our notion of agency correspond to restrictions placed on the relations among hyperplans. For instance, notice that in general one undertakes many actions at a single point of choice – viz., to choose to attend the department colloquium this afternoon is also to choose not to go home early. Whatever determinate inclusions and exclusions may obtain among choices should be reflected in the plans we use to think about them. It would also seem part of the logic of agency that wherever it is possible to plan to do something then it is also possible to plan not to do it, understood de dicto in the sense that one could plan to do what one takes to be something else. In the branching-times framework of Belnap, Perloff, and Xu (2001), employing the ‘sees to it that’ modal operator for talk of agency, this is represented by a requirement that where it is possible to see to it that something is the case it is also possible to not see to it that it is the case – there is no agency where an agent could not have done otherwise (this is the ‘negative requirement’ on agency, discussed at pp. 36–7). If that is right, and if our talk of intention inherits the logic of our talk of agency, then the following is a plausible condition to impose on the notion of a hyperplan. With ‘h1’ and ‘h2’ ranging over deontic and intentional hyperplans (introduced in Sect. 11.3), and with ‘chooses*’ ranging over ‘chooses indifferently’, ‘chooses single-­ mindedly’, and ‘chooses’ (the third case is to ensure that intentional hyperplans also meet this condition): Genuine Agency: ∀α∀A∀C∀h1[if α chooses* to A at C on h1 then ∃h2(α chooses* not to A at C on h2)] Genuine Agency says that for any plan on which one chooses to do something (perhaps under a particular attitude) there is another plan on which one chooses (under that same attitude) not to do so. Let a deontic hyperstate ⟨w, hD⟩ be an ordered pair of a possible world and a deontic hyperplan and let SD denote the set of deontic hyperstates. That is, where ‘W’ denotes the set of possible worlds and ‘HD’ denotes the set of deontic hyperplans, a deontic hyperstate ⟨w, hD⟩ ∈ SD is an element of the Cartesian product of W and HD. The contents of states of mind voiced with descriptive and prescriptive sentences are given by the world and choice components of deontic hyperstates (from here on I tend to trade talk of states of mind for the contents of the sentences used to convey them). Where ⟦φ⟧ denotes the content of a sentence φ ∈ L under a given interpretation, for descriptive sentences d we have: ⟦d⟧=def.{〈w, hD〉: d is true at w} For prescriptive sentences we distinguish three varieties, corresponding to deontic judgments concerning what is forbidden, what is obliged, and what is permitted. Schematically, the contents of the three kinds of deontic sentence are determined as follows. An assertion of ‘people are obliged to A in C’ expresses universally rejecting not doing A in C: ⟦people are obliged to A in C⟧ =def. {⟨w, hD⟩: for every α, α single-mindedly chooses to A in C on hD}

224

P. Stovall

Thus, the content of ‘my colleagues are obliged to go to colloquia’ is the set of hyperstates whose hyperplans are such that every person who is in a circumstance of being my colleague and faced with a choice of either going or not going to a colloquium chooses single-mindedly to go. This will include, e.g., plans where a colleague chooses indifferently to sit in the front of the room, chooses indifferently to sit in the back, etc., so long as these choices are also ones where the agent chooses single-mindedly to go to the colloquium. To say that my colleagues are obliged to go to colloquia, then, is to express commitment to a plan concerning what I would single-mindedly do were I anyone of my colleagues. Similarly, an utterance of ‘people are forbidden to A in C’ expresses universally rejecting doing A in C: ⟦people are forbidden to A in C⟧ =def. {⟨w, hD⟩: for every α, α single-mindedly chooses not to A in C on hD} And so the content of the thought given voice with an utterance of ‘my colleagues are forbidden from going to colloquia’ is the set of hyperstates whose hyperplans are such that every person who is my colleague and faced with the choice of going to a colloquium chooses single-mindedly not to go. Finally, the assertion of ‘people are permitted to A in C’ expresses universally rejecting rejecting doing A in C: ⟦people are permitted to A in C⟧ =def. {⟨w, hD⟩: for every α, α either single-mindedly chooses to A in C or indifferently chooses whether or not to A in C on hD} The content of the thought that my colleagues are permitted to go to colloquia is therefore the set of hyperstates whose hyperplans are such that every colleague of mine who is in a circumstance of being faced with attending a colloquium either chooses single-mindedly to go, or chooses indifferently whether or not to go (this allows for individual hyperplans where an agent sometimes chooses to indifferently to go and sometimes chooses indifferently not to go). In this case the only hyperstates ruled out are those whose hyperplans include colleagues who choose single-­ mindedly not to go. It is easily established that where negation is a complementarity operator, understood in terms of rejection for prescriptive sentences, these three kinds of state of mind bear all the right duals with negation (when something is not forbidden it is permitted, when it is not obliged one is permitted not to do it, etc. – see Manuscript for details).

11.3  A Deontic-Intentional Hyperstate Semantics 11.3.1  From Deontic Prescription to ‘We-Shall’ Intention According to this proposal, when everyone in my department shares the mental state of being in agreement about how we ought to comport ourselves at a colloquium, we are in a state of mind wherein we each agree on what we would do were we

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

225

anyone else in the department. To plan in this sort of way is to think of oneself as a member of a group whose other members we relate to as if we could make their choices for them. This feature of the planning apparatus discharges a key commitment in Sellars’ theory of moral judgment. For on this account a moral judgment that one ought to do A expresses a plan, were one anyone, anywhere, at any time (where A-ing is at issue) to choose to do A. And so the capacity for reasoning in a moral frame of mind is the universalization, over person, place, and time, of the practically rational capacity to plan one’s life. In this section I extend this treatment by providing a hyperstate semantics for individual and collective intentions.

