Tuomela on Sociality 3031226259, 9783031226250

Citation preview

PHILOSOPHERS IN DEPTH

Tuomela on Sociality Edited by Miguel Garcia-Godinez · Rachael Mellin

Philosophers in Depth

Series Editor Constantine Sandis Department of Philosophy University of Hertfordshire Hatfield, UK

Philosophers in Depth is a series of themed edited collections focusing on particular aspects of the thought of major figures from the history of philosophy. The volumes showcase a combination of newly commissioned and previously ­published work with the aim of deepening our understanding of the topics covered. Each book stands alone, but taken together the series will amount to a vast collection of critical essays covering the history of philosophy, exploring issues that are central to the ideas of individual philosophers. This project was launched with the financial support of the Institute for Historical and Cultural Research at Oxford Brookes University, for which we are very grateful. Constantine Sandis

Miguel Garcia-Godinez  •  Rachael Mellin Editors

Tuomela on Sociality

Editors Miguel Garcia-Godinez University College Cork Cork, Ireland

Rachael Mellin University of Glasgow Glasgow, UK

ISSN 2947-552X     ISSN 2947-5538 (electronic) Philosophers in Depth ISBN 978-3-031-22625-0    ISBN 978-3-031-22626-7 (eBook) https://doi.org/10.1007/978-3-031-22626-7 © The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed 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. Cover credit : EschCollection / Getty Images This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

For Raimo

Acknowledgements

Firstly, we would like to thank all the contributors for their shared commitment towards this project. Each chapter is a thoughtful and valuable addition to the ever-growing literature on social ontology. We are delighted to have been able to gather this collection of essays and are grateful to have had the opportunity to work with each of the authors. We would also like to thank Constantine Sandis for encouraging us to propose this volume and for his support during the process, as well as everyone at Palgrave Macmillan for making this possible. Finally, and most importantly, we’d like to thank Raimo Tuomela, to whose memory we dedicate this book. His generosity was as inspiring to us as his rigour. We hope he would find this a fitting tribute towards his unparalleled contribution to the philosophy of sociality.

vii

Contents

1 I ntroduction  1 Miguel Garcia-Godinez and Rachael Mellin 2 We-mode  in Theory and Action 11 Raul Hakli, Kaarlo Miller, and Pekka Mäkelä 3 Tuomela  and the Unity of Belief 37 Sara Rachel Chant 4 Joint  Actions: We-Mode and I-Mode 59 Seumas Miller 5 Towards  a Situated Approach of Tuomela’s Theory of Social Practices 79 Judith H. Martens 6 What  Is Collective Acceptance and What Does It Do?105 Arto Laitinen 7 Can  There Be Institutions Without Constitutive Rules?129 Frank Hindriks ix

x Contents

8 Institutional  Proxy Agency: A We-Mode Approach151 Miguel Garcia-Godinez 9 F  rom We-Mode to Role-Mode177 Michael Schmitz 10 Group  Morality and Moral Groups: Ethical Aspects of the Tuomelian We-Mode201 Björn Petersson 11 Tuomela  on Social Norms and Group-­Social Normativity219 Olle Blomberg 12 Cooperation  Rests on Trust—But What Is Trust and How Do We get There?243 Maj Tuomela I ndex267

List of Contributors

Olle Blomberg  Department of Philosophy, Linguistics and Theory of Science, University of Gothenburg, Gothenburg, Sweden Sara Rachel Chant  Purdue University, West Lafayette, IN, USA Miguel Garcia-Godinez  Department of Philosophy, University College Cork, Cork, Ireland Raul Hakli  University of Helsinki, Helsinki, Finland Frank Hindriks  University of Groningen, Groningen, The Netherlands Arto Laitinen  Tampere University, Tampere, Finland Pekka Mäkelä  University of Helsinki, Helsinki, Finland Judith  H.  Martens  Centre for Philosophical Psychology, Department of Philosophy, University of Antwerp, Antwerpen, Belgium Rachael Mellin  University of Glasgow, Glasgow, Scotland Kaarlo Miller  University of Helsinki, Helsinki, Finland Seumas Miller  Charles Sturt University, Canberra, ACT, Australia Uehiro Centre for Practical Ethics, University of Oxford, Oxford, UK TU Delft, Delft, Netherlands xi

xii 

List of Contributors

Björn  Petersson Department of Philosophy, Lund University, Lund, Sweden Michael Schmitz  University of Vienna, Vienna, Austria Maj Tuomela  University of Helsinki, Helsinki, Finland

1 Introduction Miguel Garcia-Godinez and Rachael Mellin

1 Raimo Tuomela on Sociality Raimo Tuomela (1940-2020) was a leading figure within social philosophy and a pioneer of contemporary social ontology. His main research and publications helped define the scholarly agenda of a philosophical investigation of social reality. Working first on (social) scientific realism and (social) action theory, he then developed an unprecedented account of collective intentionality—the kind of intentionality that individual agents require in order for them to perform intentional actions together, as a group. Based on his novel understanding of collective intentionality (or we-­ intentionality), Tuomela undertook a more extensive project: to explain the elements that are constitutive of our social world, from shared

M. Garcia-Godinez Department of Philosophy, University College Cork, Cork, Ireland R. Mellin (*) School of Law, University of Glasgow, Glasgow, Scotland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_1

1

2 

M. Garcia-Godinez and R. Mellin

attitudes, group reasons, and collective actions, to cooperation, social practices, and institutions. In numerous articles, books, and book chapters, Tuomela championed what we identify here as a general theory of sociality, the most fundamental element of which is the characteristic ‘mode’ (I-mode and we-mode) in which group members participate in intentional collective actions. Although his analysis of we-mode collective intentionality became one of Tuomela’s key research topics since the publication of A Theory of Social Action in 1984, his engagement with the philosophical discussion about methodological individualism/holism and the rationality of group reasoning were also continuous parts of his professional thinking. While elucidating every small detail of these topics, Tuomela remained clear about the general picture: we construct our social world based on our acting together. So, for him, as long as we want to understand and clarify the building blocks of our social world, we need to investigate the various ways in which people can perform intentional joint actions. In noticing the importance of groups and groupish things (e.g., group attitudes, group rationality, group agency, etc.), Tuomela established a research group in Helsinki almost entirely dedicated to elucidating the core features of social groups. It at once became one of the two or three most important spots in the world for social ontology scholarship. With this, and the publication of numerous articles and a series of books (viz., The Importance of Us: A Philosophical Study of Basic Social Notions 1995, Cooperation—A Philosophical Study 2000, The Philosophy of Social Practices: A Collective Acceptance View 2002, The Philosophy of Sociality: The Shared Point of View 2007, and Social Ontology: Collective Intentionality and Group Agents 2013) Tuomela soon became one of the best-known and most cited philosophers in the area. Without pretending to cover all or even most of Tuomela’s rich body of work, this volume aims to highlight the continued relevance of his theory. At a moment when social philosophers (including social ontologists and social metaphysicians) are mainly focused on social justice and collective responsibility, the nature of social categories (like race, gender, and disability), and the impact of social media and AI on social (as well as moral and legal) relations, a solid understanding of the way in which people come to do things together is of the essence. And here is where we

1 Introduction 

3

believe that a critical engagement with Tuomela’s view on sociality can provide the required resources.

2 Chapter Summaries We hear first from Raul Hakli, Kaarlo Miller, and Pekka Mäkelä in We-mode in Theory and Action. As three of Tuomela’s closest and longestserving collaborators, their chapter focuses on the development of his theory of group action through the lens of their cooperation with him over the years. After providing a detailed overview of Tuomela’s prolific career, research endeavours, and contributions, each author tackles a central theme in his work which also corresponds to the main topics they analysed in their joint work with him. Miller begins the discussion by outlining his and Tuomela’s work on we-intentions, practical reasoning, and joint action. The starting point for their work was Tuomela’s 1984 book, A Theory of Social Action, which introduced the idea of we-intentions as a unique notion of intention required for social action. Miller presents their joint 1988 paper, ‘We-Intentions’, as elaborating further on we-intentions by way of practical reasoning schemas before analysing what it takes for an individual to hold a we-intention (i.e., for a member of a collective to intend ‘we will do X’). Miller ends his section of the chapter by considering and responding to two challenges raised against their analysis of we-intentions, as well as all of the subsequent research that this joint work sparked. We-reasoning is one such topic that received further attention in Hakli’s joint work with Tuomela. In his section, Hakli situates their work within the backdrop of Michael Bacharach’s recent theory of team reasoning which, inspired by literature on collective intentionality, employed game theory to study decision-making as a group member as opposed to as an independent individual. The principal contribution of his research was in identifying ‘two ways that group membership can affect an individual’s decision-making’ which, Hakli points out, can be associated with Tuomela’s pro-group I-mode and we-mode cooperation. In the rest of his section, Hakli presents the distinctions between Tuomela’s famous pure I-mode, pro-group I-mode, and we-mode in game-theoretical terms

4 

M. Garcia-Godinez and R. Mellin

before showing how we-mode we-reasoning can solve the problem of maximising group utility in a Prisoner’s Dilemma game. He concludes his section by elaborating on Tuomela’s we-mode thinking and considering the importance of we-reasoning in our social world. Collective responsibility is another topic with the potential to significantly impact on our understanding of our social lives as well as our practice of attributing (moral) responsibility to various kinds of groups. In his section, Mäkelä reflects on his joint work with Tuomela which sought to advance a theory of collective responsibility based in the latter’s we-mode account of group action. Here, Mäkelä outlines their view, paying special attention to certain issues he found problematic or unclear and in need of further explication. In particular, as he puts it, all the main tensions in their work boil down to ‘differing inclinations with respect to the individualism—collectivism debate’. He ends his section by presenting several open questions for Tuomela’s view which follow from this point of conflict, before offering his own understanding of what he took to be an ambiguity concerning Tuomela’s position on the nature of agency in we-mode groups. Hakli, Miller, and Mäkelä end their chapter by detailing the legacy of ‘Tuomelianism’ with an exploration of several novel research areas that have spawned from his scholarship and which they are currently investigating. Continuing the discussion around Tuomela’s work on group action is Sara Rachel Chant’s chapter, Tuomela and the Unity of Belief. Here, she focuses on a central element of collective and joint activities: mutual belief, or the beliefs that individuals (group members) have about the beliefs of others (fellow group members). Despite the ‘almost universal’ agreement by theorists that mutual beliefs are crucial for distinguishing between genuine cases of group action and those whereby individuals merely coincidentally behave in a seemingly coordinated way, it remains largely unanalysed in the literature. While Chant notes that Tuomela is an exception to this, she is not completely satisfied by his account, as for him, mutual belief is not one single phenomenon that underpins all the various kinds of collective and joint activities. Instead, it varies in its structure from case to case, or as Chant puts it, ‘there is a potentially infinite range of types of mutual

1 Introduction 

5

belief ’. This results in what Tuomela calls the ‘level problem’—what level of mutual belief is required for the success of collective activities? Yet, Chant observes, Tuomela does not provide an answer to this question, and so in the remainder of her chapter, she sets out to provide a preliminary solution which aims to ‘respect the diversity of collective activities’ while offering a unified analysis of them under the same explanatory framework. Following this is a more critical chapter from Seumas Miller. In his contribution: Joint Actions: We-Mode and I-Mode, he calls into question the fundamental distinction between I-mode and we-mode that Tuomela uses to explain the various forms of intentional attitudes that people have when participating in joint intentional actions. Although this distinction is crucial for appreciating the distinctive character of pro-group I-mode reasoning and full-blown we-mode reasoning, Miller sustains that it is not always clear that there is a sharp contrast between I-mode and we-mode. After exploring Tuomela’s understanding of these two forms of collective intentionality, Miller goes on, firstly, to contrast Tuomela’s we-mode account with his own individualist I-mode view; and, secondly, to argue that since the notion of we-mode cannot satisfactorily be distinguished from a pro-group I-mode, a strict individualist explanation of collective intentionality purely in terms of I-mode is preferable. The next chapter, while still critical of Tuomela’s work, is more optimistic about its chances of success. In Towards a Situated Approach of Tuomela’s Theory of Social Practices Judith Martens aims to build on Tuomela’s account. As she notes at the outset, social practices are a key concept in Tuomela’s theory of sociality, helping us to understand customs, traditions, and institutions, amongst other social phenomena. Martens identifies two main elements in his analysis of social practices that she begins her chapter by investigating: we-attitudes and pattern-­ governed behaviours. By paying particular attention to how Tuomela integrates habits, customs, and skills in his theory, Martens engages with some of the lesser-explored elements of his work. While she applauds Tuomela for acknowledging the role of these elements in his theory of sociality, Martens argues that in his picture, they are too strongly dependent on, and integrated in, a context in which

6 

M. Garcia-Godinez and R. Mellin

we-­attitudes are the central explanatory component. In order to demonstrate the problem here with his view, Martens introduces some elements from the current discussion on emergent coordination and meshed cognition that she will also use to supplement aspects of Tuomela’s theory. By starting a dialogue with his work on social practices and her own framework of habits and skills in joint action, Martens provides an alternative where we-attitudes are sometimes the consequence rather than the starting point of interactions, which in turn offers a novel account of social practices. Aligned with this revisionary mood, in What is Collective Acceptance and What does It do?, Arto Laitinen introduces some puzzles to Tuomela's ‘collective acceptance view’ and then offers his preferred solution. As Tuomela made evident in a number of important publications, the collective attitude of acceptance is an essential building block in the construction of various kinds of social entities; particularly, social institutions. However, both the role and the very nature of such a key element have not been scrutinised so thoroughly. While Laitinen’s aim is to preserve Tuomela’s analysis of collective acceptance along with its explanatory function for a constructivist theory of sociality, he critically examines the consequences that follow from the way in which Tuomela formulates his main theses. More specifically, he discusses whether there can be collective acceptance of institutions when Tuomela explicitly argues that the object of collective acceptance is a sentence (e.g., ‘squirrel fur is money’). By reconstructing Tuomela’s claims, as presented in his 2002 book, The Philosophy of Social Practices, Laitinen sustains that collective acceptance should be understood as a successful notion that expresses the effective acceptance (both as a mental state and as a performed event) of a true proposition about the existence of an institution. Continuing with the discussion about institutions and social practices, Frank Hindriks sets out to distinguish them in: Can There Be Institutions Without Constitutive Rules? Answering this question in the affirmative, as Hindriks does, leads to the collapse of the idea that institutions can be contrasted with social practices based on the former’s dependence on constitutive rules as opposed to the latter’s dependence on regulative rules. Hindriks argues against this idea by introducing his ‘Transformation

1 Introduction 

7

Rule’ of constitutive rules which claims that a regulative rule can be transformed into a constitutive rule by doing little more than adding an institutional term. As a result, the distinction between institutions and social practices cannot be drawn in terms of these two kinds of rules, since it follows from his view that institutions can be constituted by regulative rules. In looking for an alternative way to make this contrast, Hindriks takes inspiration from Tuomela’s account of institutions as norm-governed social practices. This leads him to conclude instead that the distinction is to be made in terms of the normative dimension of institutions, namely, they feature deontic powers (e.g., they have characteristic rights and obligations) whereas ordinary social practices do not. Although a result of this view may appear to diminish the role of constitutive rules in understanding institutions, Hindriks does not wish to say that they have no value: ending on a positive note, he briefly explores their use in capturing the symbolic dimension of institutions. Engaged in a narrower debate in his Institutional Proxy Agency: A We-Mode Approach, Miguel Garcia elaborates on the social phenomenon of proxy agency (the capacity of an agent to act for another) and tries to situate it within Tuomela’s general framework for social ontology. Starting with a detailed discussion of Kirk Ludwig’s analysis of proxy agency in collective action contexts, Garcia introduces some challenges to Ludwig’s deflationary view when applied to institutional contexts and then suggests a robust alternative based on Tuomela’s we-mode approach to collective intentionality. After noting that Ludwig’s notion of conditional we-intentions is not strong enough to account for the irreducible group-based commitment that participants performing institutional actions (including institutional proxy actions) have, Garcia suggests looking at the constitutive elements of Tuomela’s theory of we-mode collective intentions; particularly, the collective acceptance attitude towards group reasons. So, according to his view, to explicate the we-commitment of institutional role holders qua institutional role holders, we cannot do without signing up for a we-­ mode approach to we-intentionality. Taking a step forward towards the application of Tuomela’s theory of collective intentionality to institutional contexts, Michael Schmitz

8 

M. Garcia-Godinez and R. Mellin

discusses in his From We-Mode to Role-Mode the implications of the we-­ mode account for a role-based understanding of corporations and governments as group agents. While conceding the importance and usability of Tuomela’s we-mode approach to collective intentional attitudes, Schmitz objects that it is insufficient to prevent us from thinking of institutions as fictional entities. In challenging Tuomela’s idea that performing in an institutional role (such as a government official) resembles playing a role in a fictional context, Schmitz advances a more phenomenological, subject approach to what he calls ‘role intentionality’. According to his proposal, acting in a role really amounts to ‘perceiving, acting, believing, planning, asserting and directing’ in a role, which ultimately grounds the intrinsic and non-­ fictitious agency capacity of institutional groups. Following up with the rationale of we-mode groups, Björn Petersson, in Group Morality and Moral Groups: Ethical Aspects of the Tuomelan We-Mode, aims to answer whether the implications of holding, as Tuomela does, that group members can act according to the ‘group morality’ without them behaving morally in a proper sense, are sustainable. By focusing on Tuomela’s distinction between social and moral normativity, Petersson answers in the affirmative as he thinks that we can precisify the scope of the group members’ autonomy when engaging in we-mode activities, and so provide a solution to the ethical dilemmas that arise out of competing, group vs moral, commitments. While tracking this solution back to Tuomela’s assertion that the norms characteristic of we-mode groups should not be conflated with moral norms, Petersson questions if there is a way to draw the distinction more clearly. By bringing Tuomela’s theory of we-mode intentions closer to Sellars’ theory of ethical judgments as expressions of we-intentions, Petersson suggests that we can formulate a Tuomelan theory of ethical normativity that explains both the sense of bindingness and interpersonal validity that seem to distinguish moral norms from other social norms. Pressing on a similar point in Tuomela on Social Norms and Group-­ Social Normativity, Olle Blomberg highlights the difficulties of trying to understand exactly what group normativity is within Tuomela’s framework. Although Blomberg accepts that, for Tuomela, group ‘oughts’ are irreducible to individual ‘oughts’ because they are ontologically and

1 Introduction 

9

conceptually dependent on group members’ we-mode attitudes, he attempts to clarify Tuomela’s view by presenting two versions of it. First, Blomberg considers whether Tuomela’s earlier individualistic account of social norms (as presented in The Importance of Us 1995) can make sense of our practice of holding each other accountable for acting in accordance with our group norms. By arguing that this position fails to explain how demands for social norm-compliance directed at ‘dissenting group members’ could ever be justified, Blomberg goes on to explore a second, more collectivist alternative (as developed in Tuomela’s The Philosophy of Sociality 2007). Without endorsing this view in full, Blomberg believes that this collectivist approach can help us articulate what needs to be explained by a successful account of social norms. Finally, we end the book with a chapter from undoubtedly Raimo Tuomela’s closest collaborator, Maj Tuomela, who in Cooperation Rests On Trust—But What is Trust and How do we get There?, presents her theory of trust, developed in order to fortify R. Tuomela’s scholarship. As M. Tuomela reminds us, trust plays a central role in cooperative activities, and thus in sociality, but before we can use it and understand how to further it in social contexts, we need to know what (social) trust is. In her own words, ‘an existence without trust is almost impossible to imagine’. Yet, within the landscape of R. Tuomela’s work, trust had remained an unanalysed notion. In the first part of her chapter, M.  Tuomela presents an analysis of social trust that distinguishes two main types of trust: genuine trust and predictive trust. Roughly, genuine trust is justified trust in that the trustor has subjectively rational reasons to believe that the trustee will respect their rights, and so when trust is betrayed, it is the trustee who is to blame for this betrayal. Predictive trust, on the other hand, is when the trustor ‘falls-into’ trust based on beliefs and predictions about how the trustee will act, and so there cannot be a betrayal of trust. M. Tuomela goes on to introduce in detail her account of rational genuine trust in the form of five jointly necessary and sufficient conditions, focusing in particular on defending the third of these—the ‘goodwill belief ’ which says that ‘we expect those who we trust to act with goodwill towards us’ and is grounded in our mutual respect for each other’s rights. Along the way, M. Tuomela sharpens her account by making further distinctions between different

10 

M. Garcia-Godinez and R. Mellin

types of trust, as well as contrasting these with other similar notions. In the second part of her chapter, she considers and defends three theses on the ‘two-way connection’ between trust and cooperation, as developed in her joint work with R. Tuomela. In sum, this collection of original essays deal with both foundational questions in social ontology (such as the ontology of social groups, social practices, and institutions) and with issues in collective intentionality (including his famous distinction between I-mode and We-mode, as well as his collective acceptance view). By delving deeply into various aspects of Raimo Tuomela’s philosophy of sociality, each chapter helps to showcase the philosophical significance of his outstanding contribution to our understanding of social reality.

2 We-mode in Theory and Action Raul Hakli, Kaarlo Miller, and Pekka Mäkelä

1 Introduction In this chapter we will reflect on Raimo Tuomela’s philosophy of social and group action as we learnt to understand it over the years of cooperation with him. We will focus on topics on which we did joint work with him. Our joint work unsurprisingly built on his general framework of collective intentionality and social ontology, and we will discuss that as well from the perspective of our more specific topics. Tuomela is not commonly considered exactly a bedside reading and out of respect for him we make a humble attempt to clarify some notions and ideas central to his theorising. We try to understand the development of Tuomela’s theorising as path dependent: his background in philosophy of science in our understanding structured his way of doing philosophy of action and social ontology. On the more personal and not so philosophical side we would also like to

R. Hakli (*) • K. Miller • P. Mäkelä University of Helsinki, Helsinki, Finland e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_2

11

12 

R. Hakli et al.

use the opportunity to discuss our collaboration with Tuomela as an example of his account of group action. We indeed functioned as a group, perhaps even as a we-mode group. The chapter will be structured around the topics of we-intentions, we-reasoning, and collective responsibility. We-intentions was the topic of the prominent joint paper by Raimo Tuomela and Kaarlo Miller (1988). “We-intentions” was among the first analytical philosophy papers on collective intentionality, and it contributed strongly to what was to become a growing field in philosophy nowadays called social ontology. In this chapter we will revisit the genesis of the paper and the debate that ensued and reflect upon it with a twist of our group-internal perspective. We-reasoning is reasoning in terms of a “we”, and an interesting special case of it is team reasoning, a decision-making method in which group members select their actions on the basis of the group’s goal. The theory of team reasoning was developed by economists on the basis of philosophical ideas of social action. In our joint work, we understood team reasoning in a slightly different way from the economists, and here we will reflect on some of these differences. Collective responsibility was, perhaps somewhat surprisingly, the only meta-ethical or even political philosophy topic to which Tuomela contributed seriously. In his writings on collective, group or corporate responsibility, Tuomela applied his we-mode account of group action to make sense of collective moral responsibility attributions. Our joint research on this topic reflected upon the interpretation of Tuomela’s position concerning the nature of collective agency along the lines of the collectivism–individualism debate. Indeed, this was also a source of inspiring philosophical tension in our research group over the years. We will try and provide an interpretation of Tuomela’s position both of collective responsibility and collective or group agency.

2 Tuomela’s Career Raimo Tuomela (1940–2020) was one of the most well-known Finnish philosophers. He died at the age of 80. He acted as a professor of practical philosophy at the University of Helsinki for almost 40 years from 1971

2  We-mode in Theory and Action 

13

up until 2008 when he reached the obligatory retirement age, after which he continued getting external funding and leading research projects as an emeritus. Since 2005 he was also a visiting professor at the Ludwig Maximilian University in Munich. Over the years he ran numerous research projects funded by Academy of Finland and other funders, both national and international. As a professor emeritus he was one of the leading figures in the Centre of Excellence in the Philosophy of Social Sciences (TINT) at the University of Helsinki during the years of 2012–2017. Tuomela got his PhD in 1968 at the University of Helsinki and 1969 at Stanford University. He was an active member of the international community, visited international conferences regularly, published almost 200 articles and more than ten books. Tuomela’s influence on the tradition of analytic philosophy in Finland was immense, he maintained Finnish philosophy on the global map in the continuum starting from Georg Henrik von Wright, followed by Jaakko Hintikka and then Raimo Tuomela. Tuomela was one of the founders of the research fields of collective intentionality and social ontology. He was the president of the International Social Ontology Society (ISOS) during 2013–2017. Despite Tuomela’s international orientation, he, however, edited and authored publications in Finnish as well. Indeed, these publications have played an important role in the education of future generations of philosophy in the University of Helsinki (Tuomela and Patoluoto 1976) and even in Finnish high-schools (Koukkunen and Tuomela 1997). Tuomela started his professional career as a philosopher of science professing a version of internal scientific realism which rejects the Myth of the Given (Sellars 1997) and is committed to the scientia mensura thesis, according to which science is the measure of what there is. Tuomela’s take on philosophy and science more broadly can be characterised as science-­ friendly philosophical naturalism, according to which the conceptual and explanatory frameworks that are constructed should be at least indirectly and partially empirically testable concerning their major claims. The final viability test for a philosophical theory is the success of the scientific theories stemming from it. In the 1970s, Tuomela moved on to work on human action and its explanation defending a version of causal theory of action—the purposive-causal theory of human action (Tuomela 1977).

14 

R. Hakli et al.

In the early 1980s, Tuomela pioneered in the philosophical study of social action (Tuomela 1984). This work led to the publishing of the seminal article “We-intentions”, jointly with Kaarlo Miller, in Philosophical Studies (Tuomela and Miller 1988). This paper provided a detailed analysis of the intentions of individual participants in an intentional joint action. It introduced the core idea of we-attitudes more generally and launched the study of collective intentionality that focuses on such phenomena as group action, cooperation, social practices and institutions, and construction of the social world in a broad sense. According to Tuomela, the proper understanding of human sociality requires collective intentionality in the sense of shared we-attitudes. Collectively intentional mental states and actions based on them involve reference to a “we”, a social group capable of collective reasoning and action (Tuomela 2002). Collective intentionality (“aboutness”) expresses a we-perspective, which comes in two modes, I-mode and we-mode. The we-mode is concerned with group-involving states and processes that the group itself has at least partly conceptually and ontologically constructed for itself. The I-mode covers the private level of social life. Collectively intentional states or actions can be maintained or performed in either I-mode or we-mode. The distinction between I-mode and we-­ mode can roughly be illustrated as follows: I-mode action is based on individuals’ private attitudes, and a we-mode action is based on group’s attitudes that are accepted by the group for the group. I can act for the benefit of the group by performing an action X both in the I-mode and in the we-mode. In I-mode, I did X because I aimed at the good of the group and I believed X to be good for our group. In we-mode, I did X because it was our group’s accepted intention to do X. Tuomela’s main thesis is that both modes are needed in adequate theorising about the social world and that the we-mode is not reducible to the I-mode (Tuomela 2007). Despite this collectivism, Tuomela is not professing collective agents in the strict sense; rather than full-blown agents, groups are functional collective entities that can be conceptualised as agents but ontologically are quasi-agents at best. Tuomela has provided detailed analyses of such phenomena as joint action, group action, social institutions, social practices, social norms, social roles and collective emotions in terms of the conceptual apparatus

2  We-mode in Theory and Action 

15

described above. He has a versatile menu of topics on social ontology (Tuomela 2013) and he has also extensively studied normative dimensions of social action discussing collective responsibility.

3 Our Research Group and the Establishment of the Field Raimo Tuomela obtained external research funding especially from the Academy of Finland and used those funds to establish a research group which had several members through the years. In addition to his permanent chair at the University of Helsinki, Raimo Tuomela was appointed Academy Professor between 1995 and 2000 and this position also included funds for researchers. He received several research projects from the Academy of Finland: “Social Institutions, Collective Responsibility, and Trust” (2002–2004), “The We-Perspective, Social Institutions, and Social Change” (2005–2008), “One for All and All for One: Modelling Cooperation and Solidarity” (2010–2013), “Collective Minds” (2014–2016, collaborative project with St. Petersburg State University), and “Corporate Action: A Group Agent Account” (2016–2020). Kaarlo Miller started as a research assistant for Tuomela in the late eighties. Pekka Mäkelä joined the group in the mid-nineties, and Raul Hakli around 2000. The group included several other members over the years, such as, Juha Hatamaa, Antti Kauppinen, Jyrki Konkka, Arto Laitinen, Antti Saaristo, Mikko Salmela, Gabriel Sandu, Matti Sarkia, Maj Tuomela, and Petri Ylikoski. In the course of years and Tuomela’s successful fundraising, externally funded research projects developed into an established research group with regular and somewhat versatile activities such as weekly reading groups, DAAD-projects with LMU Munich philosophers like Wolfgang Balzer and his group in particular, which included hiking, running and swimming. We also had a collaborative project with St. Petersburg philosophers, when it was still possible. All these activities instantiated we-mode in action as it were, which indeed was, for some of us, even more tempting and engaging than we-mode in theory.

16 

R. Hakli et al.

Raimo was the founding father of the Collective Intentionality and Social Ontology conference series. He organised the first Collective Intentionality meeting with Wolfgang Balzer in Munich, Germany, in 1999. This was a small meeting of less than 20 people including the research groups of Tuomela (consisting of himself, Kaarlo Miller, Pekka Mäkelä, and Petri Ylikoski) and Balzer (including Dieter Will and Solveig Hoffmann), Margaret Gilbert, Georg Meggle, and Seumas Miller. This meeting initiated the Collective Intentionality conference series, which from 2000 was organised in even years. The conference series started as a European enterprise but extended to the USA in 2008. Around that time researchers in Europe founded the European Network on Social Ontology (ENSO) which was to complement the Collective Intentionality series by organising meetings in Europe in odd years starting from year 2009. Both conference series keep attracting growing numbers of researchers, and they have effectively merged together. In 2022, the year of writing this piece, the conference was organised in Vienna, Austria with the title “Social Ontology & Collective Intentionality 2022: Psychological, Behavioural, and Applied Perspectives”, and the 2023 conference will be held in Stockholm, Sweden. In 2013, the International Social Ontology Society (ISOS) was established with Tuomela as its first president. Furthermore, a dedicated journal, the Journal of Social Ontology, and a Springer book series “Studies in the Philosophy of Sociality” were established.

4 Our Collaboration with Raimo We all three go a long way with Raimo Tuomela and obviously we came to work on a variety of topics together in the context of various academic activities from weekly reading groups to collaboration with LMU philosophers in Munich and pihlosophers in the State University of  St. Petersburg. Indeed, we can humbly say that we are relatively well read in Tuomelianism, as “commentator” was one central role in the group for each one of us. As it is well known, Tuomela published numerous books and articles. Just to keep up with his pace of writing by way of reading was a busy job. In addition, we did some joint writing with him. In this

2  We-mode in Theory and Action 

17

piece we cannot discuss all the topics of joint work, but instead, we focus on one theme for each. The order of presentation is thematic, and we write in the first-person perspective, each focusing on our own collaborations with Raimo.

4.1 Kaarlo: We-Intentions, Practical Reasoning and Joint Action My joint work with Raimo started after I became his research assistant, and already in 1985 we published a preliminary version of we-intentions (Tuomela and Miller 1985). Its central ideas were derived from Raimo’s work in action theory, in particular from his magnum opus, A Theory of Social Action (Tuomela 1984). Raimo’s basic idea was that because the notion of intention is required to account for individual action, also an adequate theory of social action requires a unique notion of intention, viz. a we-intention. In our joint 1988 paper (Tuomela and Miller 1988), the basic idea of we-intention was introduced by means of two simple patterns of practical inference: W1  1. We will do X     2. I am one of us, of the group G     3. I will do my part of X

Here, the first premise is an expression of a we-intention of the agent, the second expresses his belief, and the conclusion stands for his intention to participate in the joint action X, which eventually would lead to his doing his part. W1 is by no means logically conclusive; the premises do not entail the conclusion. Yet, the idea is that the premises give the agent a reason to participate. Not only does the we-intention figure in simple reasoning of the above kind but also in more complex reasoning such as W2 below: W2  1. We will do X     2. A is one of us     3. Our doing X requires that A does his part of X.

18 

R. Hakli et al.

  4. Unless I do Y (e.g. teach A to do something) A cannot do his part.   5. I will do Y.

As in W1, also here the reasoning starts with an end that the agent endorses, and here, too, the first premise is an expression of we-intention. The we-intention provides an end that the agent endorses. Premises 2, 3, and 4 stand for the agent’s beliefs. The conclusion 5 represents his intention to do what he has to do in order to satisfy his we-intention that is expressed in premise 1. Here, premises 3 and 4 express what the we-­ intending agent considers necessary for the satisfaction of his we-­intention 1. Therefore, he cannot consistently profess his we-intention 1 and his beliefs 2–4 and yet deny the conclusion 5. This is a kind of Aristotelian practical inference, and Kant (2009) expressed the principle roughly as follows: Who wills an end thereby wills the necessary means in his power. Kant took the truth of this principle as “analytic”. If we replace Kantian “wills” with “intends”, we get a principle applying to intention, as well: The intending agent intends what he takes necessary for the satisfaction of his intention. This principle does not apply to wants or desires. As for we-­ intentions, we later discussed reservations concerning the general validity of this principle (Tuomela and Miller 1992). There we discussed free-­ riding and reserve members’ intentions. A reserve member does not unconditionally intend to participate, and even a free-rider candidate can simultaneously entertain a conditional we-intention to participate. The preference structure of the game Chicken can serve as an example of this, whereas even a conditional we-intention is incompatible with a genuine Prisoner’s Dilemma. Given the preference structure of PD, the only viable strategy is not to contribute. The PD-preferences may of course change. Raul will discuss an important case of such preference transformation in the next section. In a Chicken (Hawk-Dove) game, intuitively, the free-rider candidate intends not to contribute, given that the other(s) will contribute, but simultaneously he is also willing to contribute, should it turn out that without his contribution the production of the common good is likely to fail (for details, see, (Tuomela and Miller 1992). Let us, however, concentrate here on the analysis of the paradigm case of unconditional we-intention.