11.3.2  Sellarsian ‘Shall’s As I mentioned at the start of the essay, I think we have good reason to question some of the commitments Sellars takes on. Perhaps most surprisingly, he held that judgments using the ‘shall’ operator do not fall under the scope of logical operators. He also repeatedly examines problems with negation without arriving at a satisfactory picture of the interaction between negation and intentional vocabulary. Sellars’ reasoning for denying that ‘shall’ sentences logically compose with other sentences is not above criticism. For one thing, it raises problems of compositionality for a language containing both descriptive and non-descriptive sentences. Koons (2019, pp. 132ff.) examines this point in some detail and concludes that we should allow for embedded ‘shall’ sentences. It is also not clear that Sellars, by his own lights, should bar ‘shall’ from embedding. For he is happy to talk about implications among statements of intention. In a footnote at p.115 of (1966b) he defends this by asserting that “the relation word ‘implies’…is not a truth-functional connective [and] is in the metalanguage, and mentions rather than uses the shall-sentence” (emphasis in the original). This is a curious position for one who, 13 years earlier (1953), had argued that metalinguistic claims about inferential relations among sentences could be understood in terms of object-language claims that use conditionals. And so insofar as Sellars allows metalinguistic implications like ‘I shall walk quickly to the colloquium talk’ implies ‘I shall walk to the colloquium talk’ one would expect him to allow object-language sentences like if I shall walk quickly to the colloquium talk, then I shall walk to the colloquium talk Perhaps that sounds stilted. But there would seem to be a perfectly sensible state of mind associated with that conditional. At any rate, I mean to allow ‘shall’ sentences to embed arbitrarily under the Boolean operators and I defer a consideration of non-­ Boolean implication for another time. There are also issues with Sellars’ discussion of negation, as it was his contention that ‘shall’ sentences cannot be externally negated in any sense that differs from their internally-negated variants. Sellars’ discussion of external and internal negation is more substantial, however, and he examines the issue across a number of

226

P. Stovall

texts (especially 1963, 1966b, and chapter 7 of 1968; see also Klemick 2018, §2 and Koons 2019, chapter 5, the latter containing an extended discussion of Sellars’ standpoint on negation and ‘shall’). His central concern is that ‘ought’ clearly distinguishes an external from an internal negation: ‘it is not the case that you ought to A’ differs in meaning from ‘you ought not A’ – the former marks a permission, the latter something forbidden. This is a problem for Sellars because he wants to analyze ‘ought’ in terms of the collective ‘shall’, and he does not think that ‘I shall not A’ and ‘it is not the case that I shall A’ marks a similar sort of distinction. His commitment to the equivalence between ‘ought’ and ‘shall’ thereby requires that he explain why, as he claims, expressions of individual intention do not distinguish internal and external negation while expressions for collective intentions, or the deontic expressions that go proxy for them, do.8 In parts 11 and 12 of (1963), an essay that is cited at a footnote of p.185 of (1968) as the point of contact for his understanding of negation in practical contexts, Sellars argues that ‘ought’ has an external negation because of the intersubjectivity of ‘ought’ claims. At the end of part 12 he writes (emphasis in the original; notice the use of ‘ordinary’ in the second sentence, apparently referencing individual rather than collective intentions – my thanks to Jim O’Shea for pointing this out): Let me now bring all these considerations together. I suggest that the fact that ought-­ statements, unlike ordinary shall-statements, have a proper negation is built on the shared intending expressed by “ought”. In other words, the syntactical intersubjectivity of ought-­ statements which [makes it possible that two ought-statements contradict each other], and which consists in the existence of the form “not-ought[X doing A]” in addition to the form “ought[X not doing A]” rests on the intersubjectivity of the intention expressed by ought-statements.

It is true that expressions of collective intention, as with expressions of moral judgment, share the feature that they are in principle part of an intersubjective space of practical reasoning within some community. But we can register this point without conceiving of moral reasoning is a species of collective intentionality, and when we do we free ourselves of the need to explain why ‘ought’ has an external negation while ‘shall’ does not. Sellars faces that need only because he defined ‘ought’ in terms of ‘shall’– in the paragraph preceding the one just quoted Sellars writes (emphasis in the original) “I suggest that ought, as an expression of intention, is a special case of shallwe”. Koons (2019, chapter 5) follows Sellars in this identification and argues that we should distinguish external from internal negation for ‘shallwe’. I do not think we should identify ‘ought’ with ‘shallwe’, or otherwise define the former in terms of the latter. These two modes of expression can be understood in terms of the general notion of planning, and in that regard they are species of the genus practical rationality. But as species they are differentiated separately  – whereas the plans that model the deontic modalities distinguish the single-minded from the indifferent choice, the plans that model the intentional modalities do not.  There are other features of Sellars’ account that a fuller story would have to look out. For instance, I will have nothing to say about Sellars’ claim that intentions can at most conflict but not contradict, nor his understand of the relationship between truth and negation. 8

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

227

11.3.3  Issues of ‘Shall’ and ‘Not’ As I see it, there are three questions in play here. First, there’s the question of whether we can frame a semantics on which sentences containing ‘shall’ and ‘ought’ have the right kinds of logical properties. This is a question to be decided by formal considerations and the metaphysics of the models we use to interpret the language under consideration, together with whatever arguments we give for whatever logical properties we think the semantics ought to have. Second, there is the question of whether ‘shall’ has a dual, as ‘obligation’ does with ‘permission’ and ‘necessity’ does with ‘possibility’. An answer to this question will be closely connected to the answer one gives to the first, on account of the fact that internally and externally negated ‘ought’ claims clearly differ in meaning, where the strong force of forbiddance, defined by an internally-negated ‘ought’ claim (one ought not A – i.e., it is forbidden to A) is distinguished from the externally-negated ‘ought’ claims that mark the weak force of permission (it is not the case that one ought to A – it is permitted not to A). Finally, there is the question of whether saying that one shall not do something is equivalent to saying that it is not the case that one shall do it – this is a question of whether ‘shall’, in either its individual or collective guise, as a matter of the grammar of English (or the pidgin-English spoken by a community of philosophers), distinguishes an internal and an external negation. Concerning the first question, the Genuine Agency condition and the fact that intentional hyperplans do not distinguish single-minded from indifferent choices ensures that for any action A an agent is capable of undertaking, an externally-­ negated ‘shall’ sentence (whether individual or collective) is equivalent to an internally-­negated one (see proof in Sect. 11.3.7).9 And so from plausible assumptions about the nature of our planning agency it follows that there is no distinct external negation for ‘shall’. While this entailment is grounded in the metaphysics of agency, there is no corresponding entailment that the absence of an intention to do something is the intention not to do it, however. That second entailment corresponds to a descriptive claim about the ontology of intentions, and it would be reflected in a relation among the world elements of hyperstates. But the equivalence between ‘shall not’ and ‘not shall’ is a claim about the logic of sentences expressing intentional agency: any time one expresses that it is not the case that one shall do A, one has thereby expressed that one shall not do A. Nevertheless, we can preserve a distinct internal and external negation for the deontic modalities, and this feature of the language is captured by the fact that the hyperplans used for interpreting the expressions of deontic judgment specify whether a choice is made single-mindedly or indifferently. The distinction between single-mindedness and indifference

 The restriction to actions the agent is capable of undertaking rules out cases like ‘it is not the case that I shall make the sun explode’, which should be understood as descriptive claims that one lacks a particular intention (the expressive intentional thought can be captured with ‘it is not the case that I shall try, etc.’). If we allow hyperplans to include actions that are apparently impossible, this restriction can be lifted. 9