2  We-mode in Theory and Action 

19

Analysis of we-intention: (WI) A member A of a collective G we-intends to do X if and only if 1. A intends to do his part of X; 2. A has a belief to the effect that joint action opportunities for X will obtain; 3. A believes that there is (or will be) a mutual belief among the participants of G that the joint action opportunities for X will obtain. (Later, e.g., in “We-Intentions revisited” (Tuomela 2005), Raimo added a fourth clause: 4. 1 because of 2 and 3)

The point of WI is to provide an analysis of the first premise “We will do X” of the above practical reasoning schemas. Besides being an expression of an intention, “We will do X” could also express what the agent believes they will eventually do, or it could express his prediction of what will happen in the future. This kind of belief or prediction can turn out to be true or false in the future, but it should not be confused with expression of we-intention. When the agent truthfully expresses his we-­intention as “We will do X”, he expresses his present intention, he tells us something about himself. The individual agent is the subject of the we-­ intention, whereas his belief or prediction concerns their joint action, and the subject of the joint action is his group. In addition to intending to do his part, the we-intending agent believes that, first, their joint action is possible (clause 2), and second, that there is a consensus in the group that they can achieve what they aim at (clause 3). A we-intention is a participatory intention. The agent intends to do his part with the purpose of the joint action coming about; he intends to do his part as his part of their joint action. The subject of the intention is a single member. No agent “can” intend to do what he believes impossible, the “can” concerning the joint action is not entailed, he cannot intend to do X alone. What the we-intending agent can intend to do is to participate in X, viz. intend to do his share of the joint action. John Searle (1990) criticised WI on two accounts. He argued that either “A member of a group can satisfy these conditions and still not have a we-intention” (Searle 1990, 404) or that WI is circular. To illustrate his first criticism, Searle gave the following example. A group of

20 

R. Hakli et al.

young economists who believe in Adam Smith’s invisible hand that pursuing their own interests leads to the benefit of all and they intend to act accordingly. Searle’s point was that they satisfy conditions 1–3 of WI, and yet fail to have a we-intention. Accordingly, such “reductive analyses fail”, he claimed. However, in our joint paper, Raimo and I explicitly argued that WI was not meant to be reductive, and later in the paper we gave an account of what “doing one’s part” amounts to, and Searle seems to have overlooked this. For example, the we-intending agent can succeed in bringing about what “extensionally” would count as his part and yet fail to satisfy his we-intention, if the other members fail to perform their parts. In WI, doing one’s part is an “intensional” notion amounting to doing one’s part as one’s part of an intentional joint action, and Searle’s economists clearly fail to satisfy this requirement. As I would now put it, a we-intention is a participatory intention and as such a unique kind of intention in the family of intentions, it is clearly in the intention mode, and not e.g., in the belief or desire mode. It is not reducible to any non-­ participatory intention (plus beliefs, wants, and what have you). When one acts to satisfy his we-intention, he acts with the intention that the joint action comes about. As for the criticism concerning circularity of the analysis, that in the analysandum we are presupposing the we-intention that the analysis was supposed to be about, let me just say this: WI is circular to the extent that any adequate analysis in terms of necessary and sufficient conditions is (cf. Paradox of Analysis). It is not viciously circular, but is just meant to elucidate the unique notion of we-intention. We ended the first section of our We-Intentions paper with the following brisk sketch for future research: “given the notion of we-intention (involving the notion of mutual belief ) the notion of intentional joint action can be formulated. With the help of we-intentions, mutual beliefs, and (intentional) joint actions one can characterise social norms. Given the notion of social norm, social roles can be analysed. Next, with the help of roles and we-intentions, one can define a strong, normative notion of a social group. From social groups one can proceed to social organisations, institutions, and finally to the notion of a social community without adding any holistic social notions (or—if you prefer to call

2  We-mode in Theory and Action 

21

our we-intentions holistic—any supraindividual notions) to the conceptual basis of the analysis” (Tuomela and Miller 1988, 372). In his later work Raimo investigated all these topics. Here we cannot discuss them in any further detail. What is missing from the above list is the notion of we-mode, and the study of we-mode groups, we-mode attitudes and we-mode reasoning. The notions “collective commitment” and “collective acceptance”, that are central in Raimo’s later writings, were not explicitly discussed in the 1988 paper, although most of the basic ideas are present there, too. Raimo’s later ideas around we-mode were for the most part consistent with my views, although I was often inclined to lean towards a more individualistic approach. For example, concerning the “existence” and agency of we-mode groups, Raimo was, in my view, dangerously flirting with obscure collectivistic ideas. Another difference of emphasis concerned the mode-content issue. Much of what Raimo was willing to account for by the mode of intention, I was inclined to put in the content of intention. In my view, it is the content of the intention that individuates the intention, it tells what the intention is about and how it is meant to be satisfied, and there is no limit to what can be put in the contents of the intentions and respective beliefs. If Raimo, by a change of intention mode, would change the intention, then, in my view, this change must somehow be reflected in the contents of the respective intentions and beliefs. In my view, a change of intention is a matter within the intentional horizon of the agent, not outside it. I hope that the above rough sketch gives at least some idea of some relevant differences between our approaches. These differences created fruitful tension and continued constructive discussion and debate that we both enjoyed, at least most of the time.

4.2 Raul: We-Reasoning When I joined the group, the book Cooperation (Tuomela 2000) was new, and Raimo gave me the task to work on presenting some of his ideas on I-mode and we-mode cooperation to the Artificial Intelligence (AI) community. One of his aims was to apply his theories of social action to

22 

R. Hakli et al.

Distributed Artificial Intelligence, robotics, and social simulation, and I was coming from computer science and had worked a bit on AI-related topics. We published one conference paper together (Hakli and Tuomela 2003), and although we focused on ideas related to preference transformations, we also cited some papers by economists Robert Sugden (2000) and Michael Bacharach (1999) who had developed the theory of team reasoning. They had been influenced by earlier literature on collective intentionality and attempted to use the tools of game theory to study decision-making as a group member, in contrast to decision-making as an independent individual. We didn’t realise their importance immediately and our paper didn’t really break new ground. For me it was mostly just a rehearsal for writing semi-philosophical papers after having only written computer science papers before. However, the study of team reasoning literature opened up a new avenue for research which came to fruition years later when Raimo was invited to speak on a memorial conference on Michael Bacharach’s work in Trento, Italy, after which we started our joint work on a paper and asked Kaarlo to join us (Hakli et al. 2010). Raimo had realised that Bacharach in his 1999 paper “Interactive team reasoning” discusses two ways that group membership can affect an individual’s decision-making (Bacharach 1999), and that these two ways can be associated with Raimo’s pro-group I-mode cooperation and we-­ mode cooperation, respectively. In both cases, an individual first adopts the group’s preferences (which in Raimo’s terminology become her we-­ mode preferences) and then chooses the best actions in light of these preferences, given her beliefs about other agents and their preferences, using standard individualistic best-reply reasoning. In the we-mode case, after adopting the group’s preferences, the individual adopts the group’s point of view and asks what the group should do in light of these preferences. Finally, she infers what is her part in this group action, and decides to do that. For me, Raimo’s distinction between pure I-mode, pro-group I-mode, and we-mode were easier to understand in game-theoretical terms, probably because it was a simplification and not entirely equivalent with his more precise definitions of the different modes: In pure I-mode, an agent selects her actions based on her private preferences that may not be affected by the group in any way. In pro-group I-mode, an agent

2  We-mode in Theory and Action 

23

undergoes a preference transformation and adopts the group’s preferences before selecting her actions. In the we-mode, she undergoes the same preference transformation, and then another transformation, which Bacharach called the agency transformation, and adopts another way of making decisions: Instead of reasoning individualistically in terms of “What should I do in this situation?”, she asks “What should we do in this situation?” Once it is clear what we should do, namely, once everyone’s part actions have been determined, it is then easy for her to select her part as her action. Assuming that everyone reasons in this way, the group manages to reach the best outcomes for the group. Bacharach proved an important theorem in his paper: The solutions that resulted from using both preference transformation and agency transformation (that is, team reasoning) were a subset of solutions that resulted from mere preference transformation (which Bacharach called team benefactor reasoning). Raimo’s idea was that this means that we-­ mode reasoning can restrict the solutions more effectively than pro-group I-mode and hence there are situations in which groups whose members act in the we-mode can reach a determinate solution unlike those groups whose members act in the pro-group I-mode. Bacharach (1999, 2006) illustrated the power of team reasoning by looking at some collective action dilemmas like the Prisoners’ Dilemma: PD C D

C 3,3 4,−4

D −4,4 1,1

Like many philosophers, social scientists, and game theorists, he had been wondering about the puzzle present in PD where individual utility maximisation leads to suboptimal results for all, and the fact that people in experiments didn’t get caught in the dilemma but were able to cooperate. He was aware of the idea of social preferences, or the idea that the numerical utilities didn’t actually represent people’s actual all-things-­ considered utilities which were affected by other players’ utilities as well, and hence should be transformed accordingly in order to represent the actual decision-making situation that the players face. This would require some kind of a preference transformation, and the most natural (or

24 

R. Hakli et al.

cooperative) one to consider was a transformation where everyone transforms their preferences into the welfare maximising preference formed by averaging the individual preferences, thus representing the group preference. Such a preference transformation turns the PD into a Hi-Lo game, in which the individuals’ preferences are perfectly aligned and there are two coordination points which are both Nash equilibria, but one is strictly better than another: HiLo H L

H 3,3 0,0

L 0,0 1,1

The astonishing observation made by Bacharach was that the Hi-Lo game was not uniquely solvable by game theory in spite of its intuitive obviousness (and also later experimentally demonstrated obviousness): One should select H. This conclusion cannot be inferred by using standard game theory, because both HH and LL are Nash equilibria and hence both H and L are allowed action alternatives for rational individual players. We illustrated this situation with the following kind of a failed attempt of practical reasoning: 1. You intend to maximise group utility. 2. I intend to maximise group utility. 3. If you choose H, my choosing H maximises group utility. 4. If you choose L, my choosing L maximises group utility. 5. Therefore, I will perform…? This was still we-reasoning in our view because it was based on group utility, that is, “our” preference, but it was in the I-mode, because the agents reason as individuals, each anticipating what the other might do and trying to come up with a best response. Because they are in a symmetric situation in which the best action depends on the other one’s action, they are unable to coordinate their actions and fail to reach a determinate conclusion.

2  We-mode in Theory and Action 

25

The agency transformation, and Raimo’s we-mode we-reasoning, can easily solve this problem: As soon as we think in terms of the group as the agent (“What should we do?”), we can infer that HH is the best solution for the group, and hence I, as an individual group member, should choose H. This kind of “top-down” reasoning seems to be the most distinctive feature of the we-mode. We are viewing our group as the agent by attributing attitudes to our group, and then inferring our own (we-mode) attitudes from those of the group: 1. We intend to maximise group utility. 2. We maximise group utility only if we choose HH. 3. Therefore, we will choose HH. 4. Therefore, I will choose H. The idea here is that group preferences and preference transformations alone are not sufficient to reach a conclusion, but we need to conceptually think of the group as the agent that has intentions and can select between alternatives: The group can select outcomes whereas individuals can only select strategies. Once the group has selected an outcome, the individuals can select strategies that lead to that outcome. A clear difference in the way we conceived of team reasoning as compared to the way it had been understood by others (Bacharach 2006; Gold and Sugden 2007) was that our version started from expressions of we-intentions whereas others had suggested that team reasoning is a method of forming collective intentions. Whereas the economists were inspired by the idea that people spontaneously adopt a group’s point of view and as a consequence start thinking in terms of a group agent, for us, the group perspective presupposed a pre-existing group with collectively accepted group attitudes like group intentions from which the team reasoning process could start. This, I think, is the key issue in Raimo’s idea of we-mode thinking: People construct groups by collectively forming groups, and by collectively accepting that these groups have various kinds of attitudes, like belief-type and intention-type attitudes. Once this is done, they can we-­ reason, by taking these attitudes as reasons for action, or, equivalently, as premises in their practical reasoning that leads to action. Such

26 

R. Hakli et al.

we-­thinking, or thinking in terms of “we prefer”, “we intend”, etc. makes a causal contribution to decision-making and enables people to coordinate their actions and act as one group. Because the conceptualisation in terms of group agents with group attitudes makes a difference to their behaviour, correct action explanations must refer to such group concepts. Hence, Raimo argued, even though it is individuals who act, social sciences must employ group concepts—even if only in the thought contents of the individuals—if they want to give an accurate account of what is happening in the social world. So, in a nutshell, the importance of we-reasoning (and of team reasoning) can be summarised in the following: People construct the social structures, including groups, group attitudes, institutions, etc., by collective acceptance. This is a bottom-up process in which individuals create the social world. The social world, on the other hand, affects their behaviour in various top-down ways, and here we-reasoning is a central mechanism: The collectively accepted groups and group attitudes like beliefs, goals, and values, etc., figure in their reasoning that leads to action, and have a causal effect on the ways they behave (often leading them to maintain or strengthen existing structures and to create new ones). We-reasoning in this picture is one of the main mechanisms by which the social world affects people’s behaviour: It creates group-based reasons for action and guides people to coordinate their individual actions in accordance with collectively accepted goals, values, and beliefs.

4.3 Pekka: Collective Responsibility I had the privilege to collaborate with Raimo Tuomela for a quarter of a century since 1995. First as his research assistant and then later as a Ph.D. student and eventually as a member of his research group. My main task as a research assistant was to read and comment on the manuscript of The Importance of Us (Tuomela 1995). When The Importance of Us was published, Georg Henrik von Wright, another significant Finnish philosopher, was happily surprised that Tuomela had moved on to do political philosophy, having read the title as referring to the significance of the United States. After the research assistant phase, we rather quickly

2  We-mode in Theory and Action 

27

started writing joint papers as well, of which we ended up writing six altogether. Our joint topics were social action (Mäkelä and Tuomela 2011), collective action dilemmas (Tuomela and Mäkelä 1998) and most importantly collective and group responsibility (Mäkelä and Tuomela 2002, 2003; Tuomela and Mäkelä 2016, 2020). Obviously, everything that I wrote together with Tuomela was based on or an attempt to further develop his conceptual framework for understanding social action and social ontology more broadly. Perhaps one might say that we were in the business of building, or at least trying to build, an account of collective or group responsibility on the basis of Tuomela’s we-mode account of group action. This collaboration over the years led to some academic publications but personally more importantly to a warm and meaningful friendship. All my academic work has been somehow related to either Tuomela’s work or been inspired by his firm but yet wisely implicit supervision. The anecdote of von Wright was, if not funny, mildly amusing, because Tuomela, despite his deep and broad interest in social action and social ontology, perhaps somewhat surprisingly, was not much into moral or political philosophy in his research. However, rather early on, in the history of the collective intentionality “movement”, Tuomela saw that the normative dimensions, e.g., trust and responsibility, are and will be interesting and important extensions of the focus of the collective intentionality research. Indeed, he encouraged me to write my thesis on collective responsibility, against my wish of writing it on free will, which was “way too demanding a topic to my brain” according to Raimo, and right he was. Tuomela was extremely well and broadly read in the wide range of philosophical topics but as a very cautious (Finnish) man he had a rather strict focus in his own writing. Raimo predicted many things which we younger researchers were arrogantly suspicious about, for instance in the turn of the millennia, Raimo encouraged some of us to invest in simulation in social sciences (Mäkelä et al. 2002) and research on AI, little did we know back then. I think the joint work on a piece in The Routledge Handbook of Collective Responsibility edited by Saba Bazargan-Forward and Deborah Tollefsen in 2020 (Tuomela and Mäkelä 2020) which turned out to be the last published article by Tuomela, was also a representative concentration of our joint philosophical work as I remember it.

28 

R. Hakli et al.

That is why, below, I will focus on that piece and the process of bringing it about. In what follows, I will take the liberty to make claims that are based on our private discussions and not anchored in academic writing. I will also interpret some part of Tuomela’s work in the way it seemed to me, no claims of objectivity whatsoever will be made. I discuss some issues that I found problematic or in need of further discussion and debating in our joint work. This discussion, even if brief, hopefully sheds some explanatory light to our joint publications. I always thought that Tuomela saw our work on group responsibility primarily as a test case for his theory of group action. What will follow from the account of group action to the central social practice of holding agents (morally) responsible? The plausibility and implausibility of the implications of the theory of social action and group action to the analysis of moral responsibility in collective contexts were to be evaluated in terms of reflective equilibrium method and to reflect back on the theory of group agency. Tuomela was a self-acclaimed science-friendly philosophical naturalist, and accordingly, he thought that at the end of the day philosophical accounts are tested by their capacity to inspire empirically supported and corroborated scientific theories in the field of research in question. This is roughly an expression of Tuomela’s commitment, which he never at least publicly renounced, to scientia mensura thesis, according to which the best explaining scientific theories determine the ontology. Our work on moral responsibility was seen along similar lines by him. Indeed, I think that work was a useful exercise and had an impact on his theorising. One of the permanent topics of debate in our collaboration concerned the question of how we should go about the “methodological” choice, inspired by the implications of the we-mode account to the practice of holding agents morally responsible, between either fiddling with Raimo’s interpretation of the group agency claims, or rather their strength, or going a bit revisionist about the criteria of agency required by moral responsibility. I was in favour of picking the first horn, but Raimo in my understanding was hesitant of both of the horns. The metaphorical terminology, which appears also in our joint work, the talk about fictitious aspects, distinction between extrinsic and intrinsic, are all reactions to the undecidedness about the way to go here.

2  We-mode in Theory and Action 

29

In my view, there is a distinction to be made in two different sources of motivation for Tuomela’s we-mode account, which sometimes makes it difficult to get a good grip of the nature of his argumentation as these sources of motivation are not always kept clearly separate. One such source is of a normative origin, say “nordic or social democratic view” of society, social groups and social action, “we are in the same boat and we have to contribute to the joint enterprise”. The “value charge” of we-­ mode is by definition positive. The other source of motivation is more metaphysical or conceptual, if you like, which stems from the idea that we need the strong conceptual apparatus of we-mode to understand some social phenomena. Along these lines the aim, of course, was to argue that the we-mode is necessary for explaining and understanding some class of social phenomena. Sometimes these sources of motivation led to “thick argumentation” having both normative and descriptive elements. In his wonderful seminal collective responsibility paper “Responsibility Incorporated” (Pettit 2007), Philip Pettit introduces an admirably simple and clear argumentative strategy. First, he introduces the severally necessary and jointly sufficient conditions for someone to be fit to be held morally responsible in a given choice: value relevance, value judgement and value sensitivity condition. He then goes on to argue that there are collective agents such that they do satisfy the criteria pretty much in the same way as individual human agents do. I admire the strategy but disagree with the conclusion, indeed, I have tried to argue against this view claiming that no account of collective agents provides us with collective agents such that it would be fair to hold them morally responsible in their own right (Mäkelä 2007). Raimo also agreed upon the strategy. Our hypothesis was that the best candidate for a group fit to be held morally responsible (in its own right) would be a group that acts in the way satisfying the conditions of we-mode action of Tuomela’s theory. I would think that this is true, and obviously so did Raimo. However, we did not agree upon whether the best candidate is good enough after all. In my view, we-mode action does not provide us with the grounds to ascribe moral responsibility to the acting group itself, but it provides us with the very strong grounds for a collective or group responsibility in the sense of binding all the group members to the moral responsibility for the group action interdependently. Raimo was not quite happy with my rendering.

30 

R. Hakli et al.

In my view he was oscillating. I simply thought that either the group is capable of bearing moral responsibility or it is not. In our discussions, I boyishly claimed that Raimo was rather of the opinion that in a sense the group is and in a sense it is not. Consequently, we spent hours and hours discussing how Raimo would and should answer such questions as: What exactly is the sense in which groups are morally responsible? What is it to be extrinsically responsible? What is it to be fictitiously responsible? What are the implications of such claims about extrinsic or fictitious responsibility for our actual practices of holding agents morally responsible? Raimo perhaps somewhat reluctantly agreed to hold the position according to which “truthmaker” of a collective attribution of moral responsibility is moral responsibility borne by the individual members of the group, qua members of the group. A part of our debate concerned the qualification “qua members of the group”—what difference does it make and what is its exact import into the discussion concerning moral responsibility for collective actions. My collaboration with Raimo was spiced up with our mild tensions, which luckily were purely philosophical, and unfortunately probably due to my incapacity to get properly on the train of thoughts driven by a seriously deeper mind than mine. Basically, the tensions had to do with differing inclinations with respect to the individualism—collectivism debate. I was never able to fully nail down Raimo’s exact position with respect to the nature of agency of we-mode groups. In my understanding, the we-mode firstly is an attribute of intentional action, of reasoning, or perhaps of holding and maintaining attitudes. I would like to think that we-mode action leads or may lead to groups being reason-giving structures rather than agents. Individual members create via their collective acceptance, decision-making procedure and commitment to group attitudes that share the ontological status of institutional facts. I, as an individual agent, am responsible for my commitments and acceptances both individual and collective, and if I participate in the maintenance of a reason giving structure, then I am to a degree responsible for ensuing actions. I cannot see that we could argue for group level responsibility with these resources. This was a permanent topic of our discussions of collective responsibility over the years. In my view Raimo was oscillating between a rather collectivist rendering of the

2  We-mode in Theory and Action 

31

we-mode groups and a more “realistic” rendering that can be characterised as interrelationist individualism of Raimo’s early work in The Importance of Us (Tuomela 1995). This boiled down to the question of whether we-mode groups are agents capable of bearing moral responsibility in their own right, that is, independently of the responsibility of the individual members of the group. My take was and still is that only individual human agents either individually or jointly can be bearers of moral responsibility (Miller and Mäkelä 2005). The oscillation and tension mentioned above, in my view, revealed themselves in our joint papers (Mäkelä and Tuomela 2002), also in the one that turned out to be Raimo’s last published work (Tuomela and Mäkelä 2020). It seems to me that we ended up claiming that group responsibility is joint responsibility of the individual members and yet, the group itself is morally responsible. Unfortunately, we ran out of time and never got a chance to settle all the tensions, but this provides me with a reason to keep studying what Tuomela really thought.

5 Future Endeavours Raimo, as one of the founders of the field of collective intentionality, has planted the seeds for the growth and development of social ontology, and thus his influence on a wide range of current research agendas is immense. As his students, we have also personally continued working on several themes that were inspired by Raimo’s work. We, Kaarlo and Raul, have been trying to fill a lacuna in the team reasoning literature concerning the nature of group preferences. Theories of team reasoning employ a notion of relative goodness for a group, and this is what the notion of collective preference tries to make explicit. Hence, theories of team reasoning postulate collective preferences, but what they are and how they are arrived at is left unaccounted for. We have tried to understand the different ideas concerning group preferences provided by Raimo, Robert Sugden (2000), Margaret Gilbert (2001), and others, and tried to connect these ideas to general ideas concerning preferences in economics, in particular with those of Daniel Hausman (2011). We could not really fathom Raimo’s ideas about people having two sets of

32 

R. Hakli et al.

preferences, I-mode and we-mode preferences, simultaneously, but thought that there can only be one all-things-considered preference ordering, which, however, may vary according to the situation, e.g., depending on the group context in which a person happens to be operating. Until now, we have presented several conference talks about collective preferences over the years, but the manuscript still awaits submission. We, Pekka and Raul, were inspired by Raimo’s writings on artificial agents (groups and corporations), in somewhat similar vein as Raimo saw the fruitfulness of the analogy between individual intentional action and intentional group action, we have tried to make sense of robots and robotics, and responsibility issues in that field in terms of accounts of intentional action both individual and social. In a way, both groups and AI systems can be seen as artificial, constructed agents whose agency is dependent on the agency of individual agents. Of course, there are differences between them as well, but the basic analogy inspired an argument according to which, similarly to group agents (Mäkelä 2007), robots and AI agents cannot be taken to be autonomous agents and hence they are unfit to be held morally responsible (Hakli and Mäkelä 2019). Even more recently, we have tried to apply Raimo’s theory of we-mode action to planning of action sequences in the case of both human agents (Hakli and Mäkelä 2017) and artificial agents (Hakli 2017). Currently, we are working on an experimental project in which we study whether it would be possible to use a social robot as a facilitator that could help group members to make better collective decisions in a team-reasoning fashion.

6 Concluding Remarks Raimo Tuomela was internationally recognised as a remarkable philosopher who pioneered in the fields of collective intentionality and social ontology. He made a long career as a professor of practical philosophy in University of Helsinki, and he fathered a research group that worked together with him for decades and whose members still continue studying the topics that he initiated. We had a wonderful journey with Raimo that started when we were young students and sincerely scared of him. He had the reputation of a harsh, no nonsense professor, but later on we

2  We-mode in Theory and Action 

33

grew seriously less scared of him. Indeed, for very many years Raimo was a supportive, reliable, very good friend who we can truly respect and whom we sincerely miss. Acknowledgments  This research has been supported by Academy of Finland and NordForsk.

References Bacharach, Michael. 1999. Interactive Team Reasoning: A Contribution to the Theory of Co-Operation. Research in Economics 53 (2): 117–147. https://doi. org/10.1006/reec.1999.0188. ———. 2006. Beyond Individual Choice. Princeton: Princeton University Press. https://press.princeton.edu/books/ebook/9780691186313/beyond-­ individual-­choice. Gilbert, Margaret. 2001. Collective Preferences, Obligations, and Rational Choice. Economics & Philosophy 17 (1): 109–119. https://doi.org/10.1017/ S0266267101000177. Gold, Natalie, and Robert Sugden. 2007. Collective Intentions and Team Agency. The Journal of Philosophy 104 (3): 109–137. Hakli, Raul. 2017. Cooperative Human–Robot Planning with Team Reasoning. International Journal of Social Robotics 9 (5): 643–658. https://doi. org/10.1007/s12369-­016-­0377-­4. Hakli, Raul, and Pekka Mäkelä. 2017. Planning in the We-Mode. In Social Ontology and Collective Intentionality, ed. Gerhard Preyer and Georg Peter, 117–140. Cham: Springer International Publishing. https://doi. org/10.1007/978-­3-­319-­33236-­9_9. ———. 2019. Moral Responsibility of Robots and Hybrid Agents. The Monist 102 (2): 259–275. https://doi.org/10.1093/monist/onz009. Hakli, Raul, Kaarlo Miller, and Raimo Tuomela. 2010. Two Kinds of We-Reasoning. Economics and Philosophy 26 (3): 291–320. https://doi. org/10.1017/S0266267110000386. Hakli, Raul, and Raimo Tuomela. 2003. Cooperation and We-Mode Preferences in Multi-Agent Systems. In Proceedings of the Workshop on Social Life at ECAL 2003, ed. P.  Dittrich and J.  Kim. Dortmund, Germany: University of Dortmund.

34 

R. Hakli et al.

Hausman, Daniel M. 2011. Preference, Value, Choice, and Welfare. Cambridge and New York: Cambridge University Press. Kant, Immanuel. 2009. Fundamental Principles of the Metaphysics of Morals. Cosimo, Inc. Koukkunen, Kari, and Raimo Tuomela. 1997. Lukiolaisen filosofia. Yhteinen kurssi. niin & näin. Mäkelä, Pekka. 2007. Collective Agents and Moral Responsibility. Journal of Social Philosophy 38 (3): 456–468. https://doi. org/10.1111/j.1467-­9833.2007.00391.x. Mäkelä, Pekka, Thomas Pitz, Solveig Hofmann, and T. Chmura. 2002. How Can We Avoid Traffic Jams?: Simulating the Social Practice of Joint Ride. 3rd Workshop on Agent-Based Simulation, 50–54. Mäkelä, Pekka, and Raimo Tuomela. 2002. Group Action and Group Responsibility. Protosociology: An International Journal of Interdisciplinary Research 16: 195–214. ———. 2003. Ryhmän toiminta ja vastuu. In Etiikka ja Talous, 372–386. WSOY. ———. 2011. Sosiaalinen toiminta. In Sosiaalisen toiminnan perusta, ed. Tuija Kotiranta, Petteri Niemi, and Raili Haaki, 87–112. Helsinki: Gaudeamus Helsinki University Press. Miller, Seumas, and Pekka Mäkelä. 2005. The Collectivist Approach to Collective Moral Responsibility. Metaphilosophy 36 (5): 634–651. Pettit, Philip. 2007. Responsibility Incorporated. Ethics 117 (2): 171–201. https://doi.org/10.1086/510695. Searle, John R. 1990. Collective Intentions and Actions. In Intentions in Communication, ed. P.  Cohen, J.  Morgan, and M.E.  Pollack, 401–415. Cambridge, MA: Bradford Books, MIT Press. Sellars, Wilfrid. 1997. Empiricism and the Philosophy of Mind. 2nd ed. Cambridge, MA: Harvard University Press. Sugden, Robert. 2000. Team Preferences. Economics and Philosophy 16 (2): 175–204. https://doi.org/10.1017/S0266267100000213. Tuomela, Raimo. 1977. Human Action and Its Explanation: A Study on the Philosophical Foundations of Psychology. Synthese Library. Dordrecht, Holland: D.  Reidel Publishing Company. https://link.springer.com/book/ 10.1007/978-­94-­010-­1242-­3. ———. 1984. A Theory of Social Action. Synthese Library. Dordrecht, Holland: D. Reidel Publishing Company. ———. 1995. The Importance of Us: A Philosophical Study of Basic Social Notions. Stanford, CA: Stanford University Press.

2  We-mode in Theory and Action 

35

———. 2000. Cooperation—A Philosophical Study. 2000th ed. Dordrecht and Boston: Springer. ———. 2002. The Philosophy of Social Practices: A Collective Acceptance View. Cambridge University Press. ———. 2005. We-Intentions Revisited. Philosophical Studies 125 (3): 327–369. https://doi.org/10.1007/s11098-­005-­7781-­1. ———. 2007. The Philosophy of Sociality: The Shared Point of View. New  York: Oxford University Press. https://doi.org/10.1093/acprof: oso/9780195313390.001.0001. ———. 2013. Social Ontology: Collective Intentionality and Group Agents. Social Ontology. Oxford University Press. http://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199978267.001.0001/ acprof-­9780199978267. Tuomela, Raimo, and Pekka Mäkelä. 1998. Sosiaalisen toiminnan filosofia ja kollektiivisen toiminnan ongelmat. In Yhteisö. Filosofian näkökulmia yhteisöllisyyteen. SoPhi. Yhteiskuntatieteiden, valtio-opin ja filosofian julkaisuja, ed. J. Kotkavirta and A. Laitinen, 199–215. ———. 2016. Group Agents and Their Responsibility. The Journal of Ethics 20 (1): 299–316. ———. 2020. A We-Mode Account of Group Action and Group Responsibility. In The Routledge Handbook of Collective Responsibility, 65–77. Routledge. Tuomela, Raimo, and Kaarlo Miller. 1985. We-Intentions and Social Action. Analyse & Kritik: Zeitung Für Linke Debatte Und Praxis 1985 (7): 26–48. ———. 1988. We-Intentions. Philosophical Studies 53 (3): 367–389. https:// doi.org/10.1007/BF00353512. ———. 1992. We-Intentions, Free-Riding, and Being in Reserve. Erkenntnis 36 (1): 25–52. https://doi.org/10.1007/BF00401963. Tuomela, Raimo, and Ilkka Patoluoto, eds. 1976. Yhteiskuntatieteiden filosofiset perusteet 1–2. Gaudeamus.

3 Tuomela and the Unity of Belief Sara Rachel Chant

1 Introduction Raimo Tuomela observes that: There are many kinds of collective and joint activities, as we all know. We can jointly write a paper, carry a table, sing a duet, and perform a toast to somebody. We can collectively conserve energy, vote in elections, love or fear supernatural beings, follow norms, create and uphold social institutions … But corresponding to the aforementioned various kinds of collective activities also the joint or collective intentions on the basis of which they are performed are different. (Tuomela 2000, p. 39)

Clearly this observation is correct—there is a wide variety of ‘collective activities’ and they differ from each other in important ways.1 Furthermore,  I shall follow Tuomela in using the phrase ‘collective activity’ to refer generally to any form of action performed by a group. 1

S. R. Chant (*) Purdue University, West Lafayette, IN, USA © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_3

37

38 

S. R. Chant

because the intentions of the group (either ‘joint,’ ‘collective,’ or what-­ have-­you) are tied closely to the generation of the collective activity, we also find a correspondingly wide variety of these intentions. Thus, the writer on ‘collective activities’ faces a difficult methodological question right at the outset: Is it better to try to construct different theories for each type of ‘collective activity’ on a case-by-case basis or attempt a single, unified theory encompassing most or all of these types of collective activity? I shall refer to these as the ‘granular’ and ‘holistic’ strategies, respectively. Each approach has its own benefits and drawbacks. The granular strategy respects the wide diversity of collective activity and has the additional benefit of yielding theories with well-defined boundaries around specific types of activity. The holistic strategy, if successful, would yield a general theory; all else being equal, generality is an explanatory virtue. However, it could turn out that there simply is no such holistic theory to be had—that the diversity of collective activities prevents this. The granular strategy is clearly favored by Tuomela and the vast majority of writers on collective action. Theories of collective activities tend to be rather narrowly focused around specific types. Although I am not aware of any explicit discussion of the granular strategy per se, I suspect it is implausible to many people that such a wide range of phenomena as ‘collective activity,’ or even a large subset of these activities, can be brought under one explanatory umbrella. I shall argue here that we can respect the diversity of collective activities while constructing a more general and unified theory than is currently available. The explanatory strategy is straightforward. We start with Tuomela’s ‘Bulletin Board View’ (2005), which relies heavily on a concept that Tuomela refers to as ‘mutual belief.’ Then we show that well-­ motivated refinements to this concept of ‘mutual belief ’ yield explanations of other types of collective activity that are not otherwise explained by the Bulletin Board View. In this way, the Bulletin Board View forms an explanatory framework and the various types of collective activity are well-motivated corollaries of the framework. The plan for this chapter is as follows. We start with a brief discussion of the Bulletin Board View. Then in subsequent sections we examine the question of what it is for a group to believe something. This general

3  Tuomela and the Unity of Belief 

39

concept of ‘group belief ’ is shown to be of central importance to a wide range of influential theories of collective activity, including the Bulletin Board View. Next, we examine how Tuomela approaches the question of group belief, eventually ending up with his distinctive concept of ‘mutual belief.’ Importantly, we shall see that (as Tuomela himself notes) ‘mutual belief ’ is not a single phenomenon; rather there is a potentially infinite range of types of ‘mutual belief.’ However, Tuomela and others leave open the central question of how these different types of ‘mutual belief ’ apply to the wide range of collective activities; this question is referred to by Tuomela as the ‘level problem.’ Importantly, it turns out that Tuomela and others leave the level problem unanswered. We shall argue that the level problem is at least partially answered by considering the contexts of the various types of collective activity, in particular, by understanding how groups structure themselves and their communications in response to the risks they face for miscoordination, and the cost of communication within their group. And finally, we show that this preliminary answer to the level problem, combined with Tuomela’s Bulletin Board View, yields a much more general theory of collective activity that has surprisingly broad explanatory power.

2 The Bulletin Board View The Bulletin Board View (BBV) is a metaphor and conceptual framework for thinking about we-mode joint action. Tuomela offers a simple story about how a group of people might jointly clean a park together next Saturday. In the story, a particular person comes up with a plan to clean the park and initiates the proposed joint action by writing the proposal on a public bulletin board, including a note that anyone who wishes to participate should sign their name on the bulletin board. Because the bulletin board can be viewed by the public, people who sign their names can see the names of other people who have committed themselves to cleaning the park. Thus, the people who sign their names know that others have also committed themselves to cleaning the park on Saturday. In principle, people could come back to the bulletin board to repeatedly view the list of names; in this way, they will acquire beliefs about the

40 

S. R. Chant

beliefs of other people who have signed the bulletin board. Furthermore, the people who sign the bulletin board have implicitly committed themselves to cleaning the park on Saturday, creating a normative obligation to follow through with their commitment. They also have a reasonable expectation that others who are similarly committed will follow through and do their part of cleaning the park on Saturday. Thus, the signing of the bulletin board creates for each person a reason to clean the park (Tuomela 2005). The central feature of BBV is the public accessibility of the bulletin board. Because everyone will see the names of the other people who have signed up to clean the park, each person will form beliefs about the beliefs of other people. So if Alice, Bob, and Carol sign up in that order to clean the park, then Carol will know that Bob believes that Alice will participate in cleaning the park (because she can assume that Bob saw Alice’s name when he signed up). Alice could even come back to view the bulletin board later and thereby come to believe that Carol believes that Bob believes that Alice will clean the park, and so on. In principle, there is no upper limit to the number of iterations of beliefs about beliefs that this process could generate. Tuomela calls these beliefs about beliefs, ‘mutual beliefs.’ He holds that mutual beliefs should play a central role in explaining collective actions or activities. This is reasonable because of a common methodological assumption that Tuomela endorses, namely, that group-level concepts have similar explanatory roles to their corresponding individual-level concepts. For example, individual intentions play an important role in the production of individual actions. Hence, we should expect collective intentions to play an equally important role in the production of collective actions or activities. As Tuomela puts the point: The notions of goal, belief, and action are linked in the case of a group to approximately the same degree as in the individual case. In the latter case, their interconnection is well established; given that the person-analogy applies to groups …, these notions apply to groups as well. (Tuomela 1992, p. 287)

3  Tuomela and the Unity of Belief 

41

In other words, because we treat groups much as we treat individual persons, and because with respect to individuals, the concepts of ‘goal, belief, and action’ are interconnected, we should expect that the group-­ level correlates of ‘goal, belief, and action’ should display similar linkages with respect to groups. If this is the right approach, then the concept of group belief should occupy a central role in any account of collective action, activity, intention, and other related concepts. After all, individual agents’ beliefs are central to the generation of intention and action. Thus, we might expect that group beliefs must play an equally important explanatory role in accounts of other group-level concepts. And what BBV illustrates is that the appropriate form of belief for explaining joint action and other collective activities is what Tuomela calls ‘mutual belief,’ which we will examine in much more detail below.

3 Group Belief and Its Importance As we’ve noted already, analyses of group-level concepts of action, intention, and agency are motivated by an analogy to their individual-level concepts. For example, consider the concept of a ‘group intention.’2 Among other desiderata, a satisfactory account of this concept should maintain some similarity between the intentions of groups and the intentions of individuals. Similar remarks apply equally well to other group-­ level concepts such as action and belief. The theoretical desideratum of maintaining an analogy to individual-­ level explanations of action is not the only reason why an understanding of group belief is so important. It is also clear to the vast majority of writers on this topic that group belief is one of the crucial features that most clearly distinguishes between cases of people who coincidentally behave in a seemingly coordinated way, and genuine cases of collective action. To take a standard example, when a group of people all coincidentally show up at the same restaurant at the same time, we would not say that they are  As with ‘collective activity,’ I will use the term ‘group intention’ to refer agnostically among a wide range of related concepts. 2

42 

S. R. Chant

engaged in a collective activity in any substantive or interesting sense. But if they have a group belief that the others are going to show up at a particular time, and that belief plays the right kind of role in bringing about their coordinated behavior, then we are well on our way to concluding that they have indeed performed a collective action.3 Although writers on collective action agree almost universally that some form of group belief involving beliefs about others’ beliefs is central to group action, it is curious that these writers often leave the concept unanalyzed. Margaret Gilbert and Michael Bratman are two good examples of this. For example, Gilbert offers the following set of conditions for a belief to be held by a group, which she calls the ‘joint acceptance’ account: (i) A group G believes that p if and only if the members of G jointly accept that p. (ii) Members of a group G jointly accept that p if and only if it is common knowledge in G that the individual members of G have openly expressed a conditional commitment jointly to accept that p together with the other members of G. (Gilbert 1987)

Here, ‘joint acceptance’ is being defined by Gilbert in terms of ‘common knowledge,’ which found its way into the philosophical literature via David Lewis’s work, ‘Convention’ (Lewis 1969). Citing Lewis, Gilbert mentions that ‘common knowledge’ is a technical term. However, she claims that ‘[e]xactly how to define it is somewhat moot’ and makes no further effort to clarify her use of it (Gilbert 1990, p. 14). Michael Bratman puts himself in the same boat with respect to common knowledge. He offers an account of what he calls ‘shared cooperative activity’ that relies on group belief. According to Bratman, the following conditions for an act-type J to be a shared cooperative activity are: (1)(a)(i) I intend that we J.  Of course, this does not by itself entail that the group has performed a collective action. For now, the point is only that the group’s having the right kind of belief is an important mark of collective action. 3

3  Tuomela and the Unity of Belief 

43

(1)(a)(ii) I intend that we J in accordance with and because of meshing subplans of (1)(a)(i) and (1)(b)(i). (1)(b)(i) You intend that we J. (1)(b)(ii) You intend that we J in accordance with and because of meshing subplans of (1)(a)(i) and (1)(b)(i). (2) It is common knowledge between us that (1).4 (Bratman 1992)

It is clear that something closely resembling Gilbert’s definition of group belief is at work here, since Bratman requires everyone to share beliefs about each other’s beliefs insofar as they relate to their J-ing. But what are we to make of Bratman’s use of ‘common knowledge’ in condition (2)? The answer is not very satisfying. He cites the same work by Lewis on common knowledge. But, like Gilbert, he goes on to disavow Lewis’s definition, stating instead that he is using ‘an unanalyzed notion of common knowledge’ (Bratman 1993, p. 103).

4 Tuomela and Group Belief It should strike us as very odd that these two influential theories each make crucial use of the concept of ‘common knowledge,’ but that neither offers an analysis of the concept. And it is even more odd that they both cite an influential and precise treatment of the concept, only to disclaim it. Tuomela does better on this score, and his work suggests why characterizing this concept is so difficult. On Tuomela’s account, no single, precise form of group belief is sufficient to underpin the wide range of collective activities. His preferred notion of group belief is what he calls, ‘mutual belief,’ which refers to the beliefs that individuals have about the beliefs of others. In this section, we will situate Tuomela’s ‘mutual belief ’ concept within the context of a broader literature from game theory and economics. This context will be useful because observations from that literature can help us to understand why it is so difficult to pin down the exact conditions required for mutual belief across a range of cases. In so doing, it suggests  I have omitted two conditions related to coercion and stability, which are not relevant to the present discussion. 4

44 

S. R. Chant

why some writers on this subject have conspicuously shied-away from attempting a more precise formulation within their own theories.

4.1 Interactive Knowledge Beliefs about another person’s beliefs are standardly referred to as ‘interactive knowledge’5 in economics and game theory literature. As we have noted above, philosophical work in collective action has long recognized that interactive knowledge is a vital component of any account of collective action or collective intention. For it is clear to many writers that across a wide range of cases, a group member’s beliefs about the beliefs of her fellow group members are at least part of what distinguishes a collective action from a mere set of actions performed by various individuals. For this discussion, it will be important for us to be as precise as possible about the concept of interactive knowledge. So we shall make a few distinctions and lay out some terminology. Consider a group of individuals and a proposition p, which all the members of the group believe. In order for the group to have interactive knowledge, it is necessary for the individuals at least to be aware that the others also believe p. Consider the statement: (1) Everyone knows that everyone knows that p.

We shall refer to this as ‘first-order interactive knowledge that p’ because there is exactly one iteration of ‘everyone knows that’ in front of the statement that ‘everyone knows that p.’ Additional levels of interactive knowledge are generated by repeatedly prepending ‘everyone knows that …’ to the statement. For example, if everyone knows that (1) is true,

 In the philosophical literature on collective action, discussions are usually centered around the beliefs of the agents. In the relevant literature outside of philosophy (e.g. economics and game theory), the discussion is typically concerned with interactive knowledge. For the present paper, I will not bother distinguishing between beliefs that p and knowledge that p. This amounts to assuming that the agents’ beliefs are true, and that they satisfy whatever criteria need to be satisfied in order for them to count as knowledge. Whatever happens when agents have erroneous beliefs, or beliefs that aren’t knowledge, is outside the scope of this paper. 5

3  Tuomela and the Unity of Belief 

45

then the group has second-order knowledge that p, and so on, ad infinitum. Two types of interactive knowledge are often singled-out for specific treatment. First-order interactive knowledge is typically referred-to as ‘mutual knowledge’ (although we must be careful to remember that Tuomela does not follow this use of ‘mutual’). The other type is when a group has interactive knowledge for every finite level. In such a case, the group has so-called ‘common knowledge.’ Between these two extremes, there are infinitely many levels of interactive knowledge that are at least theoretically possible. We often refer to these other levels of interactive knowledge as ‘interactive knowledge to degree (or level) n.’ With this terminology in place, we are now ready to examine some of the types of collective activity that Tuomela discusses, and better understand how ‘interactive knowledge’ or ‘mutual belief ’ play an important role in theorizing about these forms of collective activity.

4.2 The Simple We-Belief Account The first type of activity performed by a group is one whose members communicate with each other, but which does not necessarily have a formal structure. These are ad-hoc groups without a hierarchy of authority, formal roles, etc. Their activities include the earlier example of people meeting at a restaurant, two people walking together, lifting a heavy table together, or painting a house together. For such a group, Tuomela endorses what he calls the ‘simple we-belief account.’ The appropriate form of group belief for this account is that a group G believes that p if: (M) for every a in G, a believes that p and a also believes that Bmp. (Tuomela 1992, p. 308)

where ‘Bmp’ is understood to mean, ‘everyone in the group has a mutual belief that p.’ For the present discussion, we will leave the question of understanding Tuomela’s sense of ‘mutual belief ’ until later. For now, it suffices to say that having a mutual belief that p is to have interactive knowledge to some degree or other about p. Tuomela’s condition

46 

S. R. Chant

requiring mutual belief for various forms of collective activity is guided to a significant extent by the same intuition that is endorsed by Gilbert, Bratman, and others.

4.3 The Positional Acceptance Account The positional acceptance account applies to groups that have a formalized structure in which different people play different roles. It recognizes that within some types of group, judgments and decisions are left to a subset of individuals, and the others are under a normative obligation to accept those judgments and to act accordingly. There are many real-world cases in which the positional acceptance account fits well. We say, for example, that a large corporation ‘believes’ a proposition if the board of directors has collectively judged it to be true, perhaps by voting in accordance with the procedure specified in the corporation’s charter. Or a nation’s government may be said to ‘believe’ a proposition in this sense if the parliament has endorsed it by whatever procedure is specified in that nation’s constitution. Clearly, in order for a corporation or a government to believe something, it is not necessary for everyone in that corporation or government to share the belief; rather, the belief is attributed to the entire group if the right subset of individuals have that belief. In order to make that idea more precise, Tuomela distinguishes between the members who have the authority to accept a proposition on behalf of the group, and those who are under an obligation to accept those beliefs: (BC’) G believes that p in the social and normative circumstances C if and only if in C there are operative members A1,…,Am of G in respective positions P1,…,Pm such that: (1’) the agents A1,…,Am, when they are performing their social tasks in the positions P1,…,Pm and due to exercising the relevant authority system of G, (intentionally) jointly accept that p, and because of this exercise of authority system, they ought to continue to accept and positionally believe it; (2’) there is a mutual belief among the operative members A1,…,Am to the effect that (1’);

3  Tuomela and the Unity of Belief 

47

(3’) because of (1’), the (full-fledged and adequately informed) nonoperative members of G tend tacitly to accept—or at least ought to accept— p, as members of G; and (4’) there is mutual belief in G to the effect that (3’). (Tuomela 1992, p. 295)

Tuomela is envisioning that the group G is divided into two sets of individuals, one that he calls the ‘operative’ members and the other the ‘non-operative’ members. The operative members of G are the people who have, according to the ‘relevant authority system,’ the power to make decisions on behalf of the group. For example, the board of directors for a large corporation would typically be ‘operative’ in this sense (at least with respect to certain classes of decisions, the scope of which would be spelled-out in the company’s charter, or by relevant laws). In an extreme case, there might be only one operative member, as in a dictatorship. But typically, a relatively small, manageable group would comprise the operative members. The non-operative members are those whose roles place upon them a normative obligation to act in accordance with the decisions made by the operative members. Most employees in a large corporation, for example, would fall into this group; so would the members of the bureaucracy of a government. It is important to note that this account does not require that the non-­ operative members of G personally agree with the decisions of the operative members. For example, a specific employee may disagree with the decisions made by the board of directors. This is why Tuomela says only that the disagreeing non-operative members of G ‘tend tacitly to accept— or at least ought to accept’ the propositions that are believed by the operative members. So an employee may personally believe that a particular belief held by the board of directors is wrong, but will nonetheless be under an obligation to tacitly accept it. Minimally, this involves acting in accordance with the belief. In other words, to tacitly accept a proposition is to act as if one personally believed it, or at least refrain from engaging in ‘overt opposition which would make it difficult for G to act on the belief set by the operative members of G.’

48 

S. R. Chant

5 Tuomela on Mutual Belief Although the simple we-belief account and the positional account differ in important ways, they share an extremely important central feature. They require that the individuals in the group (or at least a subset of them) have what Tuomela calls ‘mutual beliefs’—that is, interactive knowledge to a given finite degree. He says: … let us analyze mutual belief in terms of the ‘Everybody believes’-operator Be, applying to the full-fledged and adequately informed group members: Bm=Bep & BeBep & …, with a finite number of operator-iterations. (Tuomela 1992, p. 310)

Clearly, Tuomela’s concept of ‘mutual belief ’ is almost identical to ‘interactive knowledge’ with the single difference that he is discussing belief, whereas game theorists and economists typically talk in terms of knowledge. But for our purposes here, the difference between belief and knowledge can be put to the side. The important point is that unlike Gilbert and Bratman who leave their concept of ‘common knowledge’ undefined, Tuomela is here offering an explicit formulation of ‘mutual belief ’ while also allowing that it can vary in its structure from case to case by virtue of requiring different numbers of iterations on the ‘Be’ operator. The problem of how many levels of interactive knowledge are required is what Tuomela calls the ‘level problem.’ Given the iterative account, the problem is how many iterative levels of belief are conceptually, epistemically, or psychologically needed for success in various cases … The level question is difficult to answer in general terms. However, it can be conjectured that in certain cases at least two levels (viz. n = 2), relative to the base level, are needed for mutual belief, but no more. In some other cases this is not enough. (Tuomela 2022, p. 5)

Tuomela does not try to answer the level problem. Instead, ‘mutual belief ’ is left as a placeholder for ‘whatever level of interactive knowledge is necessary in the specific circumstances.’

3  Tuomela and the Unity of Belief 

49

The remainder of this chapter is dedicated to a preliminary solution to the level problem. An important benefit to solving the level problem is that it suggests a strategy for bringing together several different types of collective activity under the same explanatory framework.

6 Answering the Level Problem The presence or absence of interactive knowledge is an important mark of collective action and intention. As we have noted, this observation has been made numerous times and is nearly universally held, even among those whose subsequent analyses differ in important ways. But when a particular observation is so obvious, it is all too easy to take it for granted without asking ‘why is interactive knowledge required at all?’ The answer to the question might seem too obvious to bother stating: if we didn’t have beliefs about each other’s beliefs, then it would be impossible to coordinate our actions. However, there are many commonplace examples showing that it is often possible to coordinate actions without any interactive knowledge at all. For example, David Lewis observed that people often unthinkingly coordinate their actions based on some undefined criterion of ‘salience.’ People simply perform the salient action, and thereby coordinate, without ever considering the beliefs of other people. For instance, consider two people who become separated in a city. It would not be mysterious at all if they were both to go to a local landmark; and they may do so without considering abstruse questions such as ‘does the other person know that I know that we ought to go to this landmark?’ Even simpler examples can easily be found. When two drivers approach a stop light which has a green light on one side and a red light on the other, they will have no problem coordinating their actions. One will drive through the green light, the other will stop at the red light. Unless the drivers are unusually neurotic, their actions will not be motivated by any considerations of interactive knowledge. Of course, it could be objected that the drivers do have some level of interactive knowledge in this case. Both have beliefs about the other’s knowledge of the conventions that govern driving. However, this does nothing to indicate that interactive knowledge plays a role in any specific

50 

S. R. Chant

instance of coordinated behavior at traffic lights. It would be used to justify why people stop at red lights and go on green lights. But there is an important difference between explaining a person’s behavior, and providing a post hoc theoretical justification of it. To make this difference clearer, my beliefs about the behavior of electrons in copper wire do not enter into my behavior when I flip on a light switch. A post hoc justification of my behavior might reference those beliefs. For example, if you were to challenge the rationality of my flipping the light switch, I might say that when I flip the switch, an electric current will be allowed to pass through the copper wire, and this is possible because copper is a conductor of electricity. But those beliefs play no role whatsoever in explaining why I flipped the light switch on this particular occasion. In the same way, a driver’s beliefs about the beliefs of other drivers do not explain coordinated driving behavior.6 If asked to theoretically justify why it’s a good policy in general to stop at red lights, I might go into a lot of detail about how the other drivers have the same beliefs about traffic laws as I do, and that we all know that if the light is green on one side, it will be red on the other, and so on. So in some cases, interactive knowledge is not required for coordinating behavior.7 But in other cases, it is indispensable. To take an example I have discussed elsewhere (Chant and Ernst 2008), if Mark and Bob need to arrange a clandestine meeting, and there would be bad consequences if one were to show up without the other, then a high level of interactive knowledge may be required. They may require not merely that a signal be sent that they should meet, but that a second signal be sent and received in order to assure the other that the first signal was received.8  Which is not to say that interactive knowledge can’t ever explain this behavior in special circumstances. If I notice that you’re distracted while you’re approaching a stop light, then I may believe that you don’t believe there’s a stop light, and my interactive knowledge will explain why I slow down. 7  This leads us to a question that deserves some attention. Is the coordinated behavior of individuals who don’t have interactive knowledge so different from the behavior of those who do, that the former doesn’t deserve the moniker ‘collective action,’ while the latter does? If the distinction turns out to be unfounded, then collective action can (and often is) performed in the absence of any interactive knowledge whatsoever. 8  In an idealized case, the need for additional signals might never end—Mark and Bob may require an infinite number of signals and counter-signals if they are to meet. This is the so-called 6

3  Tuomela and the Unity of Belief 

51

For example, suppose that Mark is to put a flag outside his apartment in order to signal that they should meet. If Mark does this, but has no way of knowing that Bob received the signal, then he may reasonably fear that Bob hasn’t received the signal. In that case, Mark may reason that it is too risky for him to go to the meeting site. Arranging for a verification signal is a mechanism for generating interactive knowledge. In this case, if Mark has a way to verify that Bob’s signal was received, then they will be able to achieve first-order interactive knowledge; both will know that the meeting is to take place, and they will each know that the other knows this. To see that the risk of miscoordination is the important factor in this example, consider a similar case, but with negligible risk. Suppose that Mark and Bob are merely meeting at a restaurant that they often enjoy going to alone anyway, and that the worst possible outcome is merely that one of them will have a pleasant meal alone at the restaurant. In this case, they could easily coordinate, even without arranging any second signal. Mark will have no reason to fear any unpleasant consequences if he were to show up at the restaurant alone, and so he will be perfectly willing to take the risk that Bob didn’t receive the signal. In a slogan, higher risk necessitates higher levels of interactive knowledge. On the other hand, if we increase the risk associated with miscoordination, we can generate cases where even higher levels of interactive knowledge are required. Consider the following variation of a well-known example (Binmore and Samuelson 2001). Two allied generals must coordinate their attacks on the enemy to be launched at the same time. Miscoordination—that is, one general attacking alone—would result in a disastrous defeat. In order to coordinate their attacks, they have pre-­ arranged for two different signals. The first signal indicates ‘attack at noon.’ Upon receiving that signal, each general is to send up a flare that indicates they received the signal. Let’s call the generals ‘Alice’ and ‘Bob.’ Suppose that Alice and Bob each see the signal, and each fires their flare acknowledging that they saw it, and they each see the other’s flare. So the plan is going perfectly thus

‘Coordinated Attack Problem,’ which is identical to the ‘Electronic Mail Game.’ For an excellent discussion of these problems, see (Binmore and Samuelson 2001).

52 

S. R. Chant

far. Each believes they should attack at noon, and each believes that the other believes they should attack at noon. However, there is a small probability that a flare could be fired without the other general seeing it. So it is possible for Alice to reason in the following way: What if Bob didn’t see my flare? Then he might believe I didn’t fire it. In which case, he has no guarantee that I saw the original signal. And if he thinks there’s a chance I didn’t see the original signal, then in his mind, there is a chance I won’t attack at noon. And because of the disastrous consequences of attacking the enemy alone, he shouldn’t attack the enemy at noon. Given that he won’t attack, I shouldn’t attack, either.

Thus, despite the signals being sent and received perfectly according to plan, Alice doesn’t attack. Clearly, Bob can reason in exactly the same way, so in fact, neither of them attacks. This story about Alice and Bob is an informal version of a problem called the ‘coordinated attack problem’ or ‘electronic mail game,’ which has been extensively discussed and analyzed elsewhere. There are several important lessons from that analysis that bear on the present problem. Most importantly, we learn that there are several factors that influence solutions to Tuomela’s ‘level problem.’ That is, there are identifiable factors that explain how many levels of interactive knowledge are required for coordinated group action. Of these factors, two stand out. The first is the level of risk associated with miscoordination; that is, how bad the outcome would be if the group fails to coordinate. The second is the cost of iteratively communicating information back and forth among the members of the group. All else being equal, group actions with higher risk require higher degrees of interactive knowledge.

3  Tuomela and the Unity of Belief 

53

7 Risk and the Cost of Communication in Collective Belief The considerations of the previous section give us the ability to make predictions about how groups will be structured, as well as how the ‘level problem’ can be answered for any specific case. The ability to make predictions and explain the structure of groups is an important mark of a good theory of group action; along the way, these explanations also help us bring different types of group action under the same holistic explanation. First, we note that some groups are structured in such a way as to spread information more quickly than other groups, and at a lower cost. A small business might have a regularly scheduled meeting of all employees to make announcements or discuss policies. Larger and more elaborate groups (such as a military organization or a multinational corporation) often have a hierarchical structure in which specific channels of communication are defined, and information is transmitted up and down the hierarchy according to accepted policies. Such mechanisms are implemented for their efficiency in helping to coordinate the group. A benefit of having these communication mechanisms is that they make high levels of interactive knowledge easier to establish. Consider the simplest possible hierarchy consisting of three people: Alice, Bob, and Carol. Let us say that Carol is at the top of the hierarchy and is the only operative member of the group. Alice and Bob receive information only from her, and when Carol communicates, she simultaneously sends information to both Alice and Bob. If Alice and Bob have interactive knowledge of the group structure to degree n, then they automatically get the same level of interactive knowledge for free, as it were, whenever Carol sends a message to them. For suppose Carol communicates some message to Alice and Bob. Without communicating with each other, Bob knows that Carol must have received the same message. This is because he knows the structure of the group, and therefore he knows that they will both receive the same information. Of course, Alice will reason the same way. So Alice and Bob get mutual knowledge of Carol’s message, even without any communication between them. In fact, Alice and Bob

54 

S. R. Chant

need never have communicated with each other at all. Clearly, this line of reasoning can be iterated as many times as necessary, up to the level of interactive knowledge that Alice and Bob possess about the structure of their group. The observation that specific kinds of group structures lower the cost of interactive knowledge, combined with the fact that risky activities increase the need for interactive knowledge allows us to make a straightforward prediction. All else being equal, groups that are engaged in riskier activities will tend to have structures in place that make interactive knowledge easier to establish. Summing up our observations so far, when the penalty for miscooperation is higher, the more we should expect the group to have structures in place to lower the cost of communication and thereby lower the cost of establishing interactive knowledge. Specifically: 1. Group members who would suffer from miscoordination should be able to establish high levels of interactive knowledge among themselves with little or no direct communication among them. 2. Their ability to establish interactive knowledge is often derived from their interactive knowledge of the structure of the group and their respective roles within it.

These two predictions are very general. It is easy to see that other features are likely to be found in groups that would incur a large penalty for miscooperation. To name just one, deviations from the set communication structure should be strongly discouraged because the availability of alternate channels of communication makes it more difficult to assume that the proper communication protocols have been followed. For example, if Alice and Bob are supposed to receive all their information from Carol, but sometimes they individually communicate directly with someone else instead, then it is difficult for Alice to know that Bob has the same information she does. Interactive knowledge becomes more difficult to establish if there are unofficial communication channels available. Counterintuitively, interactive knowledge may be easier to establish if there are fewer ways to communicate.

3  Tuomela and the Unity of Belief 

55

A very clear example of an organization that suffers large penalties for miscoordination is a military. Doing one’s part in a military action when someone else does not do their part could have devastating consequences for everyone. Establishing interactive knowledge is literally a matter of life and death. Thus, we would expect that the predictions above should hold true for military organizations; and those predictions do in fact hold. All communication channels are strictly defined; communication outside of those channels is prohibited; the structure of the military and the respective roles of each member are common knowledge among everyone; and each person’s role and status is prominently advertised to everyone. Because of these conditions, when low-ranking members of the military receive an important piece of information, they are justified in believing that it is common knowledge among those whose cooperation is necessary. In other words, the common knowledge of the structure of the organization is bootstrapped into common knowledge of specific pieces of information, enabling the group to achieve high levels of interactive knowledge quickly and at little cost. Setting up such structures also carries a cost, of course. For groups whose activities are low-risk, it would be unreasonable to set up complex communication structures and hierarchies. Hence, small groups that are carrying out low-risk activities shouldn’t generally be expected to have much structure. These considerations and examples help motivate at least a partial answer to Tuomela’s ‘level problem.’ All else being equal, higher levels of interactive knowledge (or ‘mutual belief ’) are expected in groups where penalties for miscoordination are very high. Accordingly, for such groups we expect to find mechanisms that facilitate the generation of higher levels of interactive knowledge. These mechanisms include hierarchical communication structures, explicit rules about which channels of communication are permitted, and communication channels that rely on common knowledge of the structure of the group.

56 

S. R. Chant

8 The Holistic Approach We are now ready to reconsider the original question with which we began this discussion: is it possible to construct a ‘holistic’ theory of collective activity that encompasses a wide variety of actions performed by groups? Or does the diversity of collective activity make this impossible, leaving us with a potentially large number of different ‘granular’ theories? This chapter has laid the foundation for a more holistic approach to theories of collective action. To see how, let’s return to Tuomela’s ‘simple we-belief ’ and ‘positional acceptance’ accounts. Recall that the former is generally applicable to simple ad-hoc groups that are coordinating on simple activities (e.g. lifting a heavy table, painting a house, and so on), whereas the latter is generally concerned with more complex activities undertaken by large, structured organizations. To give these two types of groups and activities separate analyses is to miss the most important explanatory commonalities between them. Each group faces the same problem, namely, establishing a high enough degree of interactive knowledge to make coordination possible. The obvious differences between the two accounts are due to the practical considerations that different types of groups face due to their specific circumstances. That is, the exact manner in which groups create interactive knowledge (or ‘mutual belief ’) is determined by balancing at least two important competing considerations: 1. The greater the level of risk faced by the group, the more levels of interactive knowledge will be necessary for coordination. 2. But putting the structures and processes in place that are necessary for generating interactive knowledge also carries a cost.

Thus, the cases encompassed by the simple we-belief account and the positional acceptance account do not require two different treatments. All the seemingly distinctive features of the positional acceptance account—the subset of ‘operative’ members, the hierarchical differences in authority and function within the group, and so on—are simply predictable responses to the different problems that groups face.

3  Tuomela and the Unity of Belief 

57

For example, take two cases that are the subjects of different theories. First, we have an ad-hoc group of two people who want to lift a heavy table together. Second, we have a sprawling, complex military organization that needs to coordinate an attack. Both groups face the same problem: how to establish mutual belief sufficient to justify their members to do their parts of the collective activity. As far as the question of why these cases ought to qualify as ‘collective’ or ‘joint’ or what-have-you, the answer is the same. Their actions are brought about in the right way by a mutual belief. Why the groups have their distinctive structures and procedures for establishing that level of mutual belief is a separate question. But it has a simple, well-motivated answer: the military organization has established its structure and communication channels in order to make high levels of mutual belief cheap and easy to attain because their risks are very high. In contrast, the ad-hoc group of two people who lift a heavy table together are operating in a very low-risk, simple environment. Hence, they do not establish the same structures you’d see in most formal organizations.

9 Conclusion The position I have advocated in this chapter is a natural one, motivated by a commonplace observation about theory formation. Namely, that the ability to explain a wide variety of seemingly disparate phenomena is an important explanatory virtue for a theory. Conversely, if explaining a range of related phenomena seems to require a large number of specific theories, this should count against the explanatory framework, all else being equal. But unfortunately, the latter more aptly describes the situation we find in the literature on collective action—a series of specific, ‘granular’ theories has been proposed to explain a range of closely related phenomena, namely, the actions of various types of group. Thus, this chapter has posed the question, ‘Are there one or more features of the different types of group which (a) makes them appear to require different theories of collective action, but (b) can in fact be given a systematic treatment within a single explanatory framework?’ The answer has turned out to be ‘yes.’ The various features of different types

58 

S. R. Chant

of group—e.g. their organizational structure, decision procedures, formal roles or lack thereof—are simply the predictable responses to a single factor which varies across groups, namely, the level of risk faced by the group as its members attempt to coordinate their actions. This is very good news for theories of collective action because it shows that there is a single explanatory framework (the Bulletin Board View) which encompasses a very wide range of collective actions that so far has required a number of disparate theories. Thus, we achieve a greater degree of explanatory unity than has been possible until now.