228

P. Stovall

thereby answers the second question by ensuring that the duality of the deontic modalities is semantically marked, while the absence of that distinction for the semantic evaluation of the intentional operator ‘shall’, together with plausible assumptions about the nature of agency, ensures that ‘shall’ has no dual. It is of note that the absence of a discussion of Gibbard’s use of this distinction between two ways of choosing led Schroeder (2008) and Dreier (2006, 2009) to accuse Gibbard (2003) of being unable to distinguish internal and external negation for the deontic modalities. But that distinction in choice types is needed for the deontic modalities precisely so as to account for the difference between these two sorts of negations, and the associated difference between the strong and weak deontic modality. But where there is no corresponding difference in negation (and no associated duality) for the intentional operator ‘shall’, there is no need to distinguish single-mindedness from indifference in mere intention, whether individual or collective. By not properly accounting for the distinction between the single-minded and the indifferent choice, Sellars (though he was apparently the first to note this distinction) made the mistake of assimilating the intentional modalities to the deontic and was then faced with the puzzle of why there is external negation for the latter. Schroeder and Dreier made the converse mistake, in their reading of Gibbard, by effectively assimilating the deontic modalities to the intentional and then puzzling over an apparent lack of the two different sorts of negation. It is only when we keep the distinction between deontic and intentional expressions in mind, by distinguishing the single-minded and the indifferent choice for the former and not the latter, that we avoid misconceiving things like the interaction between negation and the various modalities in a plan-theoretic semantics. And that, I have argued, is a matter of seeing the deontic judgment and the collective intention as two species of the genus of practical rationality rather than as species of a genus of one another. There is another reason for distinguishing the intentional from the deontic mental state. When I introduced the distinction between single-mindedness and indifference in Sect. 11.2.2, I noted some features of these attitudes that suggest the deontic frame of mind involves more cognitive sophistication than the intentional. Among these was the fact that the ability to discriminate an indifferent from a single-minded choice, via the capacity for rejection, requires that one has attitudes toward other actions that one does not undertake. The intentional state of mind, however, only requires that one be affectively directed toward whatever one decides to do. The ability to discriminate what one rejects from what one rejects rejecting, however, allows one to mark off practical species or determinations from practical genera or determinables (e.g. attending the colloquium in the front or back of the room and attending the colloquium). Once again it appears that the merely intentional mental state, whether individual or collective, is a cognitively less sophisticated form of thinking than the deontic, and this again supports keeping two sets of books on the plans they involve. Finally, concerning the question of whether or not the expression ‘not shall’ makes sense in English, or whether it and ‘shall not’ always have the same meaning, I do not think it obvious one way or the other. But the resources of modal logic help clarify the space of options. We can see this by considering analogous cases with the

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

229

temporal modality ‘will’. Because ‘it will be that A’ and ‘it is not the case that it will be that A’ (using the external negation) are clearly incompatible, we can discern cases where the external negation is not equivalent to an internal negation by looking for contexts where ‘it will be that A and it will not be that A’ (using the internal negation) is coherent. Suppose you are considering coming by my office sometime tomorrow but have not specified when. If you ask me whether I will be in, I might say ‘I will be in my office tomorrow and I won’t be in my office tomorrow’. This sounds sensible where there is an implicit quantification over moments of time, and it does not turn on features of agency. ‘There will be life on Mars and there won’t be life on Mars’ is also acceptable when ‘will’ is understood as quantifying over different moments. Similar uses might be accepted with ‘shall’. Suppose a company is deciding whether to hire a taxi service to take its employees to and from work. When asked whether an employee would use the service, a claim like ‘I shall take the taxi and I shall not take the taxi’, if a bit jarring, seems to make sense in context. As with the temporal modality, its sensibility turns on an implicit quantification over times. I grant that these examples are not dispositive. But we can make these implicit quantifications explicit, and in temporal logic it is common to treat ‘will’ as a modal operator explicitly quantifying over times either universally (‘will-always’) or existentially (‘will-sometimes’  – see the discussion at Belnap, Perloff, and Xu 2001 pp. 32, 152 and 243). Here questions of internal and external negation intersect with the issue of whether a modality has a dual. When explicitly quantifying over temporal moments, ‘will’ operators have duals: e.g., ‘it is not the case that it will-sometimes be that φ’ and ‘it will-sometimes not be that φ’ in general have different semantic values, and the strong modal ‘will-always’ can be defined thereby. There appear to be two ways forward, then. If we treat ‘shall’ as either ‘shall-always’ or ‘shall-­sometimes’ then ‘shall’ will have both a strong and a weak mode and internally and externally negated shall-sentences will receive different semantic evaluations. On the other hand, we can stick to the simple ‘shall’ locution and let our commitments concerning the metaphysics of agency determine that internal and external negations for shall sentences are equivalent.10 I choose the latter option, but principally for simplicity. It may be that ‘shall-always’ or ‘shall-sometimes’ are implicitly used in different circumstances, and attention to individual cases might allow one to discriminate features of context, or of the meanings of individual expressions, that serve to specify one or the other uses. For instance, collateral commitments about what we take actions of this sort to involve make it seem that a claim  It is also possible to eliminate the distinction between internal and external negation while using a quantificational definition for ‘shall’. We do so by requiring that the circumstance parameter for ‘shall’ specifies the moment at which the intention is to be realized, in which case any external negation for a ‘shall-always’ or ‘shall-sometimes’ sentence will be equivalent to an internal negation. Comparison with ‘will’ is again constructive: talk of what is ‘sometimes’ or ‘always’ the case becomes otiose when ‘will’ is evaluated at individual moments, and under a given interpretation two sentences like ‘it is not the case that I will be home at 4:00 pm today’ and ‘I will not be home at 4:00  pm today’ have the same semantic value. See the discussion of the At-instt operator in Belnap, Perloff, and Xu (2001, pp. 152, 242–3). 10

230

P. Stovall

like ‘I shall give to the poor’ either implicitly refers to a particular occurrence of choice (e.g., when I get my next paycheck) or existentially quantifies over circumstances – one can utter a sentence like this without meaning to express commitment to a plan to give to the poor at every point of choice where there is an option to do so. But a sentence like “I shall be more polite to my neighbors” seems to involve universal (or at least general) quantification over circumstances. Features of the meanings associated with the predicates ‘giving to the poor’ and ‘being more polite to one’s neighbors’ may play a role in determining which sort of reading one gets (this bears comparison with Koons’ discussion of material practical inferences in 2019, chapter 6).

11.3.4  Intentional Hyperplans and Non-Akratic Hyperstates Let an intentional hyperplan hI ∈ HI be a maximally determinate plan for what one would choose to do in any circumstance, including what one would do if one were other people. In this case, however, the distinction between indifferent and single-­ minded choices is irrelevant. That is, an intentional hyperplan hI is a maximally consistent plan of action such that, for every circumstance C, every agent α able to make a choice at C, and every action A that α is able to choose to perform at C, either (exclusively): 1. α chooses to A at C on hI; or 2. α chooses not to A at C on hI The set of deontic-intentional hyperstates SDI consists of a set of ordered triples of a world, a deontic hyperplan, and an intentional hyperplan. I use ‘⟨w, hD, hI⟩’ to denote elements of SDI. Finally, I ensure that the claim that one shall do something commits one to the claim that one will by semantically evaluating shall-sentences with regard to non-akratic deontic-intentional hyperstates, denoted ‘SNADI’, where a deontic-intentional hyperstate s ∈ SDI is non-akratic just in case every choice that is part of the intentional hyperplan hI of s is such that the action in question will occur in the world of s (I am suppressing a modal interpretation of sentences like will (φ), instead treating them as ordinary descriptive sentences): s ∈ SNADI iff [s ∈ SDI and ∀α∀A∀C(if α chooses (not) to A in C on hI of s, then ‘α will choose (not) to A in C’ is true at w of s)]11 I use ‘⟨w, hD, hI⟩NA’ to denote non-akratic deontic-intentional hyperstates, which I will call ‘deontic-intentional hyperstates’ from here on. One who rejects the implication that to say one shall do something is to commit oneself to the claim that one will would not use non-akratic hyperstates in the definitions that follow.  Notice that in the antecedent ‘α chooses (not) to A in C’ denotes part of a plan and in the consequent ‘α will choose (not) to A in C’ denotes a part of a world. Notational variation could disambiguate these two uses, but at the cost of a more barbarous text. 11

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

231

11.3.5  A  Hyperstate Semantics for Descriptive and Prescriptive Sentences, Revised The definitions for atomic descriptive sentences and the three kinds of prescriptive sentences can now be given in terms of deontic-intentional hyperstates. ⟦d⟧=def.{〈w,   hD,   hI〉NA: d is true  at  w} ‘people are obliged to A in C’ expresses universally rejecting not doing A in C ⟦people are obliged to A in C⟧ =def. {⟨w, hD, hI⟩NA: for every α, α single-mindedly chooses to A in C on hD} ‘people are forbidden to A in C’ expresses universally rejecting doing A in C ⟦people are forbidden to A in C⟧ =def. {⟨w, hD, hI⟩NA: for every α, α single-mindedly chooses not to A in C on hD} ‘people are permitted to A in C’ expresses universally rejecting rejecting doing A in C ⟦people are permitted to A in C⟧ =def. {⟨w, hD, hI⟩NA: for every α, α either single-­ mindedly chooses to A in C or indifferently chooses whether or not to A in C on hD}

11.3.6  Individual and Collective Intentions As the sentences that give voice to intentions use first-person pronouns, to simplify the presentation I use ‘ι’ as a variable taking either ‘I’ or ‘we’ as its value. Against this background, ‘ι shall A in C’ expresses commitment to the set of deontic-intentional hyperstates where everyone who is one of the group determined by the pronoun chooses to A in C in the corresponding intentional hyperplans: ⟦ι shall A in C⟧ =def. {⟨w, hD, hI⟩NA: ι choose to A in C on hI} Similarly, to assert ‘ι shall not A in C’ is to express commitment to the complement of this set of deontic-intentional hyperstates, which is the set where ι choose not to A in C in the intentional hyperplan of that hyperstate. Notice that I distinguish the class of agents to which an intention applies from the circumstances in which the intention is realized, rather than building the restriction to the class into the specification of the circumstance as I did with prescriptive sentences. One could accomplish the same result by continuing to use a variable ‘α’ that ranges over the unrestricted class of agents and specify that the circumstance in question is one that concerns how either I or one of us would behave, denoted ‘Cι’: ⟦ι shall A in C⟧ =def. {⟨w, hD, hI⟩NA: α chooses to A in Cι on hI} As it is useful to keep track of the agents separately from their circumstances of action when dealing with intentions, I shall stick to the former definition.

232

P. Stovall

11.3.7  S  ome Features of a Deontic-Intentional Hyperstate Semantics Various desirable features of the logic of intentions, prescriptions, and descriptions can be established on this semantics. The proofs proceed by showing that the corresponding subset relations obtain among the semantic values of these sentences under an interpretation. For instance, commitment to ‘ι shall A in C’ commits one to ‘ι will A in C’: Proof: ⟦ι shall A in C⟧ ⊆ ⟦ι will A in C⟧ By the definition of ‘shall’, ⟦ι shall A in C⟧ = {⟨w, hD, hI⟩NA: ι choose to A in C on hI}. Because these triples are non-akratic, it follows that every world in that triple is one where ι will A in C. But that means that the semantic value of ‘ι shall A in C’ is a subset of the semantic value of ‘ι will A in C’. Of course, one may sincerely assert ‘I/we shall A in C’ without oneself or one’s group doing A in C – this is once again a claim about the inferential relations among claims rather than ontological relations among facts. Just so, notice that an assertion of the descriptive claim ‘α intends to A in C’ (which does not express an intention but reports on its existence), even where ‘α’ denotes the speaker, does not commit one to the claim that ‘α will A in C’. This is reflected in the fact that the semantic value of ‘α intends to A in C’ is determined by a set of worlds, and these worlds are not in general a subset of the worlds that determine the semantic value of ‘α will A in C’. At the same time, the claim that one ought to A in C does not commit one to the claim that one will: Proof: ⟦ι ought to A in C⟧ ⊈ ⟦ι will A in C⟧ Because deontic sentences are evaluated with regard to different hyperplans than intentional sentences for any given deontic-intentional hyperstate, and because the non-akratic condition is enforced only for intentional hyperplans, the semantic value of ‘ι ought to A in C’ will include hyperstates with worlds where ι will not A in C. These worlds are not part of the hyperstates that provide the semantic value for ⟦ι will A in C⟧, and so ⟦ι ought to A in C⟧ is not a subset of ⟦ι will A in C⟧. Notice that there is also a non-trivial semantic value for ‘ι shall A in C but ι ought not A in C’. For ⟦ι shall A in C⟧ contains intentional plans where ι choose to A in C and worlds where ι choose to A in C, while ⟦ι ought not A in C⟧ contains deontic plans where ι single-mindedly choose not to A in C. Because the semantics of the former imposes no restriction on the deontic plans determined by the latter, and vice versa, there will be deontic-intentional hyperstates at the intersection of these two sentences. It is true that in such a mental state one is in conflict with oneself over what one intends to do and what one thinks one should do, as exemplified by the fact that there is no decision one can make that satisfies both sets of plans. But alas, that is a feature of the mental life of sublunary beings. Thus akrasia is possible even