References Binmore, Ken, and Larry Samuelson. 2001. Coordinated Action in the Electronic Mail Game. Games and Economic Behavior 35 (1): 6–30. Bratman, Michael. 1992. Shared Cooperative Activity. The Philosophical Review 101 (2): 327–341. ———. 1993. Shared Intention. Ethics 104 (1): 97–113. Chant, Sara Rachel, and Zachary Ernst. 2008. Epistemic Conditions for Collective Action. Mind 117 (467): 549–573. Gilbert, Margaret. 1987. Modelling Collective Belief. Synthese 73 (1): 185–204. ———. 1990. Walking Together: A Paradigmatic Social Phenomenon. Midwest Studies in Philosophy 15: 1–14. Lewis, David. 1969. Convention: A Philosophical Study. Cambridge: John Wiley & Sons. Tuomela, Raimo. 1992. Group Beliefs. Synthese 91 (3): 285–318. ———. 2000. Collective and Joint Intention. Mind & Society 1 (2): 39–69. ———. 2005. We-Intentions Revisited. Philosophical Studies 125 (3): 327–369. ———. 2022. Shared Belief. https://citeseerx.ist.psu.edu/viewdoc/download?d oi=10.1.1.25.1810&rep=rep1&type=pdf. Accessed 21 Aug 2022.

4 Joint Actions: We-Mode and I-Mode Seumas Miller

1 Introduction A core notion in the analysis of joint actions and related social phenomena is that of a so-called collective intention. There are a variety of different conceptions and theories of collective intentions. Moreover, these conceptions and theories tend to mirror different conceptions and theories of joint action; and do so in part because it is held that joint actions are actions performed in accordance with collective intentions.1 Relatedly, it is held that joint actions involve collective reasoning. At any rate, collective intentions are typically contrasted with individual intentions, it  See Searle (1995), Tuomela and Miller (1998), Gilbert (1989), Miller (2001), Bratman (2014), Ludwig (2016). 1

S. Miller (*) Charles Sturt University, Canberra, ACT, Australia Uehiro Centre for Practical Ethics, University of Oxford, Oxford, UK TU Delft, Delft, Netherlands e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_4

59

60 

S. Miller

being generally understood that there is some distinction to be drawn between these two types of intention, even by those of an individualist persuasion. One way of drawing the distinction is by recourse to the terminology of intentions in the we-mode as opposed to intentions in the I-mode. So, according to theorists of a collectivist persuasion, the collective intentions involved in joint actions are intentions in the we-mode (it being understood that joint actions also involve some intentions in the I-mode). In addition, and relatedly, the reasoning involved in the performance of joint actions is held to be reasoning in the we-mode (as opposed to I-mode reasoning). Since the concern in this chapter is with joint actions (and collective intentions and collective reasoning in so far as these notions pertain to joint actions), we initially need a brief description of joint actions. Joint actions are (prima facie, at least) not the singular actions of collective entities, such as the supposed ‘actions’ of the nation-state, Australia. Nor are joint actions the aggregated singular actions of individual human beings, such as the set of actions consisting of Australians getting out of bed in the morning. Rather, joint actions are actions involving a number of agents performing interdependent actions in order to realise some common goal. Examples of joint action are: two people dancing together, a number of tradesmen building a house and a group of robbers burgling a house. Over the last few decades, a number of analyses of joint action (and, relatedly, of collective intentions and collective reasoning) have been provided, of which Raimo Tuomela’s is one of the most important.2 These analyses can be located on a spectrum at one end of which there is so-­ called (see Schmitt 2003) strict individualism, and at the other end of which there is so-called (again by Schmitt) supra-individualism (often referred to as collectivists). The location on this spectrum of any given analysis of joint action depends in large part on the particular conception of a collective intention deployed by that analysis. Roughly speaking, Tuomela is somewhere in the middle of this spectrum in that he endorses a kind of weak supra-individualism (and, therefore, weak collectivism).  Tuomela provided a number of closely related accounts. I take his account in (Tuomela 2013) to be definitive for my purposes here. 2

4  Joint Actions: We-Mode and I-Mode 

61

Specifically, he argues in keeping with supra-individualists that collective intentions (and like collectivist attitudes) are not conceptually reducible to individual intentions (and like individual attitudes). Yet he also argues that these conceptually irreducible collective attitudes are not ontologically existent, causally operative attitudes. Rather he holds, at least in some places (Tuomela 2013, but see below), that his positing of such entities and mental states is based on heuristic considerations (e.g., simplification of reference, utility of the ‘as if ’ stance) and on the (alleged) unavailability of the relevant conceptual reductions to individualistic notions. Having introduced the related notions of joint actions, collective intentions and collective reasoning, and located Tuomela’s perspective on these notions in relation to some of the main competing perspectives, let us now turn in more detail to Tuomela’s account of collective intentions and collective reasoning.

2 Tuomela on Collective Intentions and Collective Reasoning On Tuomela’s account,3 as already stated, we-intentions are central to joint action as in the well-known example of a group painting a house, i.e., the group we-intends to paint a house. The contrast here is with individual attitudes, such as individual intentions said by Tuomela to be in the I-mode (e.g., Jones I-intends to paint his wall). In addition to these we-intentions, i.e., a species of intention in the we-mode, and common or garden individual intentions, i.e., a species of intention in the I-mode, Tuomela introduces what he calls, a pro-group I-mode attitude. According to Tuomela the possessor of such an attitude is “functioning as a “private” person but still at least in part for the group when in a group context” (Tuomela 2013, p. 5). This notion of a “private” attitude of a “private” person (or, relatedly, a “personal” attitude) is somewhat opaque.

 Many of the points made in this section are drawn from my review (Miller 2014) of Raimo Tuomela’s book, Social Ontology. 3

62 

S. Miller

Thus, Tuomela says: “I use the word “private” basically to mean (purely) personal … (i.e., not intrinsically based on a social group’s properties)” (Tuomela 2013, p. 266, n. 13). But the only relevant non-controversial attitudes (and persons) in play, whether one is an individualist or a collectivist, are attitudes such as ‘My intention that we dance’ or ‘My intention that I do my part of painting the house’ (and persons possessed of such attitudes). Such attitudes are in part constitutive of some joint or otherwise group activity, are interdependent with the like attitudes of the other members of the group in question,4 and are a matter of common knowledge among the members of that group. These joint actions and their constitutive pro-group I-mode attitudes consist both of purely individual actions and attitudes (e.g., one single human agent’s contributory individual action and individual intention to perform that action) as well as social actions and attitudes (e.g., the joint action itself and the shared interdependent end constitutive of that joint action—a pro-group I-mode attitude). That is, these joint actions and, in particular, their constitutive pro-group I-mode attitudes are social in some sense. Certainly, these pro-group I-mode attitudes are not ‘private’ or ‘purely personal’ in any ordinary sense of those terms. Indeed, evidently, they are individual human intentions, beliefs etc. which are interdependent, constitutive of human cooperative activity, the motivating reason for the contributory single actions of the human agents in question and publicly known (in the sense of being objects of common knowledge among the members of the group in question). Nor is Tuomela’s suggestion that they are not intrinsically social helpful; it merely seems to beg the question against individualists. Tuomela claims that the I-mode/we-mode distinction is not based on the content of the attitudes in question nor on attitude type, e.g., that it is an intention or a belief (Tuomela 2013, p. 67). So what is this mysterious ‘third dimension’ of mode upon which the distinction rests? Perhaps  As such, they are perhaps instances of what Tuomela refers to as joint intentions. To the extent that his joint intentions refer to such attitudes then they are individualist in character and to be distinguished from we-intentions. I note this in the context of Tuomela’s claim, “The reviewer incorrectly and confusingly speaks of we-intentions and joint intentions in my sense as a group’s intentions” (Tuomela 2014) in his comments on my review (Miller 2014). 4

4  Joint Actions: We-Mode and I-Mode 

63

it is simply that the subject of the attitude in question is not ‘I’ but ‘we’, as Tuomela suggests: “in the we-mode case the intending subject is “we”, and in the I-mode case it is “I”” (ibid.). Apparently, there is no third dimension rather there is a new subject of we-mode attitudes: a collective agent. Naturally, this collective agent will need to be irreducibly collective otherwise individualism reappears by way of intentions of the form ‘I intend that we-x’. Moreover, the mystery of sui-generis we-attitudes has only been dispelled by supplanting it with the mystery of irreducibly collective agents. Elsewhere Tuomela claims that “ontological reducibility or irreducibility are generally a posteriori matters to be decided in casu by empirical research” (Tuomela 2014). However, this does not dispel the mystery since it is unclear what would count as this not yet available empirical evidence of an ontologically irreducible collective agent, i.e., evidence above and beyond the evidence that we currently have. At any rate, Tuomela ultimately cashes out we-mode attitudes in terms of familiar notions in the literature, crucially that of “qua member of [group] g” and “collective acceptance” (2013, p. 68).5 However, the question now arises as to the adequacy of the analysis in terms of these notions and, in particular, the acceptability of the we-mode/pro-group I-mode distinction. Tuomela’s attempt to make good this crucial (for him) distinction between we-mode and pro-group I-mode thinking runs as follows. He says: “collective intentionality (we-mode in my terms) requires that agents intend to act together as a group and thus, according to my approach, for the same authoritative group reason, and also satisfy the criteria or markers of collective commitment and the collectivity condition (a kind of “necessarily being-in-the-same-boat” condition)” (Tuomela 2013, p.  6) and “Comparing I-mode thinking with we-mode thinking, the crucial differences are, respectively, the change of agency from individual to collective (or group agency) and the change of I-reasoning to we-reasoning. These differences also account for the claimed irreducibility of the we-­mode to the I-mode” (2013, p. 7).  The original notion of collective acceptance as fundamental to the construction of institutions and institutional facts was introduced by (Searle 1995). Tuomela’s account of institutions (2013, Chapter 8) is derived from Searle’s, albeit there are differences. See also (Kirk Ludwig 2017). 5

64 

S. Miller

Here there are three separable moves. The first move consists of making the distinction between pro-group I-mode and we-mode simply one of degree—notwithstanding that pro-group I-mode is cast as being in the individualist I-mode rather than in the we-mode. This move consists of explicitly adding additional non-normative conditions to bind the agents in question closer together; additional ‘glue’ is supposedly added to pro-­ group I-mode attitudes/actions. The crucial such condition is that of acting together as a group. However, such ‘additional’ coating is not sufficient to ground a fundamental conceptual difference between pro-social I-mode and we-mode such that the latter is irreducibly and intrinsically social while the former is merely ‘personal’. This is because the ‘glue’ in question may itself be reducible to individualist notions. Importantly, as I argue below,6 acting qua member of a group can be analysed as acting in accordance with an individual end which each agent has interdependently with the others (a shared interdependent end). Moreover, this condition is already implicitly met by many, if not all, pro-group I-mode attitudes. If one acts on the basis of a pro-group I-mode attitude constitutive of a joint action to which one is a contributor, then in doing so one is acting qua member of the group (the group whose members consist of those contributing to the joint action). Tuomela’s second move is to import into the definition of the we-­ mode normative notions which are self-evidently (or at least plausibly regarded as) intrinsically social; specifically, authoritative group reason and collective commitment. However, the question here is whether he is simply offering a stipulative definition of we-intentions as involving institutional and/or moral normative notions of authority and commitment (or some other non-controversially intrinsically social normative notions); or whether the normative notions in question are somehow derivable from that of a we-intention. If the former, fair enough; but the individualist can also make this move by providing pro-group I-mode attitudes and associated joint actions with a normative moral or institutional loading, e.g., specifying that pro-social I mode attitudes have moral or institutional normative  And have argued in detail elsewhere (see Miller 2010, pp. 52–54, 2013).

6

4  Joint Actions: We-Mode and I-Mode 

65

content or that there are normative constraints (e.g., laws) on certain joint actions or that their ends are normative in character (e.g., are moral goods) (See Miller 2010, 2013). If the latter, then there is much work to be done in the provision of the derivations in question and it is unclear whether this work has been successfully done. Specifically, it is very unclear that, as Tuomela claims following Margaret Gilbert (1989, p. 162), “joint intention entails collective commitment” (Tuomela 2013, p. 83). Here there appears to be an equivocation between commitment in the sense of a moral or quasi-moral obligation (or an institutional obligation) and commitment in the sense of non-revisablility—at some point immediately prior to performing an action the intention-in-action is non-revisable. The latter is derivable from a joint intention but the former is not; yet it is the former that Tuomela requires. The other derivation that Tuomela requires pertains to collective acceptance. Once again there appears to be equivocation. Tuomela says, “participants collectively accept a content (proposition) as true or correct for the group just in case they come, in a way involving their explicit or implicit agreement, to share a we-mode joint attitude” (2013, p. 126). So collective acceptance depends on explicit or implicit agreement. Clearly explicit agreements are a form of promise making and as such presuppose institutional and moral normativity. What of implicit agreements? Implicit agreements are agreements just as implicit promises are promises. Therefore, implicit agreements are a form of promise making and the problem remains. Tuomela invokes the notion of a convention to try to extricate himself from this problem (2013, p. 146). But this is unhelpful in so far as joint intentions do not entail conventions and question begging in so far as conventions are regarded as intrinsically social phenomena. Tuomela’s third move is to invoke so-called ‘we-mode reasons’. Tuomela offers an account of collective reasoning in terms of we-mode or we-­ reasoning (2013, Chapters 4 & 7). Roughly speaking, he posits a group with an integrated membership such that the group members are committed to a common goal, say, the maximisation of group utility, and this allows him to infer on the part of each group-member-reasoner that he will do his part. The strong commitment to maximisation of group utility—at the expense of individual self-interest or otherwise self-­oriented

66 

S. Miller

goals—enables Tuomela to avoid the problems that beset individual rational self-interested actor models. According to Tuomela, a person acts for a we-mode reason if he takes as his main reason for action what the group wants, intends, or believes etc.—in general something that also requires his participation as a group member. Here the reason is collective acceptance based (see above) rather than simply intersubjective in another regarding I-mode sense. If the I-mode is here stripped of the notions of acting as a member of a social group (and related normative notions). then this distinction between we-mode and I-mode reasoning made here by Tuomela will hold up. But what if it is not? What if relational individualist accounts are available to strengthen I-mode reasons, as in fact is the case (see Sect. 2 below). Such accounts are not restricted to offering “aggregatively constructed” reasons, as suggested by Tuomela (2013, p. 24). In this connection, note that I am in agreement with the general move Tuomela makes in the direction of reasoning from common goals to individual actions (common goals to which members of social groups and organisations are strongly committed). However, this move does not entail any commitment to an irreducibly collective agent or reasoner; nor does it entail a commitment to irreducibly we-mode reasons. In relation to the issue of collective reasoners, Tuomela states: “The simple answer to this is that all reasoning is at bottom done by group members: groups qua groups do not think, reason, intend, believe or act except through their members reasoning as group members” (Tuomela 2014). However, this seems to suggest that groups do reason qua groups, albeit through their members reasoning as group members. Again, we have mystery: namely, the mystery of the collective reasoner. Or, if not, then evidently we are simply left with individual human persons reasoning and, in particular, reasoning qua members of a group.

4  Joint Actions: We-Mode and I-Mode 

67

3 Relational Individualism: A Strict Individualist Account Thus far I have, in effect, argued that Tuomela’s account of the we-mode and the I-mode is inherently unstable. His standpoint of conceptual non-­ reductionism without ontological commitments is inherently unstable. Moreover, in doing so I have offered some arguments against his non-­ reductionism. Clearly, any attempt to deal with all or most of the arguments for and against reductionism would take us well beyond the confines of this article. Rather I have more limited aims. These are, firstly, to offer an alternative conception; a strict individualist relationist account, namely the Collective End Theory (CET) elaborated elsewhere.7 Secondly, I want to apply this theory (appropriately supplemented) to two key notions that Tuomela relies on and that are apparently thought by him to pose problems for strict individualism. The notions in question are acting qua member of a group and collective reasoning (in so far as it pertains to joint actions). I note before proceeding that it is sometimes argued that individualism must hold that the attitudes involved in joint actions must not make any reference to irreducibly collective entities. However, this kind of argument confuses irreducibly collective entities with irreducibly collective attitudes. An individualist account of joint action does not need to hold the implausible view that there are no irreducibly collective entities, e.g., Australia, the USA, Finland and the European Union. Clearly there are irreducibly collective entities and such entities can and do figure in the content of individual attitudes, e.g. John Searle believes that Finland but not Australia or the USA is a member of the European Union. However, the existence of collective entities does not entail the existence of collective agents, and specifically, it does not entail the existence of supra-­ individual or we-attitudes. Let me now turn to the Collective End Theory. Basically, CET is the theory that joint actions are actions directed to the realisation of a collective end. However this notion of a collective end is a construction out of the prior notion of an individual end;  See (Miller 1992, 2001, 2007, 2013).

7

68 

S. Miller

accordingly, it is an attitude in the I-mode. Roughly speaking, a collective end is an individual end more than one agent has, and which is such that, if it is realised, it is realised by all, or most, of the actions of the agents involved; the individual action of any given agent is only part of the means by which the end is realised. So a joint action simply consists of at least two individual actions directed to the realisation of a collective end.8 Accordingly, individual actions, x and y, performed by agents A and B (respectively) in situation s, constitute a joint action if and only if: (1) A intentionally performs x in s and B intentionally performs, y, in s; (2) A x-s in s if and only if (he believes) B has y-ed, is y-ing or will y in s and B y-s in s if and only if (he believes) A x-s or is x-ing or will x in s; (3) A has end e, and A x-s in s in order to realise e and B has e, and B y-s in s in order to realise e; (4) A and B each mutually truly believes that A has performed, is performing or will perform x in s and that B has performed, is performing or will perform y in s; (5) each agent mutually truly believes that (2) and (3).

In respect of this account the following points need to be noted. First, with respect to clause (2), the conditionality of the action is internal to the agent in the sense that if the agent has performed a conditional action then the agent has performed the action in the belief that the condition obtains. Moreover, the conditionality of the agent’s action is relative to the collective end. A x-s only if B y-s relative to the collective end e. So B’s y-ing is not a necessary condition for A’s x-ing, tout court. For example, it might be that A has some other individual end e1, and that A’s x-ing realises e1 (in addition to e). If so, A will x, even if B does not y. But it remains true that A x-s only if B y-s relative to the collective end e. Again, A x-s if B y-s relative to the collective end, e. So B’s y-ing is not a sufficient condition for A’s x-ing, tout court. If for some reason A abandons the end e, then even if B y-s, A will not x.

 The following definition is drawn from various previous publications of mine. See previous note.

8

4  Joint Actions: We-Mode and I-Mode 

69

Second, the notion of a collective end is an individualist notion. The realisation of the collective end is the bringing into existence of a state of affairs. Each agent has this state of affairs as an individual end. (It is also a state of affairs aimed at under more or less the same description by each agent.) So a collective end is a species of individual end. Thus CET is to be distinguished from the accounts of theorists such as Searle who favour irreducibly collectivist notions of collective intentions. A collective end is the same as an ordinary individual end in that qua end it exists only in the heads of individual agents. But it is different from an ordinary individual end in a number of respects. For one thing, it is a shared end. By shared end I do not mean a set of individual ends that would be realised by qualitatively identical, but numerically different, multiple states of affairs. Rather I mean a set of individual ends that would be realised by one single state of affairs; so the coming into existence of the one single state of affairs in question constitutes the realisation of all of the individual ends that exist in the heads of each of the agents. For another thing, although collective ends are shared (individual) ends, they are ends that are necessarily shared by virtue of being interdependent; they are not ends that are shared only as a matter of contingent fact. Suppose you and I are soldiers being shot at by a single sniper. Suppose further that we both happen to see the sniper at the same time and both fire at the sniper in order to kill him. I have an individual end, and you have an individual end. Moreover, my individual end is one that I share with you; for the death of this one sniper will not only realise my individual end, it will simultaneously realise your individual end. However, this is not yet a collective end; for the fact that I have as an end the death of sniper is not dependent on your having as an end the death of that very same sniper, and vice-versa for you. So much for the strict individualist collective end theory. Let us now turn to the two salient notions thought to be problematic for strict individualist accounts.

70 

S. Miller

3.1 Acting Qua Member of a Group Although the boundaries are not always clear-cut, the notion of acting qua member of a group should be distinguished from that of acting qua member of a category of persons. Examples of the latter might include acting as a father or acting as drug addicts do (e.g., by stealing to sustain the drug habit). The members of a category of persons have certain features in common but they are not necessarily, or even typically, engaged in a joint action or joint enterprise. In some cases, the features that a category of persons have in common might arise in part by virtue of their interaction with one another, e.g., fatherly behaviour might arise in part from exposure to fathers. However, such interaction does not necessarily constitute joint action, at least in any straightforward sense. By contrast, the notion of acting qua member of a group can be framed in terms of joint action since the group can be defined in terms of the collective end or ends which the group of individuals is pursuing.9 Consider a group of individuals building a house. Person A is building a wall, person B the roof, person C the foundations, and so on. To say of person A that he is acting qua member of this group is just to say that his action of building the wall is an action directed toward the collective end that he and the other members of the group are seeking to realise, namely a built house. Notice that the same set of individuals could be engaged in different collective projects. Suppose persons A, B, C etc. in our above example are not only engaged in building a house but also—during their holidays— in building a sailing boat. Assume that A is building the masts, B the cabin, C the bow, and so on. To say of A that he is acting qua member of this group is just to say that his action of building the masts is an action directed toward the collective end that he and the other members of the group are seeking to realise, namely a built boat. Accordingly, one and the same person, A, is acting both as a member of the ‘house building group’ (G1) and as a member of the ‘boat building group’ (G2). Indeed, since A, B, C etc. are all and only the members of each of these two groups, the membership of G1 is identical with the membership of G2.  Some groups, such as economic classes, might be defined in part in terms of a common interest which they pursue somewhat unconsciously. Such groups may have unconscious collective ends. 9

4  Joint Actions: We-Mode and I-Mode 

71

Moreover, when A is building the wall he is acting qua member of G1, and when he is building the mast he is acting qua member of G2. But this phenomenon of one agent acting as a member of different groups in no way undermines individualism. Indeed, CET is able to illuminate this phenomenon as follows. For A to be acting qua member of G1 is for A to be pursuing—jointly with B, C etc.—the collective end of building the house; for A to be acting qua member of G2 is for A to be pursuing— jointly with B, C, etc.—the collective end of building the boat. Further, let us suppose that G1 and G2 each have to create and comply with a budget; G1 has a budget for the house, and G2 has a budget for the boat. The members of G1 and G2 know that they must buy materials for the house and the boat (respectively) and do so within the respective budgets. Assume that A, B, C etc. allocate $50,000 to pay for bricks for the house. This is a joint action. Moreover, this joint action is one that A, B, C etc. have performed qua members of G1. G1 is individuated by recourse to the collective end of building the house, and the proximate (collective) end of buying bricks is tied to that group, G1, and to its ultimate end of building a house. Accordingly, A, B, C etc. are not in buying the bricks acting qua members of G2. For G2 is individuated by the collective end of building a boat, and A, B, C etc. do not qua members of G2 have any plans to build their boat from bricks! Doubtless, many groups are unified in more complicated ways than in our above example. Consider the members of an organisation; each person occupies an organisational role. Elsewhere I have elaborated an account of the joint actions performed by such organisational role occupants (Miller 2001, Chapter 5). Specifically, I have introduced the concept of a layered structure of joint actions. Roughly speaking, a layered structure of joint actions is a set of joint actions each of which is directed to a further collective end; so it is a macro-joint action comprised of a set of constituent micro joint actions. I have also supplemented this account by recourse to concepts of conventions, social norms and the like, i.e., regularities in action to which members of organisations (and members of social groups) typically conform. At any rate, the point to be made here is that my account of the notion of acting qua member of a group in terms of acting in accordance with collective ends can be complicated and supplemented to accommodate various different kinds of acting qua

72 

S. Miller

member of a group. Nor does the irreducibility of collective entities pose problems for this analysis of acting qua member of a group or, more specifically, of a collective entity. For as argued above the irreducibility of collective entities does not entail the existence of irreducibly collective attitudes. Accordingly, the notion of acting qua member of a collective entity does not necessarily bring with it any irreducibly collective we-attitudes. It might be argued the organisational and other roles pose problems for my above-elaborated individualist account of acting qua member of a group. Here, of course, we are restricting ourselves to roles in joint enterprises, such as organisations, as opposed to roles occupied by members of a category of persons. We also need to keep in mind the distinction between a role per se and an embodied or occupied role. For instance, the organisational role of Head of Department in a university might at some time be unoccupied; the role per se exists but it has not occupant. If a group were to be defined purely in terms of roles, irrespective of whether the roles in question were occupied or not then G1 and G2 (mentioned above) would have different ‘members’ in the sense of different ‘roles’. So, at least in theory, we could speak of two different ‘groups’ in the sense of two different groups of roles. However, our concern here is not with groups of roles but groups of role occupants. That is, A, B and C above refer to the individuals who occupy the roles, i.e., individuals qua role occupants; not the roles themselves. That is, our concern is with individual persons, A, B and C who act qua members of a group in the sense of an occupied (by persons) set of organisational roles. We are not speaking of A, B and C acting as ‘members’ of a set of roles. Indeed, the notion of an individual flesh and blood person as a role, i.e., as an abstraction, is scarcely coherent. Note that it is consistent with the above that G1 and G2 are, nevertheless, different groups (in the sense of sets of occupied roles) and they are different groups even if the occupants of the roles constitutive of G1 are A, B and C, and the occupants of the roles constitutive of G2 are also A, B and C. For the members of G1 (whoever they are) have a different collective end from the members of G2 (whoever they are). Note also that although A might be the occupant of a constitutive role of G1 and simultaneously the occupant of a constitutive role of G2, nevertheless, A

4  Joint Actions: We-Mode and I-Mode 

73

performs different actions in these roles (actions directed to different collective ends) and, therefore, in acting qua member of G1, A is not acting qua member of G2 (and vice-versa). Nor does it follow from this that in our example of A, B and C simultaneously occupying the roles of G1 and of G2 that G1 and G2 have different members. G1 and G2 have different constitutive roles and the occupants of those roles perform different actions to realise different ends, but (in the example under consideration) G1 and G2 have the same members, i.e. the same role occupants, namely, A, B and C.

3.2 Collective Reasoning Tuomela’s account of collective reasoning tends to equate individualist theories of this phenomenon with rational choice theories. This is, I suggest, a mistake. The rational choice model of reasoning assumes rational self-interested individuals in competition with one another. This is fine as far as it goes. However, as Amartya Sen (2002) and others have argued, it is far from being the whole of the story. Specifically, it does not leave room for rational individual action performed in the collective self-interest and/or in accordance with socially engendered moral principles and purposes10; yet the latter are ubiquitous features of human collective life, including in the economic sphere. Importantly, this one-sided fixation with individually rational self-interested action eliminates the possibility, in effect, of finding a solution to collective action problems of the kind in question. This can be seen in the failure to find solutions to the problems such as the problem of who guards the guards. As Elinor Ostrum, the Nobel laureate in economics, quipped11 in relation to the title of Mancur Olson’s famous rational choice monograph, The Logic of Collective Action: “It should have been called, “The Logic of Collective Inaction”. Tuomela’s notion of collective reasoning as we-reasoning seeks inter alia to address the deficiencies

10 11

 See (Elster 1989).  At a presentation she gave at Delft University of Technology in June 2010 at which I was present.

74 

S. Miller

of the rational choice model but in doing so, as we saw above, proffers a somewhat mysterious non-reductive account. That said, I agree with Tuomela that what is needed at this point is the acceptance that individual human agents can, and often do, engage in action-determining reasoning from collective goals and interests to individual actions: collective goals and interests to which members of social groups and organisations are strongly committed. However, in my view (as already stated), such acceptance does not imply any commitment to an irreducibly collective agent (or reasoner). Indeed, any such commitment would simply be the application of the rational choice model at the collective rather than the individual level and, as such, would not provide a solution to the problem at hand but simply elevate it to the higher level. Nor does it necessarily imply some ultimate individually rational self-­ interested reason which motivates the possession of the collective goals and interests in question; rather collective goals and interests may well in many cases function as ground-level explanations of behaviour. On this view of individual reasoning from common goals and collective interests to their own individual action, what is called for is the above-described conception of an individual human agent qua member of a social group or organisation, e.g., qua banker or soldier. In short, the individual internalises the collective goals and interests of the social group or organisation to which he or she belongs. Crucially, these collective goals and interests can, and often do, transcend the role occupant’s prior and limited, individually rational self-interested, goals and interests; moreover, the collective goals and interests in question can, and often are, embraced by the individuals in question on the grounds that they are desirable from an impartial or, at least, collective standpoint. This capacity of individuals to reason from, and act in accordance with, collective goals and interests is not without its problems. For example, the collective goals in question might be inherently morally problematic (e.g., those of the Third Reich), or the collective interest of the members of an organisation in question might be at variance with the interests of the wider society (e.g., the members of an investment bank with a business model based on constructing and on-selling toxic financial products). In order to get an understanding of reasoning of individual agents acting qua members of a social group or organisation engaging in joint

4  Joint Actions: We-Mode and I-Mode 

75

action to realise a collective end which is also a collective good it might be helpful to remind ourselves of the basic structure of means/end reasoning. Roughly speaking, individual means/end reasoning proceeds in the following manner. First, identify the end that one has. Second, identify the means to that end. Third, perform the action which is the means to the end. Notice that here ends are not necessarily goods or benefits; they are simply ends. In the case of cooperative schemes—involving as they do, joint actions directed to collective ends—matters are more complicated. But here is a rough summary of such reasoning. First, agent A has end e, and agent B also has end e, and so does C etc. Moreover, A has end e only on condition that B has end e, and so on. Second, A and B and C etc. believe that e cannot be achieved other than by A, B, C etc. each performing some action x. Third, A, B, C etc. each has the true belief that each other has performed, is performing or will perform x, and these beliefs are a matter of common knowledge. Four, A, B, C etc. each perform x. Notice once again that here ends are not necessarily goods or benefits; they are simply ends. Now let us consider means/end reasoning in relation to ends which are also regarded as goods or benefits. First, agent A comes to have e as an end for the reason that he regards e as a good or benefit, G. Second, agent B has end e, and agent C also has end e, and so does D etc. A, B, C etc. have e as an end for the reason that they regard e as a good or benefit or as a means to some other good or benefit. Note that here A has end e only on condition that B has end e, and so on. So e is a collective end. Second, if A and/or B etc. have competing ends, e overrides the other ends, and does so by virtue of the greater importance that A, B, attach to the goods or benefits that the realisation of e will bring. Third, A and B and C etc. believe that e cannot be achieved other than by A, B, C etc. each performing some action x. Fourth, A, B, C etc. each has the true belief that each other has performed, is performing or will perform x, and these beliefs are a matter of common knowledge.

76 

S. Miller

Fifth, A, B, C etc. each perform x. I suggest that this is the basic structure of practical means/end reasoning in joint actions performed to produce what are mutually believed to be benefits or goods. Moreover, this basic structure does not change in the case of complex joint actions, such as those constitutive of organisational action, e.g., actions of officers, managers, employees in a corporation in a market, or of soldiers, generals in an armed force fighting a war. Viewed from this perspective the solution to collective action problems, such as the tragedy of the commons and the so-called Hi/Lo game (discussed by Tuomela), is joint action, typically a cooperative scheme, and frequently (but not in the case of, for instance, the Hi/Lo game) an enforceable cooperative scheme, in which there is a collective end the realisation of which removes or at least curtails the collective action problem. Moreover, this collective end is a collective good and, as such, has motivational force for the participants in the joint action. Naturally, individualist self-interest remains a problem and, as stated, enforcement is often an additional requirement. However, framing the issue in this manner enables us to see: (1) collective self-interest and/or a collective good may in fact be constitutive of individual self-interest, e.g., a banker’s individual self-interest is constituted in part by his participation in and contribution to the realisation of the collective ends of the bank which employs him; (2) collective self-interest and/or collective goods may in fact override individual self-interest; individual self-interest does not necessarily dominate collective self-interest and/or collective goods, e.g. the success of the armed force to which he belongs might be more important to a soldier than even his own life; (3) in cases where collective self-interest and/or collective goods have motivational force but it is overridden by individual self-interest enforcement is necessary, but enforcement mechanisms may not be sufficient for compliance, e.g. if enforcement mechanisms have limited reach. In relation to (3) above, I note that given the motivational role of collective self-interest and/or collective goods, enforcement does not have to be sufficient for compliance. Finally, I note that framing collective action problems in this manner allows us to offer at least a partial solution to the well-known above-­ mentioned problem of who guards the guards (or who ensures the

4  Joint Actions: We-Mode and I-Mode 

77

enforcers actually enforce, if they have an individual interest in, say, free-­ riding). Here there are three relevant points to be made. First, the individual enforcers are themselves engaged in joint action to realise a collective end which is a collective good to which they are, at least pro tanto, committed, namely, that of ensuring the compliance of participants with the requirements of the cooperative scheme(s); accordingly, qua bona fide enforcers they are motivated by collective self-interest and/ or this proximate collective good, as well as by individual self-interest. Second, the enforcement team is itself a participant in and contributor to the larger collective end of the cooperative scheme(s), be it an organisation or industry, in which they enforce compliance; so they have this further collective self-interest and/or collective good as motivating factors. Third, enforcers are in fact dependent on the cooperation of the majority of those whose compliance they enforce; indeed, enforcement is itself a joint action of sorts between the enforcers, on the one hand, and the broad mass of those whose compliance with the larger cooperative scheme is required, on the other. Consider in this connection that combating fraud is dependent on the co-workers of fraudsters reporting the latter’s fraudulent activities. Fourth, and in conclusion, this way of framing collective action problems provides solutions which do not have recourse to irreducible collectivist notions.

References Bratman, Michael. 2014. Shared Agency. New York: Oxford University Press. Elster, Jon. 1989. Rationality and Social Norms. Logic, Methodology and Philosophy of Science 8: 531–552. Gilbert, Margaret. 1989. On Social Facts. London: Routledge. Ludwig, Kirk. 2016. From Individual to Plural Agency. New  York: Oxford University Press. ———. 2017. From Plural to Institutional Agency. New  York: Oxford University Press. Miller, Seumas. 1992. Joint Action. Philosophical Papers xxi (3): 275–299. ———. 2001. Social Action: A Teleological Account. Cambridge: Cambridge University Press.

78 

S. Miller

———. 2007. Joint Action: The Individual Strikes Back. In Intentional Acts and Institutional Facts: Essays on John Searle's Social Ontology, ed. S.L. Tsohatzidis, 73–92. Dordrecht: Springer Press. ———. 2010. The Moral Foundations of Social Institutions. Cambridge University Press. ———. 2013. Joint Actions, Social Institutions and Collective Goods: A Teleological Account. In Institutions, Emotions and Group Agents: Contribution to Social Ontology, ed. Anita Konzelmann-Ziv and Hans Bernhard Schmid, 99–115. Springer. ———. 2014. Critical Review of Raimo Tuomela’s Social Ontology: Collective Intentionality and Group Agents. Notre Dame Philosophical Reviews. https:// ndpr.nd.edu/reviews/social-­ontology-­collective-­intentionality-­and-­group-­ agents. Accessed 28 Aug 2022. Schmitt, Michael. 2003. Joint Action: From Individualism to Supra-­ Individualism. In Socializing Metaphysics: The Nature of Social Reality, ed. Michael Schmitt, 1–37. Maryland: Rowman and Littlefield. Searle, John. 1995. The Construction of Social Reality. New York: Penguin. Sen, Amartya. 2002. Rationality and Freedom. Harvard University Press. Tuomela, Raimo. 2013. Social Ontology: Collective Intentionality and Group Agents. New York: Oxford University Press. ———. 2014. Comments on Seumas Miller’s Review of Social Ontology: Collective Intentionality and Group agents in the Notre Dame Philosophical Reviews. April 20. Centre for Philosophy of Social Science. https://tint.helsinki.fi/materials/tuomelareply.pdf. Accessed 28 Aug 2022. Tuomela, Raimo, and Kaarlo Miller. 1998. We-Intentions. Philosophical Studies 53 (3): 367–389.

5 Towards a Situated Approach of Tuomela’s Theory of Social Practices Judith H. Martens

1 Introduction Social practices are a key concept in Raimo Tuomela’s work on sociality, they help us understand many aspects of sociality, including customs, traditions, and institutions. The key elements in his analysis of social practices are we-attitudes and pattern-governed behaviors. I am sympathetic to Tuomela’s approach to sociality to the extent that it recognizes and spells out many sufficient and necessary conditions for different types of social activity that together make up sociality. I agree with him that the complexity and multiplicity of human agency, both when acting individually as well as when acting collectively or cooperatively, indeed calls for such an analysis with multiple sufficiency conditions. I especially appreciate his acknowledgement of the role of habits, customs, and the like for sociality. His three most recent books, The Philosophy of Social J. H. Martens (*) Centre for Philosophical Psychology, Department of Philosophy, University of Antwerp, Antwerpen, Belgium e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_5

79

80 

J. H. Martens

Practices (2002), The Philosophy of Sociality: The Shared Point of View (2007), and Social Ontology; Collective Intentionality and Group Agents (2013) all pay attention to this aspect of human being. In this paper, I will zoom in on this aspect of Tuomela’s work and look at how he integrates habits, customs, and skills. In Tuomela’s words, I will zoom in on pattern-governed behaviors and social practices, focusing on Chap. 3 and 4 from The Philosophy of Social Practices and Chap. 8 from Social Ontology. What I want to focus on is the notion of habit that Tuomela uses throughout his work. On the one hand he gives it an important role and mentions customs often, as the “collective counterpart of habits” (Tuomela 2002, p.  121). At the same time, the analysis of habits and customs left me wanting. I will argue that they are too strongly dependent on and are integrated in a context in which we-attitudes are the key explanatory element. That is to say that also in his account of customs Tuomela’s we-mode account emphasizes social and shared reasons for which a group act as key elements of sociality. This entails that collectively accepted contents must be taken to be for group use, namely, collectively available and in force for the group members, and, when broadly conceived, for the benefit of the group’s goals and interests. This means that the weight of his account rests on acceptance of group’s ethos by the relevant individuals. In his own words: We-mode thinking, feeling, and acting presuppose collective acceptance of the group’s ethos (or of some element entailed by the ethos) or of some other, nonconstitutive content as the object of the group’s “attitudes.” (Tuomela 2007, p. 4)

This idea relies on the interactive knowledge condition: It is not enough for the individuals to have an intention to perform the actions that could make up a collective action. Rather, they have to possess an intention to do their part as their part of the collective action. Therefore, according to Tuomela, interactive knowledge requires communication and/or a decision procedure. To this purpose, I will start a dialogue between Tuomela’s work on social practices and pattern-governed behaviors on the one hand and my

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

81

developing framework of habits and skills in joint action on the other hand. I will first elaborate on the elements within Tuomela’s framework that are essential to understand the role he gives to routinized, or pattern-­ governed, behaviors, before I continue by pointing at some elements from the current discussion on emergent coordination and meshed cognition that supplement, but also problematize, aspects of his theory. From there I will look at the consequences of such a more situated and bottom-up approach. To get to these points, the set-up of the paper is as follows: The 2nd section will discuss the starting point we typically find in philosophy of collective intentionality and social ontology accounts, and what consequences it can have. In the 3rd section I will first discuss Tuomela’s ideas on social practices and pattern-governed behavior and how it integrates into his larger theory. The 4th section will discuss emergent coordination and agentic control through different cognitive capacities and within a framework of meshed cognition. This will give me some ways to point at alternative ways of integrating habits and skills within Tuomela’s account of sociality. In the 5th section I will then look at some unresolved issues that I have put on the table by arguing in favor of an account of meshed cognition and skills. Furthermore, I will comment on some of the metaphysical consequences of my proposal as well as the consequences for our ideas on autonomy and agentic control.

2 The Importance of Our Starting Point There is a tendency in philosophy to take reasoning as a key element of understanding human beings. Understanding human uniqueness in terms of reasoning can be understood in two ways. On the one hand we can understand such an ability to be central or present in most activities humans are involved in. Let us call this centrality-HD (high degree). On the other hand an ability can be understood to be central to humans as it explains some uniquely human capacity. This I will label centrality­U (uniqueness). Having a unique capacity does not mean we are constantly using this capacity. In other words, we should not make the mistake to go from a centrality-U claim to a centrality-HD claim.

82 

J. H. Martens

Although Tuomela starts with the promising mention of the plurality of ways in which sociality comes about, as well as a keen eye to the habituality of humans, I think his starting point of the analysis of sociality in the end does push him towards making this mistake. To quote Dan Sperber and Hugo Mercier: The elephant trunk is a type of nose. However impressive it may be, it would not make sense to think of it as the epitome of noses. Similarly, reason is one type of inference mechanism; it is neither the best nor the model of all others. (Sperber and Mercier 2017, p. 49)

They then continue with an analogy that I found striking and which I also think applies to the collective intentionality debate: The classical contrast between intuition and reasoning isn’t better justified than the old hackneyed contrast between animals and humans beings (and its invocation of reason as something humans possess and beasts don’t). To contrast humans not with other animals but simply with animals is to deprive oneself of a fundamental resource to understand what it is to be human and how indeed humans stand out among other animals. Similarly, to contrast reason with intuitive inference in general rather than with other forms of intuitive inference is to deprive oneself of the means to understand how and why humans reason. (Sperber and Mercier 2017, p. 90)

Tuomela starts his analysis of collective intentionality with thoughts that can be expressed as propositions (see Tuomela 2002 around p. 125). This is problematic because this makes these thoughts that can be expressed as propositions the epitome of collective intentionality and sociality. I will argue that other forms of sociality should not be modeled after this form of cognition. When we do that we are making the mistake of moving from a centrality-U claim to a centrality-HD claim and take on board all kinds of assumptions on the working of our cognitive capacities that underlie our sociality. This claim may initially remain hidden, as habits and skills are mentioned as important categories of human activity. But if we do not flesh out an account of sociality based on such forms of cognition there is a hidden assumption that either habitual activity should

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

83

be analyzed in analogy with such propositional attitudes, which leaves us with a centrality-HD claim, or the account—in the end—does not take habitual activity as central as initially seems to be suggested. I would like to note that this critique is similar, but not identical, to Sara Chant’s critique of the wash-rinse-repeat method of theories in collective intentionality more generally (Chant 2017). Chant points at the way philosophers have taken accounts of individual agency and “copy-­ pasted” it onto accounts of collective intentionality. We are both pointing at a wash-rinse-repeat process, but in Chant’s case this holds between individual and collective intentionality. What I am pointing at is a similar mistake, but within the domain of collective intentionality. On the one hand Tuomela’s work is full of recognition of the habitual, of the rituals, customs, traditions, and so on that help understand a lot of what our sociality consists of in. Yet, reasons and propositional attitudes remain the starting point, also when considering pattern-governed, routine, behavior. Pattern-governed behaviors are only meaningful in their larger context and by means of a we-attitude and these attitudes are researched by means of linguistically expressible attitudes: The Collective Acceptance approach will be formulated by speaking mainly of the acceptance of collective ideas and thoughts, assumed to be linguistically expressible so that we can speak of the acceptance of sentences. (Tuomela 2002, p. 124)

Tuomela suggests that in the cases of collective ideas and thoughts there is an attitude that can be expressed. In other words, in these cases there is a propositional attitude that is collectively accepted. But what about those cases where there is nothing said? Is the mental machinery supposed to function in a similar vein? Take the following example where Tuomela is talking about the acceptance of money as an entity that gets its collective-social status by being collectively created. He writes: Here the acceptance in the case of single individuals amounts either to accepting out loud (e.g. simply by saying “I accept that such and such is the case”) or internal mental acceptance that would have been accepting out loud had it been overt. (Tuomela 2002, p. 125)

84 

J. H. Martens

How this could translate to cases where a we-attitude is not of the form of a propositional attitude remains unclear. What does acceptance amount to when we are not thinking of propositions? What is a mental acceptance of, for example, know-how? Tuomela does not argue that skilled activity always involves propositional attitudes. But he does take propositional attitudes to be the starting point when reflecting on the content (and mode) of attitudes. His starting point, on the level of analysis, however, is also the end point. There are no other attitudes or mental states considered. What would such attitudes look like? How would they compare to propositional attitudes? One option to think about within Tuomela’s framework, is his use of Sellars’ notion of pattern-governed behavior. Tuomela makes use of this idea throughout his work (see e.g., 2002, 2007, 2013). However, at the same time he is committed to the idea that actions should be understood by means of the standard theory of action: According to our common-sense view, a person can be regarded as an intentional agent, and an intentional agent is conceptually to be understood in terms of our ordinary framework of agency. It contains the following central elements: Intentional agents can have representational mental states such as beliefs, wants, and intentions, and they can also have emotions and feelings with their bodily accompaniments. (Tuomela 2013, p. 21)

On this view, agency and acting for a reason are closely connected. When discussing another matter, Tuomela writes that he is “[f ]ocusing on intentional action, thus action performed for a reason” (Tuomela 2013, p.  10). This is a general tendency that we find throughout his work. On the one hand there is the acknowledgement that human agency is often also habitual, and this habitual analysis should be different. Yet this is in tension with his key analysis in terms of intentional actions and intentional states. In what follows I will first discuss his notion of pattern-governed behavior, with a focus on the work in The Philosophy of Social Practices as it discusses Sellars. The second main source I will focus on is Chap. 8 from Social Ontology, as it gives an “account of social institutions as

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

85

systems of positions and normatively governed, routine-like social practices (collective “pattern-governed behaviors”)” (Tuomela 2013, p. 20).

3 Social Practices: Pattern-Governed Behaviors and We-Attitudes Social practices, on Tuomela’s account, are one type of collective social actions. The core idea is that social practices lead to token activities that are performed for social reasons, or we-attitudes. Social practices are repeated collective social actions. Such collective pattern-governed behaviors are not reducible to aggregates of individual actions nor to aggregates of individual pattern-governed behaviors. Another important aspect of social practices, how they are grounded, are the pattern-governed behaviors of individuals and groups. Throughout his work, Tuomela writes about the “core sense” of social practices. He takes this to consist of recurrent collective social actions that are performed for a shared social reason or, in other words, shared we-attitudes. These we-attitudes can be either intentions or intentions-in-action (Tuomela 2002, p. 42).

3.1 Pattern-Governed Behavior Action, within Tuomela’s framework, comes in two forms: intentional and non-intentional. Intentional action is action performed on purpose. “Such action is performed in accordance with and (partly) because of the intention to perform that very action or a “closely related” action” (Tuomela 2002, p. 40). Nonintentional action is contrasted with both intentional action on the one hand and reflex behavior as well as mistaken actions and action slips on the other (2002, p. 41). Nonintentional action is “meaningful action in the sense of being an understandable action with a teleological point” (ibid.). Both in The Philosophy of Social Practices and Social Ontology this is illustrated by the example of scratching one’s nose while in a somewhat stressful situation. Tuomela argues that nonintentional actions are meaningful in two ways:

86 

J. H. Martens

Thus, in this control sense it is meaningful, and it is meaningful in the teleological sense just because the type of action of scratching one’s nose is a meaningful action. (Tuomela 2002, p. 41)

The meaning of these actions is derived from a larger pattern of action as an element of that larger action or activity. The pattern-governed behavior is part of a larger action, it is performed because of the whole behaving system’s achieving a goal. This means that the larger action is “not nonintentional” (Tuomela 2013, p. 216). In this case, Tuomela suggests that scratching one’s nose might be meaningful as it expresses the agent’s nervousness. Within this analysis there are no pattern-governed activities that cannot also be analyzed in terms of a goal of the organism, even if the activity does not strictly speak for, or contribute to, that goal.1 Such nonintentional actions (which are components of larger actions) are understood in terms of Sellersian-inspired pattern-governed behaviors. For this Tuomela (2002, p. 46) cites Wilfrid Sellars: Roughly it is the concept of behavior which exhibits a pattern, not because it is brought about by the intention that it exhibit a 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 has been selectively extinguished. (Sellars 1973(b), p.  489  in Tuomela 2002, p. 46)

While in Sellars’ work this type of strictly nonintentional behavior is always understood as activities by single individuals, Tuomela extends this idea to groups as well. The pattern-governed behavior of the single agents (the causal motors of collective agency on Tuomela’s account) serves to ground intentional action in the public realm (Tuomela 2013, p. 216). Having given this exposition of pattern-governed behaviors, it is now time to turn to the other central element in Tuomela’s account of social practices: we-attitudes.  Throughout Tuomela’s work there seems to be a tension between, on the one hand, the reading that I propose here and, on the other hand, the suggestion that there are other ways of integrating routine behavior. On those occasions that he mentions them they are put aside as not conceptually or philosophically interesting. 1

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

87

3.2 We-Attitudes and Social Practices Tuomela’s approach to social practices presupposes the notions of we-­ attitude and collective social action. A collective social action, on this account, is an action performed for a social reason, where this social reason is to be understood as a shared we-attitude. Collective pattern-­ governed behaviors are causally and conceptually dependent on a group context. Collective and individual pattern-governed behaviors are, on Tuomela’s account, understood to have the same grounding function; they are central building blocks of sociality. Tuomela writes: Accordingly, collective pgbs [pattern-governed behaviors] are central building blocks of the social world, especially in the sense of forming the “routine” ingredients in social practices (viz., repeated collective social actions), customs, and institutional behavior. […] This can result in collective activities and practices that are, so to speak, routine throughout. (Tuomela 2002, p. 58)

Tuomela considers customs to be on the collective level what habits are on the individual level. On both occasions the larger activity gives meaning to the pattern-governed behaviors. The we-attitudes that are grounded in pattern-governed behaviors can be both in the I-mode and in the we-­ mode. These attitudes remain present, even when action is strongly routinized. There are weaker forms of social practices in Tuomela’s work. These weaker notions still all rely on a belief of what is being collectively done or of a shared goal, but this social awareness need not be mutually known. Notice that “weaker” here does not refer to this being “less of an action”, but is rather about the strength of the sharing. It is debatable how to best interpret Tuomela’s claims about the centrality of social practices. He writes about social practices performed for a we-attitude as social practices in the core sense or core social practices. Customs and traditions, according to Tuomela, are to be analyzed and understood in terms of this combination of shared we-attitudes and repetition of the behaviors that go with them until pattern-governed behaviors are part of the explanation.

88 

J. H. Martens

Several agents can perform actions that are, in some way, spatial-­ temporally connected and in that sense could be judged to be collective actions. This, however, is insufficient to count as a social action on Tuomela’s account. A collective action is only a social collective action when the agents are socially connected in the right way. The agents have to act (at least in part) for the same social reason. When an agent has such a reason they are understood to have a suitable attitude (analyzed in terms of a belief or intention) towards the particular reason (the content) and to be acting for this reason entails realizing the attitude towards this reason. In order for a reason to be a person’s motivational, action-guiding reason and for him to act for that reason it must obviously enter and be processed by him in his intentional cognitive system. (Tuomela 2002, p. 82)

Besides the awareness that the agent needs to have of their reason they must also believe that others have that same reason and act similarly for that reason and believe that this is mutually believed. In what follows I focus on one of many examples that Tuomela introduces in demonstrating the explanatory power of his framework. This particular example sits uneasily with me because of the analysis of mimicry that is given by Tuomela. The example focuses on teenage girls dressing similarly (all wearing miniskirts in Tuomela’s example). According to Tuomela this is a case of social practice, and the we-attitude is “to want to wear similar clothing” (2002, p. 83 for a detailed discussion). To me this seems a rather quick assumption about the underlying reason why these teenage girls dress similarly. How can we say with certainty that this is the underlying reason? Especially in terms of these girls being mutually aware of that being the reason why they are wearing miniskirts. One alternative would seem to be that these girls want to be fashionable and they like what is in fashion. On this scenario they might actually dislike the fact that so many other girls are wearing the same fashionable item. On a slightly different scenario there might be some more primitive in-group/ out-group motivation playing a role to motivate these girls to dress similarly, a motivation that might not come up as a motivating reason to these girls at all.

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

89

These alternative interpretations of the reasons why these girls are all following the same fashion bring me to another point that has me confused in Tuomela’s analysis. He analyses social practices in terms of we-­ attitudes that are shared and there is mutual awareness of these shared we-attitudes. Such an analysis seems to leave less room for an understanding of these girls in terms of intentionality (more broadly construed), and instead seems to direct us towards an understanding in terms of we-­ intentions and other we-attitudes. Again, as discussed in Sect. 2, on the importance of our starting point, Tuomela leaves room for other interpretations but it remains unclear what these other interpretations would look like and why they aren’t applied to this particular example.

3.3 Routine and We-Attitudes As discussed earlier, pattern-governed behaviors, both individual and collective ones, are nonintentional, except when considered in their context (see Sect. 3.1). Yet, Tuomela argues that they are collectively intentional in the aboutness sense. This aboutness sense is contrasted with the conduct sense. Pattern-governed behavior is nonintentional in the conduct sense, which means that it cannot be performed on purpose, by forming an intention to perform it. The action, however, can still occur in a meaningful context of activities and normally does occur embedded in such a way. From this embeddedness a pattern-governed behavior can be (collectively) intentional in the aboutness sense. The pattern-governed behaviors are meaningful in the sense of exhibiting intentional aboutness in a teleologically appropriate sense (Tuomela 2002, p.  57). For collective pattern-governed behaviors, this aboutness sense can be understood as the ought-to-be norms governing a pattern-governed behavior. An ought-­ to-­be norm, in the context of pattern-governed behaviors, can be taken to represent a group goal. Ought-to-be norms are, like pattern-governed behaviors, a concept from the Sellersian framework. Sellars makes a distinction between two kinds of rules, or oughts: ought-to-do norms and ought-to-be norms (deVries 2021). Ought-to-do norms are like conditional imperatives: when in circumstance C do A. Ought-to-be norms are norms of criticism, they are an endorsement of

90 

J. H. Martens

a particular state of affairs but do not spell out how this state of affairs is to be achieved. Ought-to-be norms lead to sanctioning and/or disapproval when they are not followed and will entail ought-to-do norms, which will make it the case that actions are in line with the ought-to­be norm. Let us now return to the collective pattern-governed behaviors. An example in Tuomela’s work are children learning to cross the street when the light is green (2013, p. 217). The norm in this example is that the people in that community obey traffic rules including that you only cross the street when the light is green. The children in this example do not need to have the right representation or goal, what matters is that their behavior is correct. The intentional aboutness of the collective pattern-­ governed behavior need not be on the level of the individual, but there are group goals involved (the ought-to-be norm). These ought-to-be norms can be created or designed by a group or institution. However, they can also be the result of a natural or social selection process. In such cases the connection to we-attitudes is not clear to me. Yet, Tuomela does want to discuss them in terms of social practices, which implies such we-­ attitudes are involved. This means that after a child learns the correct behavior, they are assumed to—at some point—also learn the connected we-attitude that gives the meaning for a pattern-governed behavior as it is integrated in a specific context. Otherwise, they cannot become the teachers of the future generation. Yet, this is contradicted in his writing: As to the initiation problem, there can be collective pgb as a type of activity which requires both relevant individual and collective intention to be initiated, although the intention is “forgotten” if the activity gets fully routinized. Alternatively, there can be collective pgb that requires only relevant individual intention to get initiated. (Tuomela 2002, p. 62)

What kind of forgetting could be in play here? The most fitting possibility would seem to think that the intentions are still playing a role, but they are nonconscious. They have become implicit. This seems to fit best

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

91

to the general picture we get from Tuomela’s account.2 The problem I have with this reading is that implicit is a word with many potential meanings and ways of understanding what it would amount to. The process by which pattern-governed behaviors automatize, become habitual, and what that implies for the way in which we think they still deserve to be understood as actions, is something that plays at a level of analysis that we do not find in Tuomela’s work. It has, however, consequences for the way in which we can understand the ontological claims that Tuomela makes. I will say more about the understanding of routine and implicit cognition in the next section but let me first say a few things about why this matters.

3.4 Mind-Dependency and Realism of Sociality One of the key questions in social ontology is about the ontological status of the social realm. Tuomela holds a realist position, but simultaneously argues that group agency is partly “fictitious”. In an elucidating paper, Arto Laitinen (2017) discusses the various ways in which we can understand the mind-dependency of institutional reality. The purpose of his analysis is to understand Tuomela’s claim, which several others in the essay collection are struggling with, that in some sense collective intentionality is fictitious rather than factual or real (social reality mind-­ dependent, ontologically subjective and only “epistemically” objective). Starting from the intuitive distinction between mind and world he sketches several alternatives of understanding the relation. I will not discuss all the possible options but instead focus on the key distinctions that we will need to understand Tuomela’s view. Fictitious views are often understood as an irrealist view. An irrealist view in social ontology would hold that games, rules, organizations and tournaments are not real.

 Another way to read this is to think that the skillful agent has learned in a conscious way, but— like the ladder to climb the wall—no longer needs, and therefore forgets, any explicit knowledge about the now routinized activity (this would be in the spirit of Dreyfus and Dreyfus (1986). This, however seems problematic, because the we-attitudes get lost in this picture. 2

92 

J. H. Martens

They are “imagined”, “fictitious” rather than factual, unreal rather than real, subjective rather than objective, mental rather than physical, “made up” rather than independently existing (the other positions will argue these distinctions do not all go together). Given the strong hold that this distinction has in the philosophical imagination, the use of phrases like “ontologically subjective” or “fictitious” concerning collectives, games and organizations may suggest this irrealist position. (Laitinen 2017, p. 149)

Laitinen contrasts this view with perspectivalism. On this view the physical, material, causal world is real, but so are pains and games. Games and pains are real despite being ontologically subjective or mental in some sense. A perspectivalist has three conceptions of reality to avail themselves to: 1. Objective things: events and facts about the physical and material reality exist and are real in themselves or non-perspectivally. 2. Individually subjective or mental things: events and facts exist and are real “for the individual” or within the individual’s perspective. 3. Collectively intentional things: events and facts exist and are real “for the group” or within the group’s perspective (Laitinen 2017, p. 150). Both irrealists and perspectivalists locate the institutional reality within minds and not in the world. A realist, on the other hand, will argue that games and institutions are objectively and publicly in the world and are real in another way as understood by the perspectivalist. Another way to describe this is to say that some objective things are intentionality-­ dependent. The contrast with perspectivalists is to think that social and institutional reality is located in the world, rather than in the mind. Social reality may be mind-dependent but, for a realist, that doesn’t imply that it is mental or subjective. The realist view can be further specified by looking at the kind of mind-dependency that is assumed. Below are two of the three positions that will be relevant for our current discussion: the first view holds that collective intentionality is needed both for bringing about and maintaining the institutional reality, the second holds that it is needed only for bringing about. (Laitinen 2017, p. 151)

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

93

Laitinen argues that Tuomela seems to subscribe to a combination of perspectivalism and the first form of realism. Tuomela, in his reply to Laitinen, argues that his view is a combination of perspectivalism and the second form of realism (Tuomela 2017). This surprised me, as I would have analyzed Tuomela’s view as Laitinen does. So let’s first look a bit into the reasons why Tuomela is committed to a certain view and then look at the reasons I have to follow Laitinen’s interpretation. As will become clear, I will point at further reasons to accept Laitinen’s view, which will be argued for based on my analysis of the way Tuomela treats habits and skills in his theory. The main difference between the two realist views comes down to whether they think that some form of collective intentionality is needed to sustain the social reality, or whether they think that collective intentionality is only needed at the genesis of social reality. The second realist denies the need for continual acceptance or continual existence of the collective intentionality that was needed to get, i.e., an institution off the ground. This would still leave room for pattern-governed behavior to play a role. On this view, if an agent continues to accept the obligations past events have created, the role of their psychological attitudes is responsive to, not constitutive of, the already created obligations. (Laitinen 2017, p. 152)

How can we combine this second realist view with the idea that we-­ attitudes are essential for social practices? Understanding the we-attitudes that we have just analyzed in terms of a response to already created obligations seems to be at odds with the prominence they have in the analysis of social practices in Tuomela’s work. Especially as these we-attitudes function as motivating reasons on Tuomela’s account. What kind of because, or causal, relation is there between the social reason (the shared we-attitude) and the repeated collective social action? Tuomela identifies two relations: (a) a reason-giving causal relation and (b) each particular token of a social practice gets its (teleological) meaningfulness from the social reason for which it was performed (see 2002, Chap. 4). So, on the one hand I want to embrace Tuomela’s claim that the second realist picture also fits into his framework. It will allow us to think

94 

J. H. Martens

more about situated cognition. At the same time, I do feel that this has consequences for the way we can still conceive of the actions of the agents involved. To say more about this, I will now turn to the next section, which will discuss some current ideas in the cognitive sciences on habitual and skillful agency, allowing us to fill in some blanks and see what the consequences for the theory might be. What kind of connection is needed between these initial commitments, intentions, or we-beliefs and acting on social practices? Are these token instantiations of social practices still constitutive of action and agency? After all, Tuomela himself writes that “involuntary items, including social imitation, but also adhering to norms and social practices, do not seem to be analyzable by means of collective intentionality” (Tuomela 2017, p. 169).

4 Emergent Coordination and Agentic Control Bill Pollard (2006) has drawn attention both to the importance of habit explanations and to the fact that they do not fit the Davidsonian-Humean conception of action explanation. Having a habit of doing something is the reason why someone does something (the corresponding action), but it is typically not the reason for doing that thing. This is also the case for those habits that have been actively developed by the agent (for certain reasons). For as far as we understand habits as pattern-governed behaviors this fits Tuomela’s account. The trouble of solely understanding habits in that way is that on Tuomela’s account their meaning can now only be understood by means of the larger activity that they are embedded in. Tuomela recognizes the cognitive limitations of human cognition, yet somehow remains working with this folk-psychological notion of intentions, beliefs, and desires and takes that as his sole method of analysis. He writes: Individualism is a doctrine that privileges the individual over the group and emphasizes individual autonomy and self-determination. It is in part based on the general and correct view that in normal circumstances

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

95

i­ndividual human agents are in charge of what they do: an individual’s behavior is normally to be seen as her action (normality here excludes, e.g., reflexes and behavior fully controlled by an external agent). (Tuomela 2013, p. 9)

And later in the book: Intentions serve to initiate, control, and guide action. Furthermore, the full intentionality of an action requires that the agent intends the action (or a closely related action or state). Analogously to individual intentional action, full-blown collective action requires the participants’ relevant collective intention directed to a collective goal. (Tuomela 2013, p. 62)

The habitual activity can only be understood as meaningful as part of another action. This seems, to me, to be a step in the wrong direction. Overall, the theory as proposed by Tuomela functions very top-down. An intention, be it shared or individual, is what explains the social practices. This explanation is taken to be both causal and motivational (Tuomela 2002, p. 42). Pattern-governed behaviors and motor control then fill in the details. There is, however, reason to think of these processes and influences to not only be top-down. In what follows I sketch two routes to think about intelligent collective behavior differently. Mind-dependent, but not propositional attitude dependent, if you like.

4.1 Emergent and Spontaneous Action Our cognition is shaped in such a way that we can successfully and “easily” interact with our surroundings, be they other agents or objects that we are surrounded with. As many, including Tuomela, have pointed out, we are creatures with only limited cognitive capacities. Tuomela offers we-mode reasoning as less cognitively demanding in many social situations as compared to I-mode reasoning, but it still builds on decision-­ making and propositional attitudes. Many of the artifacts that we have constructed over time will also allow us to easily interact with them. One way to analyze this idea is in terms of action possibilities that objects and situations can offer us. Even if such devices were once built in an

96 

J. H. Martens

intentional way, there need be no reason that our interaction with them now relies on the same (we-)attitudes. Research on emergent coordination studies the coordination between agents while acting on the fly. The focus is on joint action that occurs via perception-action couplings or via lower-level cognitive mechanisms. An example is the tendency to fall into synchrony with others when we are performing the same action. This is, amongst others, happening when people rock in rocking chairs (Richardson et al. 2007). Even when the subjects are asked to remain rocking in their own rhythm, they will be inclined to change their tempo to that of the other. Such mechanisms involved in joint action are argued to function without shared intentions in place (Vesper et  al. 2010; Knoblich et  al. 2011; Tollefsen and Dale 2012). Many, if not all, of these types of influence of the presence of the other are described as automatic and hard (if not impossible) to control, when taking intentional control as the measure. They can, however, set the stage for more controlled joint agency. Either because such forms of emergent coordination led to the formation of we-attitudes, or because other forms of agentive control such as habitual and skillful activity, follows from (or develops out of ) such emergent processes. Note that in these cases the lower-level processes are influences and co-determining the higher-level processes. This is the reverse from the picture we find in Tuomela. Through several“rounds” of acting together, (a sense of ) we-agency and (a sense of ) being in control together, can become stronger, especially as one becomes better at predicting what will happen. Through multiple loops of interactions between the agents, an agent can become more certain over time that they are not mistaken, that they are sincerely acting together. To what extent this control and this acting together is dependent on we-attitudes (understood as propositional attitudes) is highly questionable. Again, Tuomela does not argue that we can only understand we-attitudes in terms of propositional attitudes, but the analysis of we-attitudes that he gives are in terms of content that can be (and is) spelled out in propositional form. This starting point of his analysis leaves the possibility of other forms open, yet as long as there is no advancement as to how these attitudes can be analyzed, either we have to assume an

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

97

analogy or we are in clear lack of a theory of how such attitudes can be shared and have the same explanatory role. Situatedness, both in terms of objects and in terms of others surrounding the agent, not only strongly influences decision-making, but it can also shift attention to novel actions and action possibilities. Sometimes also the goals or the type of activity we are involved in can develop as we go, depending on the action possibilities that the environment offers us. See for example Beth Preston’s (2012) critique of Michael Bratman’s account of planning agency as it doesn’t leave enough space for spontaneity. Borrowing (and extending) Preston’s example, we can imagine someone who drives past a farmers’ market and spontaneously buys olives. Later on during the day they run into friends and based on this purchase end up organizing a festive meal for a group of friends. The we-attitude was certainly not top-down guiding what the agent was doing, it only developed over the course of the day. The niche that we humans have constructed for ourselves is full of designs or artefacts that, intentionally or not, guide us to act in certain ways. Our cognition is dependent on the situation that we find ourselves in: the design of the environments to get people to behave in certain ways (social affordances and cognitive offloading would be the terminology we find in the situated cognition debates). In Shaun Gallagher’s words: “The expression of several minds is externalized and extended into the world” (Gallagher 2020, p.  213). Where the design might be intentional, the responses by other agents in the designed environment need not be so. And the original intention can also get lost while the institution still nudges us to act in a certain way. The key idea is that our thinking is guided by these cognitive institutions. This, however, is not the only claim. We can also assume that the cognitive processes get transformed and extended through these institutions (Slors 2019; Gallagher 2020). The social systems we create can shape our cognitive processes, while also depending on these processes. Another example can be found in the work on implicit shared intentions (Martens and Roelofs 2018).