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

233

though the hyperstates that model akratic thoughts are non-akratic in the intention position. Furthermore, concerning collective intentions, if we were to follow Sellars and interpret ‘ought’ in terms of ‘shallwe’ it would be hard to make sense of someone who said both ‘we ought to A’ and ‘we shall not A’. But there are clearly cases where one can hold that a collective of which one is a member is about to do something which one thinks they ought not.12 At the same time the semantic value of a complex strong deontic, weak deontic, and intentional sentence like ‘I ought to go to the lecture, I may sit in the front, I may sit in the back, and I shall sit in the front’ is well-defined as the intersection of the semantic values of the separate sentences: ⟦I ought to go to the lecture, I may sit in the front, I may sit in the back, and I shall sit in the front⟧ = ⟦I ought to go to the lecture⟧ ⋂ ⟦I may sit in the front⟧ ⋂ ⟦I may sit in the back⟧ ⋂ ⟦I shall sit in the front⟧ = {⟨w, hD, hI⟩NA: I single-­ mindedly choose to go to the lecture on hD} ⋂ {⟨w, hD, hI⟩NA: I either singlemindedly choose to sit in the front or indifferently choose whether or not to sit in the front on hD} ⋂ {⟨w, hD, hI⟩NA: I either single-mindedly choose to sit in the back or indifferently choose whether or not to sit in the back on hD} ⋂ {⟨w, hD, hI⟩NA: I choose to sit in the front on hI} That this is a practically rational state of mind is reflected in the fact that, unlike the state of mind expressed by ‘I ought to A in C but I shall not’, there is a choice one can make that satisfies all of these sets of plans. For the sets at the intersection of the middle two conjuncts rule out those deontic-intentional hyperstates where I choose, on the deontic hyperplan hD of the corresponding hyperstates, either single-­ mindedly to sit in the front or single-mindedly to sit in the back, and this leaves only those hyperstates whose deontic hyperplans have me choosing indifferently to sit in either the front or the back. In such a decided state the only way to conform both to one’s deontic and to one’s intentional plans is to choose indifferently to sit in the front – to do so, then, is to live up to the commitments one has expressed in making this complex deontic/intentional claim. Finally, the Genuine Agency condition, and the fact that intentional hyperplans do not mark indifference and single-mindedness, ensures that the semantic values for the internal and the external negations of intentions are equivalent: Proof: ⟦it is not the case that ι shall A in C⟧ = ⟦ι shall not A in C⟧ By the definition of ‘shall’, ⟦ι shall A in C⟧ = {⟨w, hD, hI⟩NA: ι choose to A in C on hI}. Call this set X. The semantic value of the negation of ‘ι shall A in C’ is the complement of X. It is a feature of agency, encoded in Genuine Agency, that one can only choose to do something at a circumstance where one could also choose not to do it—otherwise the choice is no choice at all. By the definition of hyperplans, then, if an agent does not choose to A in some circumstance C where she could, then  Reflection on this point first led me to distinguish the deontic from the intentional hyperplan, though I have since become convinced that doing so, and the distinction between single-mindedness and indifference in the former, is central to an adequate understanding of practical rationality.

12

234

P. Stovall

the agent chooses not to A in C. Because intentional hyperplans concern only what the agent chooses to do and not do, it follows that the complement of X is {⟨w, hD, hI⟩NA: ι choose not to A in C on hI}. But this latter set is the semantic value of ‘ι shall not A in C’. By interpreting ‘shall’ in terms of either ‘shall-always’ or ‘shall-sometimes’ this equivalence would not hold.

11.3.8  A Note on Bratman’s Theory of Collective Intentions A more extended comparison of this theory of practical rationality with current work in the literature on collective intentionality is deferred for another time. But there is one aspect of Michael Bratman’s proposal that this account nicely accommodates, and it does so in a way that illuminates a central commitment of Sellars about the identity and individuation of collective intentions. In various places Bratman discusses the norms of rationality he associates with the planning states that lie at the back of collective intentions (cf. 2006, 2014, the Introduction to 2018, and the concise statement of his position in 2017), and he contends that an account of collective intentionality should also satisfy a condition of rational adjustment on these planning states as they develop over time. From (2006, p. 3): [M]y intention that we J…imposes rational pressure on me, as time goes by, to fill in my subplans in ways that fit with and support yours as you fill in your subplans; and vice versa.

In defending the contention that rational pressure forces individuals who share a collective intention to fill in their individual plans in such a way as to realize that collective intention, Bratman writes (2010, p. 8): A fundamental ground of these human capacities for temporally extended and shared intentional agency are human capacities for planning agency. In saying that these planning capacities are a fundamental ground, I mean that the proper exercise of these planning capacities, given relevant contents of the plans, relevant contexts, and relevant inter-­ relations with past, future, and others, will realize phenomena of temporally extended and/ or shared intentional activity.

On the hyperstate analysis provided here, the Boolean logic of set theory goes some way toward explaining how the norms of rationality associated with planning mental states enable those mental states to “realize phenomena of temporally extended and/or shared intentional activity”. This can be seen with an example. Let Sα  ⊆  SNADI denote the set of deontic-­ intentional hyperstates that models the content of the mental state of the agent α. Suppose Alice and Bertolt share a collective intention to get a beer after the conference, but that no preferences for how to get to the pub have been specified (I am suppressing considerations of cognitivism  – the question of whether those who share a collective intention must believe or know that they do so). Setting ‘α’ as

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

235

‘Alice’, ‘β’ as ‘Bertolt’ and ‘C’ as ‘after the conference’ this mental state is modelled by the hyperstate S1 such that: S1 = ({⟨w, hD, hI⟩NA: α chooses to get a beer with β in C on hI} ⋂ {⟨w, hD, hI⟩NA: β chooses to get a beer with α in C on hI}) Now, let us suppose that Bertolt proposes to Alice that they walk to the pub together, and that in doing so he comes to hold his part of the collective intention that they do so. Bertold’s mental state (denoted ‘Sβ’) is now more restricted than S1, imposing as it does the additional requirement that Alice and Bertold walk together to the pub: Sβ = S1 ⋂ ({⟨w, hD, hI⟩NA: α chooses to walk with β to the pub in C on hI} ⋂ {⟨w, hD, hI⟩NA: β chooses to walk with α to the pub in C on hI}) For Alice and Bertold to share this collective intention, Alice will likewise need to restrict her hyperstates to those that include the collective intention to walk to the pub. On this account, the rational pressure to fill in one’s individual plans to satisfy a collective intention (and the explanation of this phenomenon of collective intentionality) is reflected in the fact that more precise collective planning states are given by the intersection of the original states with the planning states that arise as a collective plan becomes more determinate. Finally, notice that this semantics for collective intentions vindicates a claim made by Sellars (1968, p. 217): “The intendings are two in number, but the content of these intendings is the same”. Where Alice and Bertold share a collective intention, and where collective intentions are individuated by the individuals possessing them whereas their content is individuated by hyperstates, Alice and Bertold will have numerically distinct collective intentions with the same content.