98 

J. H. Martens

4.2 Habits, Meshed Cognition, and Agentic Control In the contemporary literature, scholars focus on the idea that automation need not oppose the idea of the agent being in control (Bargh 1994; Christensen et al. 2016; Mylopoulos and Pacherie 2019; Fridland 2021; Hutto and Robertson 2021). In other words, scholars are debating if and how automaticity and agentic control can go hand in hand. This has created an opening towards understanding habits and skills as providing opportunities for the agent to be in control. In philosophy of action there is a tradition to carve up the world actions and mere bodily movement. Within the category of action there is typically a further distinction made between full-blown intentional actions and activities (or doings). Habitual and skillful activities are good candidates for this class of activities and give us room to conceptualize being in control while also automatizing and routinizing our activities. This novel development to propose such theories has not been fully integrated with many of the traditional theories on agency and conceptualizations of, for example, autonomy and control are still mainly depending on the notion of full-blown intentional agency. Although most philosophers of action acknowledge that actions do not always issue from deliberation, a positive characterization of habitual and skillful control, and hence of habitual and skillful agency, is often lacking. Traditional approaches agree that agents do not always pause to consider their reasons for acting, nor need to have any conscious thoughts about how to proceed. But such negative characterizations merely mark a contrast between purposive agency (doings) and “full-blooded” agency. It remains underdeveloped what the characteristics are of acts that are described in this negative manner. The relevance of meshed cognition is especially visible against the background of the notion of control. Control, most particularly self-­ control, is generally recognized as an important, or even key, aspect of agency and autonomy. The notion also plays an important role in cognitive sciences, albeit that this discipline is mainly interested in control rather than self-control (Martens 2020). Recent discussions about control are partially coined in terms of meshed cognition. The key idea of

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

99

meshed cognition is that cognitive and automatic processes characteristically operate together in an intimately ensnared arrangement. In the case of meshed control, there is a distinction between three types of intricately connected (meshed) forms of control (Christensen et  al. 2016). (1) Cognitive control is focused on strategic task features and mostly future-­ oriented, (2) situated control is crucial for control of the action in relation to the immediate situation, and (3) implementation control is responsible for implementation, making sure the movement is executed in a way that is fitting to the current surroundings and previous, current, and future bodily positions and movement. Not all three levels of control must be constantly involved while the agent is engaged in certain activities, and the model allows room for other ways of conceptualizing goals and situationally adequate reactions as by means of propositional attitudes. As with the discussion on emergent coordination and situated and spontaneous action, we can now arrive at a similar conclusion: many of the activities we are engaged in may include propositional attitudes. These attitudes, however, are not a necessary element for agentic control, and they also need not be the starting point of the whole activity.

5 Conclusion In this paper I have looked at the role of pattern-governed behaviors in Tuomela’s work and supplemented and critiqued it based on the work on routine behaviors and the situation of the agent. While Tuomela’s work heavily relies on pattern-governed behaviors and in that sense is open to the idea of more routinized behaviors to play an important role for our sociality, I also discussed how Tuomela relies on we-attitudes in his analysis of social practices and only analyses these in terms of propositional attitudes. The meaning of pattern-governed behavior can, on his account, only be understood by looking at these we-attitudes. Pattern-governed behaviors must be understood as embedded in a situation in which the we-attitude gives us the meaning and intentionality. As I discussed in the previous section, this analysis seems too top-down. This was also the reason that I was surprised at Tuomela’s realist position, as discussed in Sect.

100 

J. H. Martens

3.4 on mind-dependency. At the same time, I think that the second realist position, the take that depends on a historical mind-dependency and not on continual mind-dependency, does open ways to integrate routine behaviors more strongly into the model. I do think, however, that this implies that we have to let go of the role of we-attitudes in many of these situations. In Sect. 4, I discussed some examples and provided some tools to think about alternatives. With these tools in hand we can now return to the starting point for theorizing that I discussed in Sect. 2. Does Tuomela make the centrality­HD claim? I think that, given the way he integrates pattern-governed behaviors into his account, he does. The other alternative is to think that many aspects of sociality cannot be analyzed by means of his theory. The core analyses that Tuomela gives, for example, in the collectivity condition, are in terms of what group members intend and believe and thus in terms of content (see Tuomela 2013, Chap. 3). Within this context Schmitz (2017, p. 48) has also pointed at the strong hold that the received view of propositional attitudes still has on philosophy more general, including Tuomela’s account. Schmitz rightly criticizes that the traditional view is (too strongly) “inspired by reports of propositional attitudes”. For Schmitz this has the consequence that the mode and subject fall out of the picture. For me this has the consequence that we take any “translation” to propositional reports for granted. When we take group reasons to be an important, essential even, aspect of the we-mode and sociality as Tuomela analyses it, the question is what role other forms of agentic control can play. We easily make the mistake to not only see reasons and propositional attitudes as uniquely human, but also as central to human behavior more generally: we shift from centrality-U to centrality-HD. When Tuomela talks about agency he typically connects this to the notion of control. This idea of control, then again, is linked to a reason-­ giving and causal relation between intentions and the actions that follow from it. What I have looked at in the last part of this chapter, is the possibility to think about agentic control but without necessarily depending on intentions. Intentions need not always be involved, and when they are involved they need not be the starting point. We need to move away from this top-down picture in which intentions and decision-making

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

101

procedures are always the starting point. Of course they can be, but what I hope this chapter has shown is that they need not be. Depicting sociality in this top-down manner is convenient because the rationality remains present, while also accommodating for lower-level processes to play their part. I think, however, that this is not true to life. These lower-level mechanisms have their own internal workings that, albeit not on a rational level, will inform the agent about action possibilities. These mechanisms (a) are smart in their own way and (b) work bottom-­up. We-attitudes will sometimes also be the consequence rather than the starting point of an interaction. Let me end on a small note on the relationship between this notion of shared agency and autonomy. Throughout his work Tuomela multiple times talks about the autonomy we lose when acting in and for a group. While I agree that this can indeed be true, I would like to emphasize two further points. First, I don’t think that agents have as much autonomy to begin with. Secondly, I think it is also important to emphasize the other side of the coin: groups, social practices, and institutions not only take away autonomy and control, but they also give us means for novel forms of action and autonomy we could otherwise not dream of. While keeping in mind that there is a double effect possible and also taking into account that not everyone is afforded the same actionables based on the social practices and institutions that we build, I think that the balance of these sides of the same coin is something that needs even more attention than it currently gets (see the many wonderful contributions in Mackenzie and Stoljar 2000). The general line of thought can then be phrased as follows: we are only autonomous beings because of and in relation to others and our interactions with others. This autonomy is relational, a matter of degree, and dependent on several forms of control. Acknowledgements  The author would like to thank Miguel Garcia and Rachael Mellin for their comments on an earlier version of the paper.

102 

J. H. Martens

References Bargh, John A. 1994. The Four Horsemen of Automaticity: Awareness, Intention, Efficiency, and Control in Social Cognition. In Handbook of Social Cognition, ed. R. Wyer and T. Srull. Lawrence Erlbaum. Chant, Sara Rachel. 2017. Collective Action and Agency. In The Routledge Handbook of Collective Intentionality. Routledge. Christensen, Wayne, John Sutton, and Doris J.F. McIlwain. 2016. Cognition in Skilled Action: Meshed Control and the Varieties of Skill Experience. Mind & Language 31: 37–66. https://doi.org/10.1111/mila.12094. deVries, Willem. 2021. Wilfrid Sellars. In The Stanford Encyclopedia of Philosophy, ed. Edward N.  Zalta, Fall 2021. Metaphysics Research Lab, Stanford University. Dreyfus, Hubert L., and S.E. Dreyfus. 1986. Mind over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. New  York: Free Press. Fridland, Ellen. 2021. Skill and Strategic Control. Synthese. https://doi. org/10.1007/s11229-­021-­03053-­3. Gallagher, Shaun. 2020. Action and Interaction. 1st ed. Oxford: University Press. Hutto, Daniel D., and Ian Robertson. 2021. Clarifying the Character of Habits: Understanding What and How They Explain. In Habits: Pragmatist Approaches from Cognitive Science, Neuroscience, and Social Theory, ed. Fausto Caruana and Italo Testa, 204–222. Cambridge: Cambridge University Press. Knoblich, Günther, Stephen Andrew Butterfill, and Natalie Sebanz. 2011. Psychological Research on Joint Action: Theory and Data. Psychology of Learning and motivation 54: 59–101. https://doi.org/10.1016/B978-­0-­12-­ 385527-­5.00003-­6 Laitinen, Arto. 2017. We-Mode Collective Intentionality and Its Place in Social Reality. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, Studies in the Philosophy of Sociality, ed. Gerhard Preyer and Georg Peter, 147–167. Cham: Springer International Publishing. https://doi.org/10.1007/978-­3-­319-­33236-­9_11. Mackenzie, Catriona, and Natalie Stoljar, eds. 2000. Relational Autonomy: Feminist Perspectives on Autonomy, Agency, and the Social Self. Oxford, New York: Oxford University Press. Martens, Judith. 2020. Doing Things Together: A Theory of Skillful Joint Action. De Gruyter.

5  Towards a Situated Approach of Tuomela’s Theory of Social… 

103

Martens, Judith, and Luke Roelofs. 2018. Implicit Coordination: Acting Quasi-­ Jointly on Implicit Shared Intentions. Journal of Social Ontology 4: 93–120. https://doi.org/10.1515/jso-­2018-­0018. Mylopoulos, Myrto, and Elisabeth Pacherie. 2019. Intentions: The Dynamic Hierarchical Model Revisited. Wiley Interdisciplinary Reviews. Cognitive Science 10: e1481. https://doi.org/10.1002/wcs.1481. Pollard, Bill. 2006. Explaining Actions with Habits. American Philosophical Quarterly 43: 57–69. Preston, Beth. 2012. A Philosophy of Material Culture: Action, Function, and Mind. Routledge. Richardson, Michael J., Kerry L.  Marsh, Robert W.  Isenhower, Justin R.L. Goodman, and R.C. Schmidt. 2007. Rocking Together: Dynamics of Intentional and Unintentional Interpersonal Coordination. Human Movement Science 26: 867–891. https://doi.org/10.1016/j.humov.2007.07.002. Schmitz, Michael. 2017. What Is a Mode Account of Collective Intentionality? In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, Studies in the Philosophy of Sociality, ed. Gerhard Preyer and Georg Peter, 37–70. Cham: Springer International Publishing. https://doi.org/10.1007/978-­3-­319-­33236-­9_3. Slors, Marc. 2019. Symbiotic Cognition as an Alternative for Socially Extended Cognition. Philosophical Psychology 32. Routledge: 1179–1203. https://doi. org/10.1080/09515089.2019.1679591. Sperber, Dan, and Hugo Mercier. 2017. The Enigma of Reason. Harvard University Press. Tollefsen, Deborah, and Rick Dale. 2012. Naturalizing Joint Action: A Process-­ Based Approach. Philosophical Psychology 25: 385–407. Tuomela, Raimo. 2002. The Philosophy of Social Practices: A Collective Acceptance View. Cambridge: Cambridge University Press. https://doi.org/10.1017/ CBO9780511487446. ———. 2007. The Philosophy of Sociality: The Shared Point of View. https://doi. org/10.1093/acprof:oso/9780195313390.001.0001. ———. 2013. Social Ontology: Collective Intentionality and Group Agents. New York: Oxford University Press. ———. 2017. Raimo Tuomela: Response to Arto Laitinen. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, Studies in the Philosophy of Sociality, ed. Gerhard Preyer

104 

J. H. Martens

and Georg Peter, 169–178. Cham: Springer International Publishing. https:// doi.org/10.1007/978-­3-­319-­33236-­9_12. Vesper, Cordula, Stephen Butterfill, Günther Knoblich, and Natalie Sebanz. 2010. A Minimal Architecture for Joint Action. Neural Networks 23. Social Cognition: From Babies to Robots: 998–1003. https://doi.org/10.1016/j. neunet.2010.06.002.

6 What Is Collective Acceptance and What Does It Do? Arto Laitinen

1 Introduction This article discusses Raimo Tuomela’s (2002) Collective Acceptance approach to social and institutional reality. By way of introduction, consider the following claims: 1. A business company, as a type of organization, is an institution. 2. The existence and nature of institutions is wholly explained by collective acceptance (of constitutive norms of institutions). 3. All of social and institutional reality is constructed by humans, and the CA-theory captures such social construction. 4. (Collective) acceptance is a separate kind of mental state of its own, alongside (collective) belief, intention, desire; perhaps with a double direction of fit.

A. Laitinen (*) Philosophy; Faculty of Social Sciences, Tampere University, Tampere, Finland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_6

105

106 

A. Laitinen

5. An institution or a social entity exists, if it is collectively accepted as existing; unlike trees and stars that exist independently. (Performativity; acceptance as an achievement -notion.) 6. An institution is collectively accepted as existing if it exists (Reflexivity) Perhaps surprisingly, Tuomela (2002) explicitly denies the first four claims, in ways to be discussed below. Further, instead of discussing existence-claims like 5 and 6, Tuomela in the main text applies collective acceptance mainly to something seemingly very different: to sentences being “collective-social” and correctly assertible in a group (having “forgroupness”). This is, at the surface level at least, a thesis about sentences having some technical features, not about institutions existing: (CAT) A sentence s is collective-social in a primary constructivist sense in a group g if and only if it is true for group g that (a) the members of g collectively accept s, and that (b) they collectively accept s if and only if s is correctly assertable (or true) (collective acceptance thesis). (2002, p. 132)

That something is correctly assertible in a group need not have ontological consequences and may be downright false: in Tuomela’s example, in a society of flat-earthers, it is correctly assertible and “true for the group” that the earth is flat (p. 133). One may wonder, then, whether Tuomela’s Collective Acceptance View is a view about the existence of institutions at all; and whether people interested in the social ontology of institutions, including business corporations, should read something else than Tuomela? The pessimist conclusion is too hasty. This paper argues first of all that Tuomela’s Collective Acceptance View does indeed include claims 5 and 6. Although Tuomela (2002) discusses these ontologically committal versions only in an appendix, he certainly holds them  (see also Hindriks 2015). And further, it may be a good strategy to approach the ontologically committal theses like 5 and 6 in a roundabout way via (CAT) about sentences. The roundabout way in which Tuomela approaches the question by emphasizing that institutions are always for some group or another, and starting from acceptance of sentences for a group enables

6  What Is Collective Acceptance and What Does It Do? 

107

him to build an impressive and distinctive “weakly communitarian” theory of institutions, where especially the we-mode attitudes have a central role (For a briefer account of the main ideas, see Tuomela 2003). Thus, further central claims for Tuomela include: 7. The key to social and institutional reality is the way in which organizations, other institutions, social practices, collective goals and beliefs, are always for a group 8. Collective attitudes had in the we-mode (entailing collective commitment and the collectivity condition) are central in how things are for a group. This paper does not focus on these two further claims, however, but on the nature of collective acceptance. What is it and what does it do? Section 2 is introductory, explaining why Tuomela does not endorse claims 1–3 above. To anticipate, the problem with (3) is just the phrase “all of ”. Social and institutional reality does have important features constructed in collective acceptance, but some features are generated by unintentional consequences, perhaps by invisible hand mechanisms, and they may be theorizable for example as game theoretical equilibria. Further, the problem with (2) is the word “wholly”: social institutions consist not only of collectively accepted social norms but also of social practices, activities and patterned collective behaviours. What about (1)? While role-organizations from hospitals to post offices are one of the three main types of institutions (the first two being general conceptual institutions like language and calendars, and then society-wide institutions with deontic powers, like marriage or property), it follows from Tuomela’s we-mode analysis that an organization must be for the benefit of all, the whole group, to count as a collective-social institution in Tuomela’s strict sense. By contrast, “[a] business company… may not be an institution even for its employees (and still less for the larger community within which it operates)” (Tuomela 2002, p. 160). I would phrase this point so that an organization is of course an institution, but it need not involve we-mode collective acceptance. In any case, Tuomela’s classifications of types of institutions are very helpful for assessing the aims of CA-theory (see Sect. 2).

108 

A. Laitinen

Sections 3 and 4 of this paper then identify puzzles about collective acceptance and try outlining answers to them, without arguing in detail for the suggested answers. The claim (4) goes to the heart of the collective acceptance view: what is acceptance in the relevant sense? If squirrel fur is money if and only if it is collectively accepted as money, what kind of thing is the “acceptance” in question, that is reflexively and performatively related to squirrel fur being money? Is it a special kind of attitude alongside beliefs and intentions? Now it is pretty clear that Tuomela regards acceptance as involving and leading to either beliefs or intentions, so it is not a new type of attitude side by side with them with a nature of its own. It rather somehow consists in them or is realizable in them. But then again, Tuomela (2002, p. 127) does write about the distinction between (voluntary) acceptance-­ beliefs and involuntary beliefs that arise from a belief-forming mechanism, e.g. from perceptions and other experiences (see Cohen 1992; Gilbert 2002; Hakli 2006). The “acceptance” in CA may or may not be such (below I will argue that Tuomela’s states of collective acceptance are indeed best seen as acceptance-beliefs insofar as they are attitudes at all; defending this will amount to revising some other claims made by Tuomela). Tuomela characterizes the relevant sort of acceptance in many ways, which do not all seem compatible. For example, he writes that “[a]n acceptance state is a disposition to act intentionally” (2002, p. 127) but also that “there is no conceptual room … for agreeing to accept s and not accepting s” (2002, p. 1, 30). But if accepting s is a disposition, surely there is conceptual room for agreeing to accept s (say, agreeing that everyone at workplace stops telling jokes about mathematicians) and failing to form the related disposition. It turns out that Tuomela distinguishes between the event of coming to accept and the state of accepting, and more can be said about that distinction to strengthen Tuomela’s position. The idea that acceptance is an achievement notion like knowledge, creates a further tension, as does the question of whether acceptance is a mere attitude, or includes a disposition to act, perhaps a commitment to act, or perhaps action? And while Tuomela elsewhere (e.g. 2007, pp. 185–6) toys with the idea that the relevant attitudes might have a “double direction of fit”, it is better to put aside that idea altogether, as

6  What Is Collective Acceptance and What Does It Do? 

109

arguably no representation can have both directions of fit with the same content (see Laitinen 2014). It is admittedly somewhat puzzling that sometimes acceptance has the direction of fit of beliefs, and sometimes the direction of fit of intentions, but it is not conceptually impossible, like the idea of having both would be. These questions about the nature of “acceptance” in Tuomela’s collective acceptance view  will be discussed in Sects.  3 (“Puzzles”) and  4 (“Answering”).

2 Which Aspects of Social Reality Are Socially Constructed? Tuomela puts forward his CA-view as an attempt to capture socially constructed aspects of the social and institutional reality. He contrasts such constructed aspects, dependent on collective intentionality, with ones that arise from invisible hand mechanisms. (a) Collective intentionality is required for the creation and maintenance of at least some social structures, most importantly, social institutions. (Especially, the maintenance of social institutions will be argued … to require strong collective intentionality.) (b) Nevertheless, there are structural features—that especially economic theory has taught us about …— which are only unintended consequences of individual actions. These latter two claims are factual claims about our social world as we now have it. We can add to them the parallel and equally obvious factual theses that (c) collective intentionality does not generate all of society (social structures) and that (d) invisible hand mechanisms do not as a matter of fact generate all of society. (2002, p. 122)

Tuomela states that in principle all of social reality and collective order could result in only one of those mechanisms: [I]t is conceptually possible to construct (“design”) at least social institutions—if not all social structures—on the basis of collective decision making and thus collective intentionality, and, as we all know, many of our

110 

A. Laitinen

institutions have in fact been so created. On the other hand, it is also a conceptual possibility that they are created and maintained by some kind of invisible hand process that does not involve collective intentionality. Thus, collective order can in principle (as a conceptual possibility) come about either by collective intentionality or by unintended consequences of I-mode intentional and nonintentional action. (2002, p. 122)

Examples of phenomena where collective acceptance plays a direct role include: (a) Tom and Jane love each other. (b) Tom and John are friends. (c) Jim is our leader. (d) It is our collective goal to get our lake cleaned up. (e) Squirrel furs are money for us. (f ) Marriage obligates the spouses to take care of their children. (g) The policeman shows his badge to give impetus to his order. (h) The Deutsche Bank is a bank. (2002, p. 135)

Structurally cumulative features such as wealth, power or unemployment typically depend on something else being collectively accepted, but are not themselves targets of collective acceptance, and can be unintended consequences of directly collectively accepted phenomena (2002, p. 136). But there are also many phenomena which are fully independent of collective acceptance; Tuomela mentions social influence, social emotions such as envy or shared fear, sociological findings about social strata or social cohesiveness. Tuomela discusses these as test cases for his Collective Acceptance Thesis. He is satisfied that it captures the phenomena which indeed are dependent on collective acceptance, but not others  (2002, p. 136). Similarly, when discussing social institutions, Tuomela stresses that there are two equally important aspects: collective acceptance of norms that are constitutive of the institutional entities, properties and activities, and behaviour in accordance with those norms—social practices. I take an institution at bottom to be a system of social practices governed by a system of norms. The institution is one for the group, viz., for the benefit and use of the group in question. (2002, p. 171)

6  What Is Collective Acceptance and What Does It Do? 

111

Tuomela’s account of social institutions (in the “standard sense”, by which Tuomela means that it covers the three important cases of conceptual institutions such as language or calendar system, institutions with deontic powers such as money or property, and organizations consisting of roles with tasks and rights; but all which go further than mere behavioural regularities) is as follows: (SI) Sentence s expresses a social institution (in the “standard” sense) for collective g if and only if (1) s expresses or entails the existence of a social practice (or a system of interconnected social practices) and a norm or a system of interconnected norms (including a constitutive one) for g, such that the social practice generally is performed at least in part because of the norm, (2) the members of g rationally collectively accept s for g with collective commitment; here it is assumed that collective acceptance for the group entails and is entailed by the correct assertability of s. (2002, p. 170)

We will below discuss Tuomela’s approach via “sentence s” and some possible misunderstandings it may create. The basic idea is that the system of norms (e.g. rules of a game) can be expressed linguistically in a possibly very long sentence s, and those norms are in force when the sentence is accepted by the relevant community. The social institution does not consist in those norms or rules alone, but either in the social practice guided by those norms or rules, or in the pair of the practice and the rule. Thus, a social institution is not an abstract entity (only) but consists at bottom (also) of certain kind of action. “Institutions are special kinds of ordered pairs of social practices and norms (or systems of norms)“ (Tuomela 2002, p. 156).1 And as clause (2) of SI suggests, the

 Or also: “In colloquial talk a social institution can be of various ontological kinds. As examples from literature and common parlance witness, there are at least the following possibilities for what ontological kind of entity a social institution prima facie is: (a) social practice—for example the old practice of sauna bathing on Saturdays in Finland; (b) object—money (as notes, coins, etc.); (c) property of an individual—for example being an owner; (d) linguistic entity—for example natural language; (e) interpersonal state—for example marriage; (f ) social organization—for example the national postal system, a university” (2002, p. 161). He continues: “I will accept that social institutions can be conceptualized in these various ways, but I emphasize that they are all related to norm-­ governed social practices and signify some elements in them. Indeed, one can say even that a social 1

112 

A. Laitinen

members are collectively committed to the behaviour, which is relevant for the continuing existence of the institution. As a terminological clarification, Tuomela mostly writes that what is collectively accepted (for example the social norms) does not include the social practices, but corresponding social practices (repeatedly performed collective social actions) are essential in satisfying and maintaining the attitudes in which the acceptance consists: Social institutions conceptually depend on collective acceptance, viz., on the group members’ holding a relevant we-attitude, and on the social practices satisfying and maintaining those we-attitudes. (2002, p.  7; italics added) Central aspects of sociality (and, as a consequence, of social reality), including social norms and social institutions, are created and maintained by collective acceptance and the social practices that the maintenance of collective acceptance requires. (2002, p. 7; italics added)

This suggests that mere attitudes (something inner, mental, subjective as it were) in the absence of the practices (something lifeworldly, located in the shared, public, social reality, for which Tuomela coins the term “groupjective”) do not suffice. It is not true that, say, squirrel fur is money if and only if the group members hold relevant attitudes towards squirrel fur being money. The attitudes do not suffice for effective collective acceptance—the attitudes must be satisfied and maintained by social practices. Should the members have the attitudes in the absence of the social practices, their (fallible) attitudes would be faulty, mistaken, unsatisfied. And as Tuomela defines collective acceptance as an achievement or success term, in those cases there would be no collective acceptance. We will look in more detail at this in Sect. 3. Before that, let us add that Tuomela is highly critical of the game-­ theoretical view that “a social institution is an equilibrium point in a repeated game” (2002, p. 157). He discusses Schotter’s (1981) analysis and identifies no less than ten problems in it: for example, it misses the institution ultimately amounts to a special kind of norm-governed social practice (or set of practices)” (2002, p. 161).

6  What Is Collective Acceptance and What Does It Do? 

113

central role of social rules or norms and the related normative requirements for members, and it misses authority-based rules, and often there may not be an equilibrium at play (2002, pp. 255–6). It might be more in line with Tuomela’s analysis to suggest that such game-theoretical considerations or the real-world patterns of incentives may help explain why some institutions persist, but they are not an analysis of what social institutions are. (On the other hand, Tuomela is also critical of Searle’s analysis of institutions, which misses out too much of social conventions.) Tuomela distinguishes three kinds of “standard” institutions that the CA-view aims to make sense of: (1) “Language can be regarded as the most fundamental, underlying social institution” (2002, p. 159). “People must rely on the correct use of concepts in all their other institutional and conceptual activities. Concepts can be regarded as rules to be mastered by people as skills” (ibid.). (2) “Money, marriage and property are examples of institutional social objects and social institutions that typically are general, viz., society-­ wide” (2002, p. 160). They are also examples of institutions where rights and obligations of the members are relevant. (3) “Specific organizations involve specific positions and “task-right systems” (ibid.). The reason why business companies may not be institutions in Tuomela’s intended sense is not that they would not come with rights, obligations or specific positions, or that they would not be socially constructed. It is rather that they would need not be “for the use of the group, and to further the group’s basic purposes” (2002, p. 133) in the relevant sense of being accepted by the group, and work in benefit for the group, to satisfy Tuomela’s Collective Acceptance View. By contrast, as quoted above, ”[a] business company […] may not be an institution even for its employees (and still less for the larger community within which it operates)” (2002, p. 160). I would phrase this point so that an organization is of course an institution, and belongs to the constructed aspects of social reality, but need not involve we-mode collective acceptance.

114 

A. Laitinen

3 Puzzles about Collective Acceptance So far, so good. Let us now turn to the puzzles about collective acceptance. One puzzle about collective acceptance is related to collective acceptance being an achievement- or success-notion like knowledge. Take the case that squirrel fur is money if and only if it is collectively accepted as money. There can presumably be cases where squirrel fur is agreed to be money, and when asked, people could report that this is indeed so and would have no objections to it being so, but nonetheless would not act accordingly, or would not have dispositions to act accordingly, or would not be subjectively committed to act accordingly—for example out of lack of trust. It is conceivable that squirrel fur is in principle accepted as money, but nonetheless people are only motivated and disposed to engage in direct barter. I think Tuomela would say that in such a scenario, squirrel fur would not be money, because the practical, behavioural element is missing in people’s actions. If so, we would have to conclude that squirrel fur was not genuinely collectively accepted as money in the scenario. It is not collective acceptance if it does not deliver. We can save the biconditional by stating that collective acceptance (of squirrel fur as money) is whatever is needed for squirrel fur to be money. That “something” will presumably have to include actions in one way or another, or if acceptance in the relevant sense will turn out to be an attitude of some sort, it will need to be realized, satisfied, effective attitude (e.g. an intention on which one has acted).2 So, the first puzzle is: is CA some kind attitude, or “anything” that works? Another puzzle concerning “acceptance” concerns the kind of propositional attitude that it is or involves—what sort of entities does the “acceptance box” include, to use Tuomela’s metaphor? Tuomela (2002, pp. 137–38) is clear that the acceptance can lead to adding a new item to the “subbox” of beliefs, or to the “subbox” of intentions. Acceptance is one thing, and the resulting attitudinal state is another thing—sometimes an intention and sometimes a belief. Analogously, sometimes one buys apples and sometimes oranges, and the “buying” need not be an  As in Tuomela (2002, p. 152): “Genuine holding of an attitude obviously also requires appropriately acting on the attitude in question.” 2

6  What Is Collective Acceptance and What Does It Do? 

115

apple or an orange. But then, Tuomela also writes that acceptance (as a continuing state) is either the state of belief, or the state of intending. Tuomela further discusses “acceptance as true” as a kind of belief (acceptance-­belief ), but seems to relate that not to the (collective) acceptance (i.e. what leads to new items being in a box, analogous to buying), but to the beliefs that enter the “subbox” (i.e. the attitude formed, analogous to apples bought). In a later work Tuomela (2007, pp. 185–6) toys with the idea of double direction of fit, but Tuomela (2002) is clear that acceptance is either intention-like or belief-like.3 A third distinction that Tuomela makes is between the event of accepting and the state of acceptance. When connected to the previous two puzzles, a solution starts to emerge: arguably it is the event of accepting that is performatively related to the institution coming to existence, whereas the continuing state of collective acceptance is at any time required for the institution remaining in existence, and finally the event of collective acceptance being withdrawn (say, because the minimal threshold of sufficient collective acceptance, or sufficient practical realization, is no longer met) has a constitutive connection to something no longer being the case. One can further propose a clarification on two kinds of CA-events: on the one hand, intentional, agreement-like events, and on the other hand, possible datable “events” of the amount of acceptance states coming to meet the criterion of sufficiency at some point in time, so that the correlated institutional effect takes place. A fourth, and main, source of puzzlement in Tuomela’s approach is that the collective acceptance thesis is officially about “a sentence s” being “collective-social in a primary constructivist sense” and something being “true for a group”; to repeat:  See e.g. Tuomela (2002, p. 151), where CA is linked to we-attitudes with one or another direction of fit, and the claim is made that the CA is constituted by coming to hold and holding the we-­ attitude: “collective acceptance always results in and involves that the members of the collective come to hold either a conative we-attitude (one with the world-to-mind direction of fit) or a doxastic we-attitude (one with the mind-to-world direction of fit or the world-to-mind direction of fit—the latter when the belief is a constitutions institutional one) and in general also maintain, and act (or at least are disposed to act) on the basis of, the we-attitude in question. As argued, in the context of the formula (CAT) the we-attitude must be in the we-mode (at least for a substantial number of group members). We can then say that collective acceptance in a sense is constituted by the corresponding we-attitude (or, more precisely, by coming to hold and holding it)” (italics added). 3

116 

A. Laitinen

(CAT) A sentence s is collective-social in a primary constructivist sense in a group g if and only if it is true for group g that (a) the members of g collectively accept s, and that (b) they collectively accept s if and only if s is correctly assertable (or true) (collective acceptance thesis). (2002, p. 132)

As noted above, this may create the impression that collective acceptance is not anything remarkable, or anything so remarkable as to be able to lead to ontological changes in social reality. It seems merely a matter of some sentences gaining a certain status or standing in some group. Well, one could say, no doubt a sentence acquires the status of being “accepted (as true) in the group” if it is accepted (as true) in the group—how could it be otherwise? Further, when Tuomela (2002, pp.  9, 102, 128–134; 139; 142; 207), illustrates what he has in mind in something being true or correctly assertible for a group, he imagines a Flat Earth Society, that collectively accepts the sentence s = “The earth is flat”, and the members are thereby obligated to state that view (whether objectively true or not), when asked about the shape of the earth. That would suggest that what is at stake is merely the standing of sentences in a group. Squirrel fur need not be money or the Earth need not be flat, even though the sentences stating so enjoy some status in the group. If that is so, how would Tuomela’s CAT be of help in shedding light on whether and how squirrel fur is money? That is the fourth puzzle, especially when combined with the following: In turns out that in the appendix, Tuomela nonetheless subscribes to the relevant ontological claims: If squirrel fur is money then it must be (collectively) accepted in g to be money”. (Tuomela 2002, p. 197)

And Squirrel fur is money if it is (collectively) accepted to be money. (Tuomela 2002, p. 198)

Thus, Tuomela does think that e.g. history books written by outsiders can make true claims such as that squirrel fur was money in Medieval

6  What Is Collective Acceptance and What Does It Do? 

117

Finland—there are facts (facts-for-anyone as it were) about how things have been in different group contexts. Such facts are not group-­ independent, so their grounds are not “objective” in the same sense as facts about stars and trees, but they are not subjective either, and Tuomela coins the term “groupjective” (close to what normally is called intersubjective validity) for the relevant kind of objectivity.4 Given this, the puzzle is: what is the connection between collective acceptance of sentences (as collective-social for the group) and of real institutions or entities (as existing, or having some social property)? Is it so that the latter is analysed in terms of the former, or is it in the other way around? In the next section we will try to answer these questions.