11.4  Summing Up and Looking Ahead In this essay I have provided a hyperstate semantics sufficient to distinguish the descriptive, prescriptive, and intentional contents of various sentences and the mental states their assertions give voice to, and I have examined and put to criticism some of Sellars’ views in this vicinity. In doing so I hope to have underwritten my contention that Sellars deserves to be placed alongside the ‘big four’ of collective intentionality as one of the significant figures in this tradition, and that Sellars’ views are deserving of more consideration in the literature on collective intentionality. Beyond its historical relevance, this semantics offers the possibility of a common language for further investigation into issues of rationality, intentionality, and our shared social spaces. For on the view developed here practical reasoning, moral reasoning and individual and collective intentions can all be understood through a common framework that handily accommodates various features of theoretical and

236

P. Stovall

practical rationality. It also, I have suggested, sheds light on the nature of practical rationality as it is manifest in deontic and intentional cognition. Acknowledgements  Work on this chapter was supported by the joint Lead-Agency research grant between the Austrian Science Foundation (FWF) and the Czech Science Foundation (GAČR), Inferentialism and Collective Intentionality, GF17-33808L. Portions of these ideas have been presented at various conferences and workshops over the last 2 years, and I owe particular thanks for discussion with Bill deVries, Ulf Hlobil, Jeremy Koons, Ladislav Koreň, Cathy Legg, Ronald Loeffler, Jim O’Shea, Jarda Peregrin, Mark Risjord, Carl Sachs, and Luz Christopher Seiberth.

Appendix: De Re and De Dicto Modality Consider the de dicto claim ‘it ought to be that every K chooses to A in C’ (e.g., it ought to be that anyone who is a member of my department attend the lecture). We formalize this claim with kind-restricted quantification as follows: Ought(∀K, α)(α chooses to A in C) Intuitively, this says that, no matter who the Ks actually are, anything that were a K ought to A in C. The de re claim ‘every K ought to choose to A in C’ (everyone who is a member of my department ought to attend the lecture), by contrast, says that every actual K is such that it ought to A in C: (∀K, α)Ought(α chooses to A in C) To discriminate the semantic values of de re and de dicto occurrences of the deontic modals, it is correspondingly necessary to relativize semantic values to points of evaluation. Each point of evaluation assigns an object as an extension to every term, a set of objects as the extension of every quantifier, an n-tuple of objects to every n-place predicate, etc. In order to evaluate de re modal claims, principles of identity and individuation are then needed in order to track individuals across different points of evaluation. Consideration of possible-worlds analyses helps illustrate the point. In a possible-worlds semantics the de dicto claim is true at a world w just in case every morally-best world w1 accessible from w is such that the following holds: every K at w1 chooses to A in C at w1. In this case one only looks at how otherworldly Ks behave at those other worlds. By contrast, the de re claim is true at a world w just in case every morally-best world w1 that is accessible from w is such that the following holds: every K at w chooses to A in C at w1. De re modality thus involves a principle of transworld identification – to evaluate de re occurrences of a modal we have to consider what the actual Ks (the individual Ks at the point of evaluation) are doing at other worlds. Even though a hyperstate semantics makes use of plans to evaluate the contents of prescriptive and intentional mental states, in addition to worlds, the difference between de re and de dicto deontic modality can be modelled on a hyperstate semantics if we relativize the semantic evaluation of sentences to worlds alone. Letting

11  Understanding What We Ought and Shall Do: A Hyperstate Semantics…

237

⟦φ⟧w denote the semantic value of φ at a world w, and where ‘@’ denotes the world of evaluation (in the usual case, the actual world), the de dicto occurrence of the strong deontic modality is interpreted as follows: ⟦ought(∀K, α)(α chooses to A in C)⟧@ =def. {⟨w, hD, hI⟩NA: every K in w single-­ mindedly chooses to A in C on hD} The de re occurrence receives the following interpretation: ⟦(∀K, α)ought(α A chooses to A in C)⟧@ =def. {⟨w, hD, hI⟩NA: every K in @ single-­ mindedly chooses to A in C on hD} De re and de dicto talk of permissions and prohibitions can be understood in the same manner. Notice that just as the interpretation of de re modality on a possible-worlds semantics requires tracking individuals from the actual world to various possible worlds, so does de re modality require tracking agents from the actual world to the plan-elements of various hyperstates. But on a hyperstate semantics a principle of identifying and individuating individuals across worlds and plans is already needed for de dicto modality, and in future work I hope to examine such principles in more detail.

Bibliography Belnap, N., Perloff, M., & Xu, M. (2001). Facing the future: Agents and choices in our indeterminist world. New York: Oxford University Press. Bratman, M. (1993). Shared intention. Ethics, 104(1), 97–113. Bratman, M. (2006). Dynamics of sociality. Midwest Studies in Philosophy, XXX, 1–15. Bratman, M. (2010). Agency, time, and sociality. Proceedings and Addresses of the American Philosophical Association, 84(2), 7–26. Bratman, M. (2014). Shared agency: A planning theory of acting together. New  York: Oxford University Press. Bratman, M. (2017). Rational planning agency. Royal Institute of Philosophy Supplement, 80, 25–48. Bratman, M. (2018). Planning, time, and self-governance: Essays in practical rationality. New York: Oxford University Press. Chant, S. R., Hindriks, F., & Preyer, G. (Eds.). (2014). From individual to collective intentionality: New essays. New York: Oxford University Press. Dreier, J. (2006). Negation for expressivists: A collection of problems with a suggestion for their solution. In R.  Shafer-Landau (Ed.), Oxford studies in metaethics, volume 1 (pp.  217–233). New York: Oxford University Press. Dreier, J. (2009). Relativism (and expressivism) and the problem of disagreement. Philosophical Perspectives, 23, 79–110. Gibbard, A. (2003). Thinking how to live. Cambridge: Harvard University Press. Klemick, G. (2018). Sellars’ metaethical quasi-realism. Synthese, published online 18 May, 2018. Koons, J. (2019). The ethics of Wilfrid Sellars. New York: Routledge. Koons, J. (this volume). Sellars on rational agency as presupposing collective attitudes. In L.  Koreň, H.  B. Schmid, P.  Stovall, & L.  Townsend (Eds.), Groups, norms, and practices: Essays on inferentialism and collective intentionality. Cham: Springer.

238

P. Stovall

Ludwig, K. (2016). From individual to plural agency: Collective action, volume 1. New  York: Oxford University Press. Mele, A. R. (1989). Intention, belief, and intentional action. American Philosophical Quarterly, 26(1), 19–30. Olen, P., & Turner, S. (2015). Durkheim, Sellars, and the origins of collective intentionality. British Journal for the History of Philosophy, 23(5), 954–975. Peregrin, J. (this volume). Normative mindshaping and the normative niche. In L.  Koreň, H.  B. Schmid, P.  Stovall, & L.  Townsend (Eds.), Groups, norms, and practices: Essays on inferentialism and collective intentionality. Cham: Springer. Schroeder, M. (2008). Being for: Evaluating the semantic program of expressivism. New York: Clarendon Press. Schweikard, D.  P., & Schmid, H.  B. (2013). Collective intentionality. In E.  N. Zalta (Ed.), The Stanford encyclopedia of philosophy. https://plato.stanford.edu/entries/collectiveintentionality/ Sellars, W. (1951). Obligation and motivation. Philosophical Studies, 2(2), 21–25. Sellars, W. (1953). Inference and meaning. Mind, 62(247), 313–338. Sellars, W. (1963). Imperatives, intentions, and the logic of “ought”. In H.-N.  Casteñeda & G. Nakhnikian (Eds.), Morality and the language of conduct (pp. 159–214). Detroit: Wayne State University Press. Sellars, W. (1966a). ‘Ought’ and moral principles. Transcribed into hypertext by Andrew Chrucky in June 2005 from a copy saved by Willem deVries. http://www.ditext.com/sellars/omp.html. Accessed May 9, 2015. Sellars, W. (1966b). Thought and action. In K.  Lehrer (Ed.), Freedom and determinism (pp. 105–139). New York: Random House. Sellars, W. (1967a). Form and content in ethical theory (the 1967 Lindley lecture). Department of Philosophy, University of Kansas. https://kuscholarworks.ku.edu/bitstream/handle/1808/12383/Form%20and%20Content%20in%20Ethical%20Theory-1967.pdf. Accessed 9 May 2018. Sellars, W. (1967b). Science and ethics. In W. Sellars (Ed.), Philosophical perspectives: metaphysics and epistemology (pp. 191–209). Atascadero: Ridgeview Publishing Company. Sellars, W. (1968). Science and metaphysics: variations on Kantian themes. Atascadero: Ridgeview Publishing Company. (1992). Sellars, W. (1969). Are there non-deductive logics? In N. Rescher (Ed.), Essays in honor of Carl G. Hempel (pp. 83–103). Dordrecht: D. Reidel Publishing Company. Sellars, W. (1976). Volitions re-affirmed. In M.  Brand & D.  Walton (Eds.), Action theory (pp. 47–66). Dordrecht: D. Reidel Publishing Company. Sellars, W. (1980). On reasoning about values. American Philosophical Quarterly, 17(2), 81–101. Silk, A. (2015). How to be an ethical expressivist. Philosophy and Phenomenological Research, XCI(1), 47–81. Stovall, P. (2021). Normative attitudes, collective intentionality, and discursive cognition. In L. Townsend, H. B. Schmid, and P. Stovall (Eds.), The Social Institution of Discursive Norms. New York: Routledge. Stovall, P.. (Manuscript). A unified semantics for a language of theoretical and practical reasoning. Stovall, P. (Forthcoming). The Single-Minded Animal. New York: Routledge. Tuomela, R., & Miller, K. (1988). We-intentions. Philosophical Studies, 53, 367–389.

Index

A Acceptance, 8, 11, 25, 29–31, 34, 77, 95, 121, 163–166, 186 Actions, 5–10, 22–24, 32–33, 66–70, 92–97, 120–124, 141–142, 175–178, 184, and chapters 6, 8, 10, and 11, passim Affordances, 148–168 Allocentric reference frame, 152, 155, 167, 168 Assertion, 2–3, 42 (note), 53, 56, 62–65, 72–78, 103, 108, 134, 151, 163–164, 173–174, 179–181, 183, 187, 207–208, 216–217, 221, 223–224, 232, 235 Assessment normative, 107, 163 Attitudes collective, 6, 24, 25, 29, 32, 35, 190–212 deontic, 8, 221, 228 indifferent, 176, 220, 228, 230 normative, 1, 3, 4, 8–10, 62–64, 68–71, 76, 94, 96, 103, 139, 150, 164 single-minded, 220, 221, 224, 226, 228, 230, 233 Authority, 3, 35, 42, 57, 77, 103, 107, 108, 134 Awareness I-awareness, 13, 148, 150, 155, 164 we-awareness, 13, 148–168 we-part awareness, 148–150, 152, 154, 156, 158, 162, 164–166, 168

Belief collective, 10, 22, 23, 25, 27–33, 35, 69, 202 observation, 206 Brandom, R., 1–5, 9, 11, 13–14, 22, 42 (note), 53, 91, 92 (note), 93, 120, 134–137, 204–205 and chapters 4 and 6, passim Bratman, M., 2, 6–9, 67, 101, 103, 105, 106, 109, 149, 216–218, 234–235

B Behavior pattern-governed, 14, 190–199, 203–207, 209–211

D Davidson, D., 60, 92, 103 Deliberation, 4, 23–28, 31–34, 137 Dennett, D., 5, 174

C Causality causal explanation, 131 causal reasoning, 89 Claims making of, 100–116 (see also Assertion) Cognition, 8, 86, 89, 92, 95, 101, 104, 105, 112, 114, 120–124, 154, 216, 236 Commitment joint, 6, 8, 40, 45–47, 49–52, 69, 149, 200, 202, 206 (see also Entitlement) Communication, 8, 9, 12, 14, 55, 60, 62, 95, 111, 120, 124, 139, 141, 143, 153, 164, 173–187, 205 Conceptual (inferential) role semantics, 2, 63, 74, 134 See also Inferential roles Culture, 12, 85–87, 89, 90, 92, 97, 164, 167

© Springer Nature Switzerland AG 2021 L. Koreň et al. (eds.), Groups, Norms and Practices, Studies in the Philosophy of Sociality 13, https://doi.org/10.1007/978-3-030-49590-9