4 Answering the Puzzles about Collective Acceptance Let us now try to tackle these puzzles. The aim is to preserve Tuomela’s analysis and its ontological ambitions with a reconstruction that can answer the puzzles; this will require rejecting some of Tuomela’s individual claims. I will tackle the puzzles in reverse order. Tuomela (2002, p. 11) introduces collective acceptance by noting the difference between collective behaviour that is not conditional on what other think or do, for example everyone opening their umbrellas when it starts to rain, and collective-social behaviour proper, where people’s actions are socially dependent on others, for example on social norms that have been accepted in the group. Conversely, all action for social reasons is not collective action—individuals can act socially in some sense of the word on their own. Tuomela focuses on collective-social action, which differs from collective but not social action (such as everyone opening their  See Schweikard and Laitinen (2023) for discussion on four definitions of “realism” in social ontology. There is ample conceptual room for realism in social ontology concerning mind-dependent (and thus not strictly speaking “objective”) properties and entities; Tuomela seems here more guarded than he needs to be. Below I will suggest that the fact that squirrel fur was money in medieval Finland (FM) makes the sentence “squirrel fur is money in group FM” true for anyone, it is not merely true-for-the-group. Yet, the fact is grounded in something group-dependent, whereas the shape of the earth is not. As facts they are equally objective, but the ground of one involves facts about a group, whereas the other does not. 4

118 

A. Laitinen

umbrellas) and minimally social but not collective action (such as an individual doing something for the sake of another individual). Tuomela approaches collective-social behaviour via cases where some sentence s is accepted as true for the group.5 Different groups have a different realms of concern—library committees are interested in issues related to libraries, and football teams are interested in football, stamp collectors’ clubs are interested in stamps and so on (2002, pp.  37–9). Different groups can have the very same members, if they have different realms of concern. Two groups sharing the same general realm of concerns, say, two book-reading clubs, may nonetheless accept different beliefs and intentions; there may be different contents in their “intentional horizon” that consists in the adopted attitudes (2002, pp.  82, 124, 137–9). The mechanism by which the collective accepts beliefs and intentions is collective acceptance. Tuomela analyses this as acceptance of sentences as true, for the group. Thus, the first context of application of collective acceptance concerns groups which gradually build up their intentional horizon, accept new entries to their “subbox” of beliefs and their “subbox” of intentions (Tuomela 2002, pp. 137–38). In more complex cases, the accepted sentence may express a system of constitutive norms, which the collective accepts for itself, thereby creating an institution (2002, ch. 6). Thus, what is directly the object of acceptance is a sentence, which expresses a rule, while the rule together with the members’ actions of following the rule, constitutes the institution. Accepting a sentence is accepting it as true for the group. Thus, even though we can speak about collective acceptance being directed at entities such as squirrel furs, or at institutions such as property, or at persons in such roles as a leader, these acceptances can be analysed via acceptance of sentences as true for the group. That is at least how I understand Tuomela’s (2002) account, for better or worse. We can say a norm is accepted in a group, or in force in a group, and this amounts to the sentence expressing the norm being accepted as true for the group.  Cf. also Tuomela and Balzer (1998, p. 176): “We formulate our approach by speaking mainly of the acceptance of collective ideas and thoughts, assumed to be linguistically expressible so that we can speak of the acceptance of sentences (with certain meanings or uses).” 5

6  What Is Collective Acceptance and What Does It Do? 

119

Thus, the answer to the fourth puzzle is that acceptance of sentences is the core phenomenon, whereas acceptances of institutions as existing and norms as valid and persons as leaders are to be analysed in those terms.6 How then do any interesting ontological consequences follow, if acceptances concern only sentences? One option would be to say that in the case of specifically collective-­ social phenomena, a sentence is accepted as true for the group if the sentence is true, and the sentence is true if it is accepted as true for the group. Squirrel fur being money seems to be that kind of case: squirrel fur really is money only if it is accepted as money. By contrast, the earth being flat or not is not that kind of case. The shape of earth is not dependent on collective acceptance. Thus, the suggestion might go, concerning collective-­social phenomena such as acceptance of squirrel fur as money, acceptance of sentences has important ontological consequences, but accepting sentences about phenomena that are not collective-social but natural (e.g. earth not being flat) does not have ontological consequences.7 In many ways, this might be a good solution to the puzzle. However, this is not the way Tuomela uses “collective-social”. For him, collective acceptance of the sentence “the earth is flat” is collective-social in the relevant sense; it differs from collective non-social action such as everyone opening their umbrellas for non-social reasons. (There can of course be a group whose members independently of each other come to hold that the earth is flat, but in Tuomela’s example they collectively accept the sentence as assertible and “premisible” in the group). As a result, there may be obligations for members to assert that the earth is flat, when asked; there is collective commitment to holding the sentence true. Thus, the difference between accepting “squirrel fur is money” and “the earth is flat” cannot be is that one is collective-social while the other is not: the collective acceptance of either sentence leads the sentence acquiring the status of “collective-social” for the group. So, we need a slightly different way to capture the difference.  There may be benefits to sticking to propositional attitudes, rather than attitudes towards entities (such as recognition of persons, or acceptance of institutions or recognition of norms), although for other reasons these are important attitudes to analyze. See Ikäheimo and Laitinen (2007). 7  I thank Raul Hakli for formulating this suggestion. 6

120 

A. Laitinen

Another option (the one I would personally favour) would be to say that there are facts about group-dependent entities such as money and facts about natural phenomena such as the shape of earth. As facts they are equally objective, and they serve to make anyone’s (including outsiders’ and future historians’) sentences or thoughts true: “squirrel fur is money for the group FM“ is not merely true for the group, but true-for-­ anyone; or true, period. The difference is not in their being facts, but in what grounds the facts. Institutional facts depend on collective acceptance of relevant sentences (and thus what is “groupjective” is the ground, whereas the resulting fact—if it is a fact—is equally objective as any fact), but natural facts do not depend on collective acceptance (and thus their grounds are ontologically objective in the sense of being mind-­ independent, and not in Tuomela’s terms “groupjective”). My understanding is that this parallel between facts about institutions and facts about natural events would be too strong for Tuomela, who would have favoured some kind of “internal realism” or retained some perspectivalism about truth-for-groups; but it is one way of linking  the collective acceptance of sentences and ontological consequences.8 This takes us to the question of events of acceptance and states of acceptance, and the third puzzle. It seems overwhelmingly plausible to link CA-events to ontological changes in the institutional world (coming to accept and squirrel fur becoming money; ceasing to accept and squirrel fur ceasing to be money) and CA-states to maintaining the ontological states in the institutional world. That is, events go with events and states with states. Approached in this way, something comes to existence or acquires institutional properties (say, squirrel furs acquire the property of being money) because the suitable CA-event takes place; and it is the CA-event that is a sufficient condition for squirrel furs becoming money. The  See the exchange between Laitinen (2017) and Tuomela (2017) about realism and perspectivalism. The view outlined above can be called realism about (facts about) institutions. Tuomela wants to retain a stronger form of group-perspectivalism: “my ontological view is a combination of the second form of realism and group-perspectivalism”. A number of Tuomela’s other publications, such as (2007) and (2013), characterize his way of understanding the ontological difference between group-phenomena and natural phenomena. For him, reality is tied to causality, whereas others would allow for facts that are epiphenomenal or only normatively, but not causally, relevant. For a systematic discussion of realisms in social ontology, see Schweikard and Laitinen (2023). 8

6  What Is Collective Acceptance and What Does It Do? 

121

suitable CA-event can in principle be of two sorts: first, it can be an intentional or explicit agreement, where the sentence is accepted as true for the group. There is no great puzzle on how agreements can have ontological consequences (they would be rather pointless, if agreements would not bring such things as contracts into existence). But secondly, it can in principle be that a sufficient number or quota of members comes to have the suitable member-level states, at some determinate point in time. At that very point, the ontological consequence takes place and squirrel furs acquire the property of being money. No visible change in the universe need to take place, as long as the sufficient conditions for squirrel fur being money have come to be met. This can be because of psychological changes in the members, say thanks to prior communication which has now led, perhaps after time-consuming deliberation by some, to the formation of suitable attitudes and behavioural dispositions in sufficiently many operative members of the group. Squirrel fur becoming money is an ontological consequence, but of what exactly, in this latter case? I suggest that we postulate that a CA-event must have taken place (when the relevant criteria were met), and thereby collective CA-state must have come to existence, grounded in the member-level states (we will below discuss what kind of state it is; perhaps it is a merely nominal or attributed state rather than a real intentional state of the collective; but even then, there are several options as to what kind of attributed state it is). That CA-event (collective acceptance as true of the sentence “squirrel fur is money”—or of a related Sellarsian (1981) dot-sentence expressing the same content as “squirrel fur is money” expresses in English) was sufficient for it to be true for the group that squirrel fur is money. And as it is true, there is a group-relative fact that squirrel fur is money, and as the fact includes the property of being money, this property exists for the group. And that fact and property, then, are the ontological consequences—squirrel fur has (really) acquired the property of being money for the group, and future historians can state truly that squirrel fur was money and became money then-and-then. (Even though it is mind-­ dependent and group-dependent, it can be real; see Schweikard and Laitinen 2023). Corresponding to the CA-state of the collective, there are further beliefs and intentions by the collective and by the members. There is a

122 

A. Laitinen

normative requirement that the acceptance of the sentence shows up in other beliefs and intentions; especially, the members should be behaviourally disposed to act accordingly. Perhaps they are not, and in that case one could say that they violate “requirements of rationality” (Broome 2013) analysable as ought-to-be -norms (Tuomela 2002, pp.  45–7, 178–80). If I understand Tuomela’s view correctly, there is a collective intentional state of acceptance, had by the collective, concerning sentence s, which entails a collective commitment to act accordingly, and it leads (in the happy cases) to the formation of suitably related behavioural dispositions and beliefs in the members, as well as to member-level commitments to act accordingly and sanctions those who do not.9 Without the members’ commitments and dispositions, they would not act accordingly, and the institution would be in danger. This leads us to the scenario of the members’ dispositions deteriorating over time, and at some point it may be that so few have the required member-level states, that the collective CA-state vanishes, with the ontological consequence of squirrel fur no longer being money. Again, there are two events constitutively related: the CA-state vanishes, and squirrel fur ceases to be money. Neither the squirrel fur or the collective vanishes, but they cease to have a property or they cease to be in a state. On this reconstruction of Tuomela’s views, we have a CA-event leading to a CA-state, persisting until a negative CA-event takes place. It is the CA-events that are constitutively tied to the institutional events of squirrel fur becoming money, and in the negative case, squirrel fur ceasing to be money. (The CA-state between the events need not have any ontological constitutive relation to institutional reality—we can think of the continuing existence as automatic until cancelled by the negative CA-event.).10 The primary target of the CA is a sentence, but it has ontological  In whatever sense it is that collectives have intentional states—Tuomela later makes explicit that group’s attitudes are in some relevant sense fictitious (see Tuomela 2013). 10  Promises are a good example of how events make a difference, but the state does not need to be sustained by anything. The event of promising creates the obligation, and the promisor’s or promisee’s attitudes are not needed for sustaining the obligation, and the obligation vanishes only when fulfilled (an event), or when the promisee releases the promisor of the obligation (an event). It is possible to argue that similarly only CA-event have ontological consequences, and not the CA-state. 9

6  What Is Collective Acceptance and What Does It Do? 

123

consequences on what could perhaps be called indirect targets, such as social and institutional entities and properties. The CA-events and CA-states are located at the level of the collective, but they are constitutively and normatively related to member-level events and states. That concludes our answer to the third puzzle. With this clarification and reconstruction of CA-events and CA-states, let us now address the question whether the CA-state is an attitude, a disposition to act, an intention, or a belief. On this proposed reconstruction, it remains true that the CA-event always leads to a CA-state. Thus, we side with Tuomela’s claim that “there is no conceptual room here for agreeing to accept s and not accepting s” (2002, p. 130). But the collective CA-state is not a behavioural disposition—that claim by Tuomela needs to be revised. It is the members who have behavioural dispositions, backed by commitments to act. The collective CA-state is, like the CA-event, something at the “groupjective” level of the collective; just like norms are. The collective CA-state is not an intention or a goal either: the collective CA-state is that of acceptance as true of a sentence, to be accompanied by a collective intention, and the member-level intentions that come with it. Similarly, the CA-state at the collective level is to be accompanied by member-level beliefs. The nature of the collective CA-state seems to be that of “acceptance as true” i.e. acceptance-belief, so pace Tuomela’s formulations in which it is e.g. a disposition to act, or sometimes an intention, a coherent reconstruction takes it to be what it seems to be: acceptance as true.11 That is the answer to the second puzzle. (Tuomela in later books and articles developed his accounts of group-­ level and individual-level attitudes, and their constitutive, normative and causal relations further; I take the reconstruction that locates the CA-states and CA-events at the collective level to be in line with Tuomela (2007, 2013); with the proviso that Tuomela (2013) regards the intentional states of collectives as fictitious. See also Tuomela et al (2020). So far, this reconstruction has tried to make sense of collective acceptance being an attitude of sorts, perhaps involving or rationally requiring  Even though such a state has ontological consequences and so is in the business of “making” things and not merely “taking in”, the temptation to say that it thereby has the direction of fit of intentions, is to be resisted. See Laitinen (2014). 11

124 

A. Laitinen

dispositions to act, or commitments to action, and located at the level of the collective and not at the level of the members. The final challenge (or the puzzle mentioned first in the previous section) is to make sense of reflexivity and performativity: it seems that squirrel fur being money requires not only Collective Acceptance of the related sentence, but also something like action, practice or patterns of behaviour. Here, it seems that the best answer, in requiring minimal changes in Tuomela’s claims, is the No True Scotsman–answer. The claims about reflexivity and performativity are true of effective collective acceptance, that is, collective acceptance accompanied by appropriate action. It is conceivable that otherwise similar collective acceptance, with membership-level dispositions, would exist, but countering dispositions (created e.g. by distrust) would block effective action. In that case we might say that there was in all other respects “genuine” collective acceptance which led for example to the formation of the relevant dispositions and commitments, but effective collective acceptance was lacking: people did not act on the assumption that squirrel fur is money, and so squirrel fur was not money. Thus, whatever the state was that the collective and its members were in, it cannot have been effective collective acceptance. Reflexivity and performativity are true of effective collective acceptance only. And insofar as collective acceptance is an achievement notion, we must bite the bullet and grant that only effective collective acceptance is collective acceptance at all; the otherwise “genuine” state which is not effective, can only be an attempted collective acceptance, or ineffective “collective acceptance”. All collective acceptance is effective collective acceptance.

5 Conclusion This reconstruction of Tuomela’s (2002) views has ended up defending the following claims: 1*) Organizations are institutions, with roles that come with rights and tasks, but some organizations such as business corporations may not have been designed for the benefit of the group, or collectively accepted

6  What Is Collective Acceptance and What Does It Do? 

125

by the group in the we-mode, so they do not meet all of Tuomela’s criteria for we-mode collective-social institutions. 2*) The existence and nature of institutions is explained not only by collective acceptance (of sentences expressing constitutive norms of institutions) but also by related social practices. 3*) Much of social and institutional reality is constructed by humans, and the CA-theory captures such social construction. Tuomela allows, that other aspects of social and institutional reality may be generated e.g. by invisible hand mechanisms. 4*) Collective Acceptance events are the formations of Collective Acceptance states (and negative CA-events are the disappearances of CA-­states). The CA-states are located at the level of the collective, and the attitudes of members are normatively, constitutively, and causally related to them. The collective CA-state is a collective version of acceptance as true of a sentence, i.e. a belief-acceptance. CA may require also collective intentions and goals, but the collective CA-state of acceptance as true of a sentence is a separate state from those practical states. As nothing can arguably have a double direction of fit, the collective CA-state does not have it either. 5*) An institution or a social entity exists if it is collectively accepted as existing; unlike trees and stars that exist independently. Talk about collective acceptance of institutions, persons, objects, properties is to be understood in terms of collective acceptance of sentences as true for a group, as in CAT. “(CAT) A sentence s is collective-social in a primary constructivist sense in a group g if and only if it is true for group g that (a) the members of g collectively accept s, and that (b) they collectively accept s if and only if s is correctly assertable (or true) (collective acceptance thesis)” (Tuomela 2002, p. 132). Of the different ways of connecting collective acceptance of sentences to the existence of institutions, the realist version distinguishing group-dependent and group-independent grounds of facts was favoured.

126 

A. Laitinen

6*) An institution is collectively accepted as existing if it exists. There is no way for an institution to exist if it is not collectively accepted. While other aspects of social reality may exist independently of collective acceptance, institutions are not among them. Further, if an institution does not exist, the corresponding collective acceptance-state does not exist. This only holds of effective states of collective acceptance; but as collective acceptance is an achievement-notion, there are no other forms of collective acceptance. And, to repeat, there were two other claims mentioned but not examined in this paper: 7) The key to social and institutional reality is the way in which organizations, other institutions, social practices, collective goals and beliefs, are always for a group. 8) Collective attitudes had in the we-mode (entailing collective commitment and the collectivity condition) are central in how things are for a group. Acknowledgements  I would like to thank Raul Hakli, Pekka Mäkelä, Rachael Mellin and Miguel Garcia for encouraging feedback and critical comments on a draft version. I hope it is not inappropriate to also express my heartfelt gratitude to Raimo Tuomela (1940–2020) for welcoming me in his group’s research meetings and projects especially for the decade or so since 2007.

References Broome, John. 2013. Rationality Through Reasoning. Wiley-Blackwell. Cohen, L. Jonathan. 1992. An Essay on Belief and Acceptance. Oxford: Oxford University Press. Gilbert, Margaret. 2002. Belief and Acceptance as Features of Groups. ProtoSociology 16: 35–69. Hakli, Raul. 2006. Group Beliefs and the Distinction Between Belief and Acceptance. Cognitive systems research 7 (2–3): 286–297. Hindriks, Frank. 2015. Do Social Institutions Require Collective Acceptance? Metascience 24: 467–470.

6  What Is Collective Acceptance and What Does It Do? 

127

Ikäheimo, Heikki, and Arto Laitinen. 2007. Analyzing Recognition: Identification, Acknowledgement and Recognitive Attitudes Towards Persons. In Recognition and Power, ed. Bert van den Brink and David Owen, 33–56. New York: Cambridge University Press. Laitinen, Arto. 2014. Against Representations with Two Directions of Fit. Phenomenology and the Cognitive Sciences 13: 179–199. ———. 2017. We-Mode Collective Intentionality and Its Place in Social Reality. In Social Ontology and Collective Intentionality. Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. Gerhard Preyer and Georg Peter, 147–167. Leiden: Springer. Schotter, Andrew. 1981. The Economic Theory of Institutions. Cambridge: Cambridge University Press. Schweikard, David, and Arto Laitinen. 2023. Realism in Social Ontology. unpublished ms. Sellars, Wilfrid. 1981. Mental Events. Philosophical Studies 39: 325–345. Tuomela, Raimo. 2002. The Philosophy of Social Practices: A Collective Acceptance View. Cambridge: Cambridge University Press. ———. 2003. Collective Acceptance, Social Institutions, and Social Reality. American Journal of Sociology and Economics 62: 123–166. ———. 2007. The Philosophy of Sociality: The Shared Point of View. Oxford: Oxford University Press. ———. 2013. Social Ontology. Oxford: Oxford University Press. ———. 2017. Response to Arto Laitinen. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. Gerhard Preyer and Georg Peter, 169–178. Leiden: Springer. Tuomela, Raimo, and Wolfgang Balzer. 1998. Collective Acceptance and Collective Social Notions. Synthese 117: 175–205. Tuomela, Raimo, Raul Hakli, and Pekka Mäkelä, eds. 2020. Social Ontology in the Making. De Gruyter.

7 Can There Be Institutions Without Constitutive Rules? Frank Hindriks

1 Introduction Institutions are constituted by rules. In this respect, they resemble games. A referee applies the rules of the game. Similarly, police officers enforce the laws of a country. Given that such rules constitute institutions, it makes sense to call them ‘constitutive rules.’ However, this term is often used in a more restrictive way for rules that define institutional actions, are logically prior to them and make such actions possible (Rawls 1955; Searle 1969, 1995, 2010). This can be confusing. One puzzle pertains to the distinction between regulative and constitutive rules (Searle 1969, 1995, 2010). Consider the fact that, in many contexts, people greet each other by means of a handshake. As it regulates the activity, the rule ‘shake hands when you meet’ is a regulative rule. At the same time, it is constitutive of a greeting practice. Furthermore, this practice seems to be an institution. But how can a regulative rule be constitutive of an institution?

F. Hindriks (*) Department of Ethics, Social and Political Philosophy, University of Groningen, Groningen, The Netherlands e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_7

129

130 

F. Hindriks

According to John Searle (1969, 1995, 2010), this is impossible. As I discuss in more detail below, he defends this view in terms of the following two claims: 1 . Constitutive rules differ categorically from regulative rules. 2. Institutions are systems of constitutive rules. To support these claims, Searle compares the rules of etiquette to those of chess. He points out that the activities involved in etiquette can be performed independently of its rules. In contrast, the rules of chess constitute the activities to which they apply. He takes this to imply that the two activities are very different. Because of this, it makes sense to regard playing chess as an institutional activity, whereas the activities involved in etiquette form social practices instead. Searle argues that constitutive rules are logically prior to the activities they define, create and regulate. In a similar vein, Raimo Tuomela maintains that constitutive rules “create and define” institutions and “enable new kinds of activities.” (2020, pp. 2, 4) So, Tuomela accepts Searle’s first claim: (1) constitutive rules differ categorically from regulative rules. However, as I discuss in more detail below, Tuomela rejects the second claim (2). He does not require that institutions involve constitutive rules, but allows for institutions that depend on regulative rules. This means that the distinction between social practices and institutions has to be drawn in other terms. Tuomela (2002, 2003, 2013) has argued that, in contrast to ordinary social practices, institutions are governed by social norms. The view that the distinction between regulative and constitutive rules is a categorical one has been heavily criticized (for an overview, see Hindriks 2009). The main criticism that has been voiced is that it is merely a linguistic distinction instead (Ransdell 1971; Warnock 1971). I have developed this critique elsewhere in terms of the “Transformation View” of constitutive rules (Hindriks 2005, 2009). Its core claim is that a regulative rule can be transformed into a constitutive rule by doing little more than adding an institutional term. Here I argue that, because of this, the distinction cannot be used to distinguish institutions from ordinary social practices. Thus, I reject both of Searle’s claims (1 and 2).

7  Can There Be Institutions Without Constitutive Rules? 

131

Both Searle (2015) and Tuomela (n.d., 2017) have heavily criticized the Transformation View. In this chapter, I respond to their criticisms. In Sect. 1, I introduce the Transformation View. In Sect. 2, I present and respond to their criticisms. This serves to provide further support for the claim that the distinction between constitutive and regulative rules is not a categorical one (1 is false). In Sect. 3, I criticize Searle’s claim that it can be used to differentiate between social practices and institutions (2 is mistaken). None of this means, however, that the notion of a constitutive rule has no value. In fact, in spite of my differences with Searle and Tuomela, I believe that it can play an important role in the analysis of institutions. Constitutive rules are particularly useful for understanding complex institutions that are rich in symbolism.

2 The Transformation View According to Searle (1969, 1995, 2010), institutions exist due to constitutive rules. What they are can be explained by contrasting them with regulative rules. Such rules regulate behavior that is possible independently of them. Consider the rule ‘eat with a knife and fork.’ This is something people can do irrespective of whether it is regulated by a rule. The same holds for wearing a tie to a formal dinner as well as for shaking hands when meeting someone. Regulative rules require the performance of such actions. However, constitutive rules do more than that. They do not only (1) regulate behavior. They also (2) define the behavior they regulate, (3) are logically prior to it, and (4) make it possible (see also Rawls 1955; Searle 1969, 1995, 2010). For instance, the rules of chess regulate how a game is to be played. But they also logically prior to the game, define it, and make playing it possible. It follows that, without those constitutive rules, it is not possible to engage in the activity. As features (2)–(4) are distinctive of constitutive rules, they differ categorically from (ordinary) regulative rules. The two kinds of rules also have a different structure. Regulative rules have the form: If X, do Y. In contrast, the structure of a constitutive rule is given by the counts-as locution: X counts as Y in C. For instance, ‘Bills issued by the Bureau of Engraving and Printing count as money in the

132 

F. Hindriks

United States.’ (Searle 1995, p. 28)1 Searle is not very explicit about what the phrase ‘counts as’ is supposed to mean. But to count something as Y is presumably a matter of classifying it as such. Furthermore, the constitutive rule that licenses doing so has to be in force, which means it must be accepted. In light of this, I take ‘X counts as Y’ to mean that X has status Y because the constitutive rule according to which this is the case is generally accepted.2 Searle realizes that characterizing the distinction in terms of a difference in structure carries risks. One might, for instance, say that ‘shaking hands counts as greeting’ and that greeting is not possible independently of the rule that requires it. As this rule has the structure of the counts-as locution, it seems to be a constitutive rule, which it is not. ‘Greeting’ is merely a term for complying with a regulative rule. For a rule to be constitutive, there must be ‘consequences’ attached to the status (Searle 1969). What this was supposed to mean remained unclear for a long time. However, his more recent work reveals that statuses come with deontic powers, to wit rights or obligations. Someone who greets someone else does not thereby incur further deontic powers. Thus, constitutive rules feature statuses that involve deontic powers. In fact, having such powers is part of the concept of the status (Searle 1995). Just as there are terms for complying with a rule, there can be terms for violating them. For instance, ‘a ghost driver’ is someone who violates the rule of driving on the right side. Now, being a ghost driver can have consequences. One can even incur an obligation, that to pay a fine. Yet, this is not part of the concept of being a ghost driver. Regulative rules can also feature rights or obligations. For instance, the rules of etiquette specify actions that people are supposed to perform. Their basic structure is: ‘If

 It is not entirely clear whether this structural feature should be regarded as part of what constitutive rules are. Although Searle mentions it throughout his work, he has also made a few claims that suggest a weaker view. In particular, early on he claimed that constitutive rules can have the same structure as regulative rules (Searle 1969, p. 34). Strikingly, this comes close to the claim that I defend below, that regulative rules can constitute institutions as well. 2  Searle (1995) and Tuomela (1995, 2013) maintain that institutions must be accepted collectively. By this they mean that they depend on attitudes that are irreducibly collective. Tuomela takes this to imply that people must be collectively committed to the relevant rules. Searle (2010) has abandoned this requirement and regards general acceptance as sufficient. 1

7  Can There Be Institutions Without Constitutive Rules? 

133

X, it is obligatory to do Y.’3 Even so, these deontic powers are not part of a status. Thus, constitutive rules have two more features. First, they have their own structure. Second, they feature statuses that implicate deontic powers. This confirms that constitutive rules differ categorically from regulative rules. Because of this, Searle proposes that institutions are ‘systems of constitutive rules’ (1969, p.  51). Regulative rules or systems thereof do not constitute institutions. Presumably, they form social practices instead. The upshot is that ordinary social practices depend on regulative rules, while institutions exist due to constitutive rules. The main objection that has been voiced against Searle’s account of regulative and constitutive rules is that it merely marks a linguistic distinction (Ransdell 1971; Warnock 1971). The considerations concerning greeting and ghost driving amount to a version of this objection. As discussed, Searle seems to have an adequate answer to it. However, the Transformation View presents a deeper challenge to it. Its key claim is that regulative rules can be transformed into constitutive rules without adding anything that is qualitatively different. All you need to do to formulate a constitutive rule is to add a status term to a regulative rule. If this is correct, constitutive rules do not create new activities. Instead, they are regulative rules dressed up in institutional language (Guala and Hindriks 2015, p. 189). To illustrate the Transformation View, I start from the following ordinary regulative rule (Hindriks and Guala 2015, p. 472): [R] If one is the first to occupy a piece of land, one has the right to its use.

This rule R can be split into two parts, the antecedent and the consequent: If one is the first to occupy a piece of land; One has the right to its use. The next step is to introduce an arbitrary term, such as ‘property*’. (I add an asterisk in order to emphasize that this is a hypothetical or toy  More complex regulative rules feature permissions, prohibitions or other kinds of obligations or rights. Note that Searle takes regulative rules to have the structure of an imperative. And some such rules do indeed to have this surface structure. Think, for instance, of ‘wear a tie to a formal dinner’ or ‘keep to the right.’ However, such statements specify actions that people are supposed to perform. In light of this, I believe that they are better explicated in terms of deontic powers (cf. Hart 1961). 3

134 

F. Hindriks

example, that should as such not be confused with the actual institution of property, which is considerably more complex.) This term is to be used to turn both halves into rules on their own, as follows: [B] If a person first occupies a piece of land, then it is her property*. [S] If a piece of land is someone’s property*, she has the right to use it.

I call ‘B’ a base rule, and ‘S’ a status rule. This transformation process is perfectly general and can as such be applied to any regulative rule. As another (toy) example, consider the rule that the person with the hammer should open the meeting. This can be transformed into two rules by adding the term ‘chairperson*’. The base rule of this institution is: ‘The person with the hammer is the chairperson*’. And the status rule is: ‘The chairperson* should open the meeting.’ The transformation is achieved by introducing a status term. The modifications of the two parts of the original rule are conservative in that they do not add anything that is fundamentally new (Belnap 1993). All they do is add an arbitrary label. Base rules are rather similar to constitutive rules as Searle conceives of them. To see this, consider a slight reformulation of the main example: A piece of land is the property* of the person who is the first to occupy it. When a context is added and the counts-as locution is incorporated, the rule is as follows: In the Wild West, a piece of land counts as the property* of the person who is the first to occupy it. Similarly, such an individual counts as the owner* of this piece of land. The chairperson* rule can be transformed in a similar manner: People with a hammer count as chairpersons* (in the context of a meeting). It follows that constitutive rules can indeed be derived from regulative rules just by introducing an arbitrary status term. In other words, regulative rules can be transformed into constitutive rules without adding anything that is genuinely novel. The upshot is that the distinction between regulative and constitutive rules is not a categorical one. Instead, the distinction is a linguistic one. As mentioned above, constitutive rules are indeed regulative rules dressed up in institutional language. However, this does not disqualify constitutive rules in any way. Status terms are meaningful. And they refer to institutional entities. In

7  Can There Be Institutions Without Constitutive Rules? 

135

fact, status rules explicate the meaning of status terms, at least in part. And base rules specify the application conditions they have in the context at issue. In this way, the Transformation View illuminates the meaning and significance of status terms. A second virtue of the Transformation View is that it accommodates all four features of constitutive rules. Status rules (1) regulate and (2) define behavior. And base rules are (3) logically prior to that behavior and (4) make it possible. Strikingly, Searle’s own proposal does not accommodate all of them. The thing to appreciate is that, as Searle conceives of them, constitutive rules need not mention behavior at all: both the X-term and the Y-term in the counts-as locution can refer to things other than actions. Now, if a rule is to regulate or define behavior, it should mention it. So, Searle’s proposal fails to account for (1) and (2). This is supported by one of Searle’s own examples: ‘Bills issued by the Bureau of Engraving and Printing count as money in the United States.’ Such rules neither (1) regulate, nor (2) define, behavior. This is why it is so important to complement such rules with status rules. What about the other two features of constitutive rules? I have just argued that, in spite of Searle’s claim to the contrary, constitutive rules do not regulate behavior. One might be tempted to conclude from this that they cannot be (3) logically prior to the behavior they regulate or (4) make it possible. But this would be too quick. The thing to appreciate is that constitutive rules or base rules can be logically prior to the behavior that status rules regulate. And they can also make such behavior possible. For instance, the constitutive rule of money can be logically prior to the use of dollar bills as a means of exchange, and it can make this activity possible. Thus, Searle’s view captures at best two of the four features that he attributes to constitutive rules. In contrast, the Transformation View accommodates all of them.4 The proposed analysis has important ramifications for the relation between institutions and rules. As just discussed, Searle claims that institutions depend on constitutive rules. One implication of the  Because of this, Guala and I use the term ‘constitutive rule’ for a combination of a base rule and a status rule, a B and an S that pertain to one and the same status (Hindriks and Guala 2015). To avoid confusion, I follow Searle and Tuomela here and use the term for base rules only. 4

136 

F. Hindriks

Transformation is that, if an institution involves such a rule, it will also feature a status rule. The reason for this is that, as just argued, constitutive rules do not regulate behavior, whereas status rules do. A second even more important implication is that institutions do not require constitutive rules at all, at least not as Searle conceives of them. They can depend on regulative rules instead. Land can be someone’s property and someone can be a chairperson even if people do not have words for these statuses. It will just be more cumbersome to talk about them, as doing so will requires mentioning the relevant status rules and, in some cases, the base rules as well. Base rules provide a specification of the features that something needs to have in order to have a status. But those features also figure in the antecedent of a regulative rule. And the relevant behaviors feature in their consequents. In light of this, I conclude that regulative rules can be constitutive of institutions.5 At this point, one might object: But if institutions are constituted by rules, surely those rules will be constitutive rules. In response, I distinguish between two conceptions of constitutive rules, a narrow and a broad one. On the narrow conception, only base rules are constitutive rules. Institutions can indeed be constituted by base rules, as long as they are combined with status rules. On the broad conception, ordinary regulative rules can also qualify as constitutive rules. In fact, institutions can be constituted by regulative rules only. Constitution is a relation of necessity. And there is indeed a necessary connection, for instance between shaking hands and the rule that regulates this activity. Given that this rule is generally accepted, it is necessarily the case that, when you meet someone, you ought to greet them by shaking their hand (in most Western countries). This rule fixes the identity of the institution. According to the Transformation View, an institution is constituted either by a regulative rule or by a base rule along with a status rule.6 The first claim may sound odd to those familiar with Searle’s notion of a constitutive rule. But I believe they can and should get used to it. The claim that institutions are constituted by constitutive rules is even stranger.  When a regulative rule is in force, there will also be a status, even if only implicitly. For instance, money and property function as such even if people do not have words for these statuses (Hindriks and Guala 2015, pp. 473–477). 6  As they regulate behavior, status rules are a special kind of regulative rule. 5

7  Can There Be Institutions Without Constitutive Rules? 

137

And the Transformation View provides a way to avoid it. It has two virtues. First, it accommodates all features that are commonly attributed to constitutive rules. Second, it provides a crystal-clear account of how such rules relate to regulative rules. For these reasons, I propose, it is to be preferred to Searle’s account. In spite of this, it has met with strong opposition from both Searle and Tuomela. I now go on to discuss and respond to their criticisms.

3 Criticisms of the Transformation View Searle (2015) and Tuomela (n.d., 2017) characterize the Transformation View (TV) as an attempt to reduce if not eliminate constitutive rules.7 They argue that it fails to do so. In particular, TV does not account for the fact that constitutive rules enable new forms of behavior and support concomitant deontic powers. To respond to Searle’s criticisms, I consider his examples of chess and etiquette in more detail. The idea that they concern fundamentally different activities is very intuitive. Searle uses this intuition to support his claim that chess is an institution and etiquette an ordinary social practice. However, this move is not without its problems. First, even if they are very different, why would this mean that etiquette is a social practice rather than an institution? Famously, Norbert Elias (1969) regarded etiquette as one of the most important institutions in civilized societies. Furthermore, just as other institutions, etiquette is a system of rules that specify how people are supposed to behave in social settings. Someone might object that etiquette is informal. However, Searle does not draw the distinction between social practices and institutions in these terms. For instance, even though they are formal, Searle does not take traffic rules to be institutions. In fact, he criticizes Guala and me for regarding them as institutions: “Driving on the right is not an  Tuomela presented his criticism in two places. In a response to a critical appreciation that I made of his book Social Ontology (Tuomela 2017). And in an unpublished comment that was intended to be included as an appendix to his paper ‘Social Corporations as Social Institutions’ (Tuomela 2020). I include the comment as an appendix to this paper. And I thank his wife, Maj Tuomela, who has been so kind as to grant me permission to use and reproduce it here. 7

138 

F. Hindriks

institution. Why not? It is a legally required practice in the US, but it does not generate deontologies in the way of institutions like private property.” (Searle 2015, p. 511) However, laws are paradigmatic examples of institutions. More importantly, traffic rules do feature deontic powers, including for instance the right of way.8 In light of this, it should not be taken for granted that etiquette is a social practice rather than an institution. Second, chess is not a paradigmatic example of an institution. It is a game. And the activities that are part of games rarely make much sense outside of that context. For instance, trying to place a ball within the confines of a net has little or no value outside of sports like basketball or soccer. But a lack of external value is not a general feature of institutions. For instance, bringing food and drinks to people can count as waiting on tables. And learning about a topic by reading books and listening to experts can count as studying. But these activities also have value independently of these institutional statuses. So, the powerful intuition that chess differs fundamentally from etiquette may in fact be unrelated to the distinction between social practices and institutions.9 Searle (2015 p. 510) also criticizes Guala and me for ‘smuggling’ deontic powers into our transformation. But we do nothing of the sort. We start from Searle’s own conception of regulative rules as imperatives (Hindriks and Guala 2015). Because of this, we introduce deontic powers as a separate step. This is not a deficiency in our derivation. Instead, this step is needed due to a shortcoming of Searle’s conception of regulative rules (see note 3). Tuomela argues instead that “regulative norms cannot do the central job of constitutive norms”, which is to “create behavior that was not earlier normatively (indeed, deontically) possible.” (2017, p. 216) However, it is unclear to me what he means by this. After all, regulative rules or norms feature deontic powers, just as status rules. So, it seems that they can do what Tuomela claims they cannot do. To be  Earlier Searle claimed that the rule ““drive on the right side of the road” regulates driving; but driving can exist prior to the existence of that rule.” (1995, p. 27) 9  See Raz (1986) for an insightful discussion of internal value and social meaning. MacIntyre (1981) uses the contrast between internal and external values to distinguish between social practices and institutions, but exactly in the opposite way. He argues that values are internal to social practices and external to institutions (which he takes to be complex organizations). 8

7  Can There Be Institutions Without Constitutive Rules? 

139

sure, at some point institutional reality is so complex that it is too cumbersome not to switch to a framework that features base and status rules. But in principle there is nothing the latter can do that the former cannot do. Searle subscribes to the narrow conception of constitutive rules. It is less clear whether Tuomela does so, which makes it difficult to assess some of his other criticisms. Just as Searle, he claims that many constitutive rules have the structure of the counts-as locution (Tuomela, appendix, 2017, p. 215). This leaves open the possibility that other rules, such as status rules or ordinary regulative rules, might be constitutive rules as well, as the broad conception has it. Unfortunately, he sets aside the question whether regulative rules can be constitutive of institutions (Tuomela 2017, p. 215). Shortly thereafter, he claims that he “is skeptical about whether also regulative norms could [enable] even in some cases.” (Tuomela 2017, p. 217) This suggests that he feels the attraction of the broad conception, but is more sympathetic to the narrow conception. At the same time, he claims that our regulative rule about property* is “also constitutive as to its form because having the right involves a right-­ involving status function entailing deontic power.” (Tuomela, appendix) But this is a regulative rule. Hence, Tuomela’s claim that it is constitutive suggests that he subscribes to the broad conception of constitutive rules after all. All in all, it is not clear where he lands on this issue.10 He goes on to argue that “form need not distinguish them but rather their different functions.” (Tuomela, appendix) By this he means that: “A constitutive rule should primarily say what something X is constituted as, while the regulative rule says how the constituted item or fact is supposed to be used or to function.” (Tuomela, appendix) However, this is exactly what base rules and status rules do. And it makes me wonder how deep our disagreement goes. Then he claims that Guala and I “incorrectly mixed  Tuomela claims that “R does not say how the institution of property can be used.” (n.d.) However, the right to use that which you own is in principle unrestricted. So, R is formulated as it should be. To be sure, it is incomplete as an account of property, because it does not feature the right to exclude others from using it and the rights of transfer. But this is because it is a toy example. Note that Searle (2015, p. 510) disqualifies our account of property as unrealistic. But he fails to note that we are not concerned with the everyday institution of property, which is why we add an asterisk to ‘property*’, as I have done in Sect. 1. 10

140 

F. Hindriks

constitutive rules with regulative ones.” (Tuomela, appendix) I wonder what he meant by this. Maybe he takes issue here with the fact that we characterize the combination of a base rule and a status rule as a constitutive rule. But this is just a terminological issue. As I mentioned in Sect. 1, I have no problem reserving the term for base rules. All this suggests that Tuomela’s own stance on these issues is not entirely clear, which makes it difficult to assess and respond to his criticisms. Tuomela argues that “the Hindriks-Guala proof does not succeed in eliminating constitutive rules in favor of regulative ones” (n.d.) But this is not what we are trying to do.11 To be sure, the Transformation View reveals how people could in principle do without constitutive rules. But this does not mean that eliminating them altogether is a good idea. Importantly, transformation is not elimination.12 In fact, it provides a form of validation. It reveals that status terms can be introduced as a conservative extension of the language. And this implies that they refer to entities with a status (Hindriks and Guala 2015, p.  477; Lewis 1983, p. 79). At the same time, it does imply that the distinction between regulative and constitutive rules is not a categorical one. However, as discussed in Sect. 1, it preserves the idea that constitutive rules are meaningful. TV reveals that regulative rules feature all ingredients of a constitutive rule, except for status terms. However, according to TV, such terms are not required for entities to have statuses (see note 5). Someone can, for instance, own a house or chair an audit committee even if the relevant practices are sustained by regulative rules that do not feature terms such as ‘property’ and ‘chairperson.’ This is not to deny that words are important for social reality (see Sect. 3 for more on this). Instead, it is to say that labels are not strictly necessary for things to have statuses. The statuses can already be there, even if the rules that constitute them leave  In fact, our main goal is to integrate the notion of a constitutive rule in a unified theory of institutions (Guala and Hindriks 2015; Hindriks and Guala 2015). This theory combines it with regulative rule theories as well as equilibrium theories. See Ross (1957) for an attempt to eliminate constitutive rules avant la lettre. He regards status terms as meaningless. 12  In a similar vein, Copp (2020) observes that reduction is not elimination. 11

7  Can There Be Institutions Without Constitutive Rules? 

141

them implicit. In this way, TV supports the claim that regulative rules can be constitutive of institutions. Unfortunately, Tuomela does not discuss this part of TV as it “concerns a topic that goes beyond my explicit discussion in [Social Ontology].” (2017, p. 217) Yet, this part is crucial for my claim that TV does account for the fact that constitutive rules enable new forms of behavior. A distinct feature of Tuomela’s account is that “a constitutive rule… is an irreducible higher-order normative presupposition (a group-level ought-to-be rule)” (appendix) The claim is that group members ought to uphold and promote an institution and thereby make it the case that the relevant entities have a status.13 He adds: “This is an important point against the kind of reduction of constitutive norms to regulative ones that Hindriks advocates” (Tuomela 2017, p. 216)14 And indeed, this is a clear difference between our views. Base rules are not higher-order ought-to-be rules. But it is unclear what such rules are exactly and why there would be any such rules. Tuomela claims that constitutive rules are higher-order norms because they require compliance with regulative rules. It requires acting as if something has the status that features in such a rule. To this end, people must comply with first-order regulative rules—status rules in my framework. However, regulative rules feature deontic powers. As such, they require compliance. This raises the question of what exactly constitutive rules are meant to add. For instance, acting as if someone is a chairperson seems to be nothing other than acting in accordance with the relevant

 Elsewhere, Tuomela elaborates on this as follows: “Given this, we can say that there is a constitutive norm “E counts as the ethos of g” (e.g. E could express the norm that g be democratic). This norm is a constitutive, conceptually necessary ought-to-be norm for g: Things ought to be so that the ethos E be upheld and promoted in g by the members through their related activities, and accordingly there will be relevant regulative norms specifying what kinds of democratic activities are normatively expected from the members of g.” (2017, pp. 215–216) 14  Tuomela (2017, p. 216) goes on to claim that Searle makes the same point. However, Searle’s account does not feature higher-order rules or ought-to-be norms. Instead, his point is that we start with imperatives and end with deontic powers, as I discussed earlier in this section. 13

142 

F. Hindriks

regulative or status rule.15 In light of this, the most distinctive feature of Tuomela’s account strikes me as mistaken.16 In sum, constitutive rules do not add anything to regulative rules that is genuinely new. To be sure, they introduce status terms. But this does not make them sufficiently different to regard the distinction as a categorical one. In this respect, it is fair to regard TV as a reduction of sorts. However, in other respects, this characterization can be misleading. For one thing, I do not claim that constitutive rules are ‘nothing but’ regulative rules. After all, they do introduce status terms. Furthermore, those terms are meaningful and have referents.17 This insight lies at the heart of the claim that I defend in the next section, where I argue that constitutive rules serve to explicate the symbolic dimension of institutions.

4 Social Practices and Institutions According to Searle (1969, 1995), institutions are systems of constitutive rules. Such rules “enable the creation and maintenance of status functions” (Searle 2015, p. 507) Ordinary social practices depend on regulative rules instead. I have just argued that constitutive rules do not differ categorically from regulative rules and that institutions can be constituted by regulative rules. It follows that the distinction between the two kinds of rules cannot be used to differentiate institutions from social  Garcia (1987) defends the somewhat similar idea that constitutive rules come with classificatory obligations. The idea is that people ought to classify, for instance, dollar bills as such and that they ought to correct others who misclassify them. However, if there are such obligations, they are generic linguistic obligations that are not specific to status terms. 16  Tuomela also claims that constitutive rules or norms constitute and create “groups capable of acting as groups” (2017, p. 216) But in what sense? Earlier, Tuomela (2013) had argued that group agents are both causally efficacious and fictitious. As these claims are difficult to square, I pressed him on this (Hindriks 2015, 2017). In response, he has admitted that they are only ‘loosely fictitious’, the point being that their intentional attitudes are extrinsic instead of intrinsic (Tuomela 2017, p. 215). (The claim that they are fictitious no longer appears in the paperback edition of Social Ontology.) This leaves me with the question of why, as Tuomela claims, regulative rules cannot do the ‘job’ of constituting group agents (2017, p. 216). Perhaps this is because, on his view, constitutive rules are ought-to-be rules. 17  What is more, I regard entities that have statuses as irreducibly institutional. As I have argued elsewhere, they are materially constituted by entities at a lower level (Hindriks 2012, 2013). 15

7  Can There Be Institutions Without Constitutive Rules? 

143

practices. Here, I ask first, whether and how the notion of a constitutive rule can be used to enrich our understanding of institutions. And second, how institutions do differ from social practices. Constitutive rules contain status concepts. As they feature in both, such concepts connect base rules with status rules. This is mirrored by the fact that statements of such rules contain status terms. Status concepts and terms make it easier to navigate the social world. They are defined, at least in part, by their characteristic deontic powers, which are specified in the relevant status rules. Because of this, people who are familiar with the relevant terms and possess the corresponding concepts will know without reflection that the statuses come with certain deontic powers. This is why, for instance, someone who is in a meeting need not recall the relevant regulative rule to realize that the chairperson is the one who gets to determine who speaks when and about what.18 Status concepts and terms form the core of what I regard as the representational or symbolic dimension of institutions. This dimension is important for various reasons. I have just argued, in effect, that status concepts and terms economize on cognitive and linguistic resources (cf. Tuomela 2013). But there is more. Base rules specify ‘base properties’, which an entity must have in order for it to have a particular status. For instance, one-dollar bills have an image of George Washington on them. This is readily observable. Similarly, traffic signs are easy to spot and to interpret. Thus, base properties help people recognize status entities. In some cases, they serve to distinguish genuine instances from fakes or counterfeits. However, not all statuses are readily observable. You cannot see whether someone is a doctor just by looking at her. Except if she wears a white coat. To double check, you might look for a certificate of a medical degree on her wall. As this example reveals, people often rely on status indicators to recognize otherwise invisible statuses (Searle 1995). Think also of  This is due to the fact that concepts involve dispositions to draw certain inferences (Peacocke 1992). However, someone might have an imperfect understanding of a status concept and lack a full grasp of the deontic powers that are involved. This is particularly likely when status terms have a legal meaning, which can be quite technical. Even then, however, status terms and concepts can be useful as placeholders. Participants will have some sense of what they are all about and can consult experts if need be. 18

144 

F. Hindriks

badges, gavels, passports and wedding rings. Status indicators have an epistemic function. Furthermore, they facilitate coordination. They signal how people can converge on convenient ways of interacting, which is why Guala and I refer to them as ‘signaling devices’ (Guala and Hindriks 2015; Hindriks and Guala 2015).19 Thus, status concepts and terms fulfill cognitive and linguistic purposes. Furthermore, status indicators have epistemic and coordinative significance. Finally, all these aspects can contribute to the stability of an institution. This reveals that the symbolic dimension of institutions is rather significant. And it can be captured in terms of constitutive rules or, more specifically, in terms of base rules and status rules. Yet, for reasons just discussed, such rules cannot be used to differentiate institutions from social practices. Furthermore, symbols play a much more substantial role in some institutions as compared to others. In light of this, the distinction has to be drawn in other terms. The alternative is to do so in terms of the normative dimension of institutions (Hindriks 2019, 2021; Tuomela 2002, 2013). The idea is that institutions feature deontic powers, while ordinary social practices do not. To be sure, social practices may have other normative features. For instance, there will be a correct and an incorrect way to respond to a signaling device. Furthermore, social practices can realize certain social values, such as efficiency or safety. And it might be morally wrong to participate in some of them. But they do not, as such, feature characteristic rights and obligations. Such deontic powers, I propose, are distinctive of institutions. Thus, I propose that institutions consist not only of social practices but also of social norms. Furthermore, such norms are constituted by regulative rules. Finally, they govern the social practices. Thus, institutions are social norms that govern social practices (Hindriks 2019, 2021). In this respect they differ from ordinary social practices, which are not governed by social norms. In a similar vein, Tuomela (2003, p. 123; see also 2002, 2003, 2013) characterizes institutions as norm-governed social practices.  Base properties and status indicators can also be captured by (the antecedents of ) regulative rules. However, if they are sufficiently complex, it is more convenient to specify them in separate base rules. 19

7  Can There Be Institutions Without Constitutive Rules? 

145

Furthermore, he adds that those norms can be either regulative or constitutive. In other words, he distinguishes, in effect, between regulative-rule institutions and constitutive-rule institutions. It follows that, in spite of our differences, Tuomela and I agree on the core issue. Both of us provide an affirmative answer to the question of this paper: Yes, there can be institutions without constitutive rules.20 In sum, institutions differ from ordinary social practices in that they consist of rules that feature deontic powers. Those rules can be regulative or constitutive. Instead of being distinctive of institutions as such, constitutive rules distinguish constitutive-rule institutions from regulative-rule institutions. I have argued that the notion of a constitutive rule is best explicated in terms of two other kinds of rules: base rules and status rules. Such rules are particularly useful for capturing complex institutions with a rich symbolic dimension.

5 Conclusion According to the Transformation View, regulative rules can be transformed into constitutive rules, by introducing a status term. This reveals that, pace Searle and Tuomela, the distinction between the two kinds is not a categorical one. A fortiori, it cannot be used to distinguish between ordinary social practices and institutions, as Searle does. Instead, institutions are to be distinguished from social practices in terms of their normative dimension. The claim is that, in contrast to ordinary social practices, institutions involve social norms, which feature deontic powers. Thus, there can be institutions without constitutive rules. Relatively simple institutions will often depend on regulative rules instead. At the same time, constitutive rules are rather useful for capturing the symbolic dimension of institutions. Status concepts and terms promote the economy of thought and language. And status indicators help people to  An important disagreement between us concerns what it takes for a norm to govern a practice. According to Tuomela, this requires that the participants are collectively committed to the norm. I allow for a wider range of motivations, including for instance peer pressure. Importantly, their motivation can be low and need not induce compliance (Hindriks 2019, 2021). 20

146 

F. Hindriks

recognize statuses and enable them to converge on mutually beneficial courses of action. Thus, the status terms that feature in base and status rules serve important cognitive, linguistic, epistemic and practical purposes. Acknowledgements  I am very grateful to Francesco Guala, Arto Laitinen and Maj Tuomela, as well as the editors of this volume, for their comments on this chapter.

Appendix: Tuomela on the Transformation View Tuomela presented part of his criticism of the Transformation View in an unpublished comment that was intended to be included as an appendix to his paper ‘Social Corporations as Social Institutions’ (Tuomela 2020). I thank his wife, Maj Tuomela, for her kind permission to present it here in full.

 onstitutive Rules Are Not Reducible C to Regulative Ones Hindriks and Guala (2015), pp. 472–474, claim that constitutive rules are conservative (viz. non-creative) and a kind of eliminable transformations of regulative rules. But this contradicts at least my account because a constitutive rule in my account is an irreducible higher-order normative presupposition (a group-level ought-to-be rule, cf. Searle 2015). It says what the Y-term in the Searlean formula “X counts as Y in context C” is and also that X constitutes it. The constitutive rule is relatively stable in contrast to a regulative rule that primarily says how Y functions in an institutional context S. The constitutive rule, C, in the Hindriks-Guala (H-G) paper (p. 472) says that if a person is the first to occupy a piece of land, it is his property, and that he hence has the right to use it. This can be seen as a general constitutive and defining feature of property. In these authors’ view the

7  Can There Be Institutions Without Constitutive Rules? 

147

regulative rule (R) involved here is a conjunction saying in effect that if one is the first to occupy a piece of land then, in a quasi-Lockean sense, it is his property (corresponding to Y in Searlean constitutive norms); and if this is the case the resulting rule is claimed by the authors to entail that he has the right to use the piece of land (see Searle 2015, for criticism of this point). But R does not say how the institution of property can be used, so to speak, which is what a regulative rule should do at least according to my account. However, as will be shown below, in the H-G account C and R on the basis of their sameness of form have basically the same content. This implies that neither one is logically or conceptually primary, lacking a difference in the sense that I have claimed above concerning constitutive and regulative rules. I will below state the H-G regulative rule R in the abbreviated form “If Occ, then Right” (p. 472). Fully formulated we have this: “If Occ, then Prop, and hence Right”. The rule C has the form “If Occ, then Prop and if Prop, then Right”. The two conjuncts jointly entail “If Occ, then Right”, i.e. the rule R. R and C thus are equivalent in relation to their consequences and in this sense have the same content in these authors’ account, given the assumptions made above. (What the authors regard as a theoretical term, viz. here ‘property’—corresponding to the “theoretical” term Y in the Searlean formula “If X then Y in C”—does not get eliminated by non-theoretical terms; see Guala and Hindriks (2015) and below for this.) The central critical point against the H-G account as viewed from the point of view of my approach should be laid bare: the authors have in my view incorrectly mixed constitutive rules with regulative ones in their form-based treatment of the rules in question. However, at least in my account form need not distinguish them but rather their different functions. A constitutive rule should primarily say what something X is constituted as, while the regulative rule says how the constituted item or fact is supposed to be used or to function. Rules such as C and R qua, respectively, constitutive and regulative ones, are generally supposed to answer different questions and have different contents, which, however, the authors’ above simple logical formalism does not tell us but rather regards as equivalent. These authors’ rule C is indeed a constitutive one but their regulative rule R is also constitutive as to its form because having the right involves a right-involving status function entailing deontic power. Yet rule R does

148 

F. Hindriks

not show how to concretely use the institution of property as it should qua being a regulative rule. It can be concluded that the Hindriks-Guala proof does not succeed in eliminating constitutive rules in favor of regulative ones. What has above been said about the case of property applies mutatis mutandis to the institutions of money and marriage, etc.

References Belnap, N. 1993. On Rigorous Definitions. Philosophical Studies 72 (2–3): 115–146. Copp, D. 2020. Collective Obligations and the Point of Morality. In Routledge Handbook of Collective Responsibility, 93–112. Routledge. Elias, N. 1969. The Civilizing Process. Blackwell. Garcia, J.L.A. 1987. Constitutive Rules. Philosophia 17 (3): 251–270. Guala, F., and F. Hindriks. 2015. A Unified Social Ontology. The Philosophical Quarterly 65 (259): 177–201. Hart, H.L.A. 1961. The Concept of Law. Clarendon Press. Hindriks, F. 2005. Rules and Institutions: Essays in Meaning, Speech Acts and Social Ontology. Erasmus University Rotterdam. ———. 2009. Constitutive Rules, Language, and Ontology. Erkenntnis 71 (2): 253–275. ———. 2012. But Where Is the University? Dialectica 66 (1): 93–113. ———. 2013. The Location Problem in Social Ontology. Synthese 190 (3): 413–437. ———. 2015. Social Ontology: Collective Intentionality and Group Agents, Raimo Tuomela. Oxford University Press, 2013, xiv + 310 pages. Economics and Philosophy 31(2): 341–348. ———. 2017. Group Agents and Social Institutions. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with his Responses, ed. G. Preyer and G. Peter, 197–210. Springer. ———. 2019. Norms that Make a Difference: Social Practices and Institutions. Analyse & Kritik 41 (1): 125–146. ———. 2021. Rules, Equilibria and Virtual Control: How to Explain Persistence, Resilience and Fragility. Erkenntnis: 1–23. https://doi. org/10.1007/s10670-­021-­00406-­9. Hindriks, F., and F. Guala. 2015. Institutions, Rules, and Equilibria: A Unified Theory. Journal of Institutional Economics 11 (3): 459–480.

7  Can There Be Institutions Without Constitutive Rules? 

149

Lewis, D.K. 1983. Philosophical Papers Vol 1. New York: Oxford University Press. MacIntyre, A. 1981. After Virtue: A Study in Moral Theory. University of Notre Dame Press. Peacocke, C. 1992. A Study of Concepts. MIT Press. Ransdell, J. 1971. Constitutive Rules and Speech-Act Analysis. The Journal of Philosophy 68 (13): 385. https://doi.org/10.2307/2025037. Rawls, J. 1955. Two Concepts of Rules. The Philosophical Review 64 (1): 3–32. Raz, J. 1986. The Morality of Freedom. Clarendon Press. Ross, A. 1957. Tu-tu. Harvard Law Review 5: 812–825. Searle, J.R. 1969. Speech Acts. Cambridge University Press. ———. 1995. The Construction of Social Reality. The Free Press. ———. 2010. Making the Social World: The Structure of Human Civilization. Oxford University Press. ———. 2015. Status Functions and Institutional Facts: Reply to Hindriks and Guala. Journal of Institutional Economics 11 (3): 507–514. Tuomela, R. 1995. The Importance of Us. Stanford University Press. ———. 2002. The Philosophy of Social Practices: A Collective Acceptance View. Cambridge University Press. ———. 2003. Collective Acceptance, Social Institutions, and Social Reality. American Journal of Economics and Sociology 62 (1): 123–165. ———. 2013. Social Ontology: Collective Intentionality and Group Agents. Oxford University Press. ———. 2017. Response to Frank Hindriks. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. G. Preyer and G. Peter, 211–217. Springer. ———. 2020. Institutions in Action, The Nature and the Role of Institutions in the Real World. In Institutions in Action: The Nature and the Role of Institutions in the Real World, ed. Tiziana Andina and Petar Bojanić, 1–8. Springer. Warnock, G.J. 1971. The Object of Morality. Lethuen.

8 Institutional Proxy Agency: A We-Mode Approach Miguel Garcia-Godinez

1 Introduction Proxy agency is the capacity of an individual or group to act for another individual or group in specific social transactions. For example, an attorney defending her client’s rights in a court room, a company’s spokesperson voicing an official message to a targeted audience, and an MP raising their constituents’ concerns in Parliament are all cases of proxy agency. Yet, although a very common social phenomenon, proxy agency has not yet received enough philosophical treatment. A prime exception to this, though, is Ludwig (2014), who, relying on a deflationary, we-content approach to collective intentionality (à la Bratman), has developed a detailed account of this social capacity. While recognising the importance of his account, I argue in this chapter that Ludwig’s approach is too weak to explain proxy agency in institutional contexts, where individuals and groups formally authorised to perform rule-guided group activities for the principal hold (and are taken

M. Garcia-Godinez (*) Department of Philosophy, University College Cork, Cork, Ireland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_8

151

152 

M. Garcia-Godinez

to hold) a strong we-commitment to act. In order to elucidate this feature, I propose instead an account of proxy agency based on a more robust, we-mode approach to collective intentionality (à la Tuomela). Although the general purpose of this work is to frame the discussion of proxy agency within Tuomela’s sophisticated account of sociality, it also aims to motivate further research on this topic. In particular, concerning the value of proxy mechanisms for a thorough discussion of authorisation, participation, representation, and responsibility involved in legal and political contexts (e.g., the acting of legal organisations and private agents for the state).1 To present this alternative account of institutional proxy agency, I structure the chapter as follows. In §2 I introduce Ludwig’s account of proxy agency and explain in §3 the deflationary, we-content approach to collective intentionality upon which it is based. In §4 I elaborate on why such a we-content approach cannot explicate the strong we-commitment involved in the performance of institutional actions and outline in §5 an alternative view, viz., Tuomela’s theory of we-mode intentionality. In §6 I piece together a robust conception of institutional proxy agency based on this we-mode approach and explain how it fares better than Ludwig’s proposal. I summarise in §7.

2 Ludwig’s Account of Proxy Agency in Collective Action An individual or group acting through another individual or group is an exercise of proxy agency. Although relatively common, as Ludwig (2014, p. 76) recognises, all cases of proxy agency are puzzling. For one thing, it is not clear how an individual or collective agent can be attributed with an action performed by another individual or collective agent. While both individual and collective agencies have already received plenty of  Ludwig (2020b) has already argued for an understanding of legal organisations (the Legislative, Executive, and Judicial branches of Government) as proxies for the citizenry (see also Garcia-­ Godinez 2022), while Cordelli (2020) has recently discussed the possibility of private agents (e.g., civilians and corporations) acting on behalf of the state. 1

8  Institutional Proxy Agency: A We-Mode Approach 

153

attention from both philosophers of action and social ontologists,2 proxy agency remains largely unexplored. For bringing the discussion about proxy agency to the forefront, Ludwig’s analysis is a welcome and necessary addition. Besides presenting a very accessible understanding of how this social capacity operates, he provides a systematic account of its constitutive elements (viz., constitutive rules, status functions, and we-intentions). However, as I aim to show in the next few sections, his analysis is still defective in one important respect: It is based on a deflationary account of collective intentionality, which is too weak to explain the collective commitment that characterises institutional role holders. Before turning to this, though, let me introduce his proposal in some detail. Essentially, Ludwig’s account of proxy agency consists in answering the following questions: What is proxy agency? What makes an (individual or collective) agent a proxy? How does proxy agency work? And what is involved in proxy action? Although he does not explicitly say that he wants to answer any of these questions, his analysis is intended to help us get to grips with the functioning of the proxy mechanism—especially, in collective action contexts (2014, pp. 76–77). In discussing such contexts, his point is to argue for a ‘pluralist’ conception of proxy agency, according to which, when a proxy acts for a group, not only the proxy, but also the group (through all and only the group members) participates in its performance (2014, p. 78).3 As we shall soon see, this is meant to oppose the view that proxy action only involves the proxy’s action.

 Roughly, individual agency is the capacity of an individual to perform intentional actions (e.g., raising an arm, having a shower, or voting), whereas collective agency consists in the capacity of a group of individuals to perform intentional actions together (e.g., playing chess, dancing a tango, or having a conversation). For an overview on individual agency, see (Schlosser 2019), and on collective agency, see (Roth 2017). 3  I use ‘pluralist’ here to name the ‘multiple agents account of collective action’ that Ludwig develops in his (2016) and (2017a). I do not mean to say that he endorses a ‘plural subject theory’ of group agents. 2

154 

M. Garcia-Godinez

2.1 What Is Proxy Agency? “Proxy agency”, Ludwig says, “is a social phenomenon” (2014, p. 77). But more specifically, we can say that it is a social artefact, the purpose of which is to extend an (individual or collective) agent’s capacity to act by allowing them to act through another (individual or collective) agent. As such, then, as it happens with any other kind of artefact, there are various reasons for which an agent may need to use it (e.g., because it lacks the time, expertise, or material resources to act for itself ). To illustrate this, Ludwig considers the case of a spokesperson: “The person designated by the announcing group speaking at the appropriate time and place to the appropriate audience, acting under [a relevant] agreement between the announcing group and audience, constitutes the group’s conveying their message to the audience” (2014, p. 89). That is, by ‘designating a spokesperson’, the announcing group can convey a message through her speaking, thus performing a proxy assertion (i.e., a particular instance of a proxy action). As Ludwig explains: In proxy assertion one person or group (the principal) asserts something through another (the proxy) who speaks on the principal’s behalf […] When an individual or a group’s spokesperson speaks in her capacity as spokesperson, then the individual or group is credited, it seems, with a speech act, with having asserted, or explained, or questioned, or ordered something, and so on (2020a, p. 307).

Thus, as we can see, the rationale here is that when the principal acts through a proxy, it performs an action, though not by means of its own agency, but through the agency of the proxy. So, the special feature of the proxy mechanism is that the principal (not the proxy) owns, and should therefore be attributed with, the performance of the action. Though I shall elaborate on this when explaining how proxy agency operates, I shall consider first what it takes for an agent to be a proxy.

8  Institutional Proxy Agency: A We-Mode Approach 

155

2.2 What Makes An (individual or collective) Agent a Proxy? Acting through a proxy involves what Ludwig calls a ‘social transaction’ (2014, p. 77). That is, a social situation where the proxy is taken (by all the relevant participants) to be acting for the principal. For example, an attorney can act as my proxy only if there is a legal agreement between us conferring the appropriate power on her to act on my behalf before any interested party, or a company’s spokesperson can deliver an official message in the company’s name only if there is an official agreement between the announcing group and the spokesperson in virtue of which she is granted the appropriate power, etc.4 ‘Agreement’ here is an umbrella term, which is intended to capture the various ways in which a principal authorises the proxy to act on its behalf. As Ludwig