239

Index

240 Desires, 4, 14, 28, 29, 124, 154, 173–179, 181–185, 199 Disagreement and collective subject, 25, 157 and we-awareness, 68, 148, 164, 167, 168 Dispositions linguistic, 15, 22 Dreier, J., 220, 228 E Entitlement to a commitment, 3, 5, 6, 61, 62, 92, 115 Epistemic commitment, 11, 40, 42, 43, 45–47, 49–52, 54, 57, 206 Epistemic responsibility, 11, 44, 47, 48, 53, 54 Expected utility, 176, 177 Experience, 2, 4, 9, 10, 13, 70, 77, 78, 132, 148–168, 177 F Folk psychology, 174 Form (way) of life, 112, 159–163, 165, 168, 204 Fricker, M., 11, 40–55 G Game (practice) of giving and asking for reasons, 1–3, 5, 13, 22, 23, 30, 34, 35, 100, 107, 108, 120, 121, 134–141, 144, 198, 199, 210 Gibbard, A., 15, 218–220, 228 Gibson, J.J., 159, 160 Gilbert, M., 2, 5, 6, 8, 15, 22, 25, 28, 40, 45–47, 49–52, 67, 69, 101, 109, 149, 200–203, 206, 216, 217 Grice, P., 8 Group action, 5, 66, 67, 70, 100, 149, 150, 153, 162, 202, 204–207 H Habermas, J., 106–108 Harmony See also Conservativeness Haugeland, J., 113 Henrich, J., 86, 93 Hornsby, J., 55 I Inferentialism

Brandomian, 115, 137 Sellars, 14 and stereotype-based exclusion, 151 Inferentialist, 8, 10, 12–14, 91, 100–104, 106–108, 115, 116, 121, 134–138, 140, 144, 163, 165 See also Conceptual (inferential) role semantics; Inferentialism Inferential relations, 3, 135, 198, 225, 232 Inferential roles, 2, 14, 63, 74, 76, 100, 134 Inner speech, 181, 183, 186 Intention collective, 66–67, 69, 95–96, 121, 148, 207–211, and chapters 1, 6, and 11, passim group, 156 I-mode, 7, 201 individual, 2, 5–7, 100, 107, 143, 201, 226 joint, 7, 104, 105, 110, 149, 200 shared, 6, 9, 95, 96, 100–106, 110, 113, 141, 153, 155, 234 we-mode, 7, 15, 69, 72 Intentional stance, 5 Intuition, 23, 121, 122, 132, 138 K Kant, I., 60, 207, 216 Knowledge, 6, 7, 9, 11, 34, 35, 41, 60, 62, 63, 66, 70, 71, 105, 120, 149, 154, 155, 160, 206, 220 Kukla, R., 1, 2, 56, 100 L Lance, M., 1, 2, 56, 100, 190 Language and culture, 12, 86 learning (acquisition), 12, 79, 111, 116 rules (norms) governing language (linguistic communication), 60, 210 Language-entry/exit transitions, 3, 193, 195, 199, 207, 209 Language games, 60, 134, 193 See also Game (practice) of giving and asking for reasons M McDowell, J., 192, 197 Meaning normativity of, 2 theories of, 14 as use, 150

Index Mercier, H., 13, 24, 120, 129 Mindreading, 9, 11, 12, 62, 65–66, 79, 86, 87, 105, 113, 174 Mindshaping, 12, 85–97 Modality, 154, 222, 228, 229, 236, 237 Money, 176, 178, 185–187, 202, 203, 206 Moran, R., 41, 42, 54–56 N Naturalism naturalistic theories of reasoning, 86 Negation, 63, 217, 224–229, 233 Niche construction, 12, 87–90 Normativity, 40, 52, 53, 62, 64, 90–93, 103, 107, 109, 115, 163, 165 Norms (rules) explicit and implicit, 70, 78, 164 governing language (linguistic communication), 210 objective, 62, 64, 74, 75, 77, 93, 107, 115, 185–187 O Objectivity, 4, 68, 69, 77–79, 104, 137, 138 P Perception, 14, 123, 153–155, 159–161 Pinker, S., 86, 89 Plan, 121, 122, 218–225, 230, 235 Planning semantics, 220, 228 Practices claim-making, 100–116 normative (norm-governed, rule-governed), 4, 8, 13, 22, 23, 34, 100, 103, 107, 109, 114 social, 1, 2, 4, 5, 8, 9, 12, 14, 35, 60, 100–103, 108, 110, 113, 115, 138, 148, 150, 160, 168, 203, 204 Proprieties inferential, 3, 5, 194, 197, 209 R Rationality, 25, 33, 62, 79, 136, 137, 176–178, 198, 199, 216–219, 228, 233–235 Realism, 160, 173, 175, 185–187 Reason space of, 3–5, 14, 134, 135, 139, 192, 198 Reasoning argumentative theory of, 23–25, 34, 35, 120, 125, 126, 143

241 moral, 15, 132, 216, 218, 226, 235 Representation linguistic, 76 mental, 105, 109, 110 Representational states, 67, 72 Responsibility, 3, 5, 11, 41, 44, 46–49, 53, 54, 77, 133, 134 Rouse, J., 88, 101 Rule-following, 4, 102, 150, 190, 191, 193–196, 207, 211 Rules governing communication, 185, 198 of inference (inferential), 2, 3, 14, 63, 74, 76, 100, 134 social rules, 100, 182 S Schroeder, M., 228, see Manifest image of man Scorekeeping, deontic (practices of), 3, 5, 8, 13, 102 Searle, J., 2, 5, 7, 8, 53, 70, 101, 149, 216 Sellars, W., 2, 10, 12, 14–15, 22, 60, 68 (note), 69, 90–91, 100, 120 (note), 175, and chapters 10 and 11, passim Semantics, 15, 216 Space of reasons logical, 4, 14 Speech acts, 40–44, 47, 49, 50, 53–56, 62, 75 Sperber, D., 13, 24, 120, 125 Status deontic, 8, 11 normative, 3–5, 8, 41, 43, 70, 100, 102, 114 T Testimony assurance view of, 41, 44, 46 collective, 11, 39–57 Thought (thinking) conceptual, 210, 211 propositional, 60, 67 Tomasello, M., 8, 9, 12, 13, 95, 96, 100, 101, 103–106, 110–112, 114, 121, 141–144, 150, 153, 160, 217 Transparency, 27, 31 Trust epistemic, 40–47, 49–52, 54 Trustworthiness, 11, 40, 42–47, 49–52, 54, 57 Truth commitment to, 29, 40, 53

Index

242 Tuomela, R., 2, 5–8, 14, 15, 23, 67, 69, 70, 149, 190, 200–203, 206, 216–218

W Wittgenstein, L., 4, 60, 102, 108, 109, 150, 166, 191

V Value scales, 177, 178

Z Zawidzki, T.W., 86–88, 92, 114, 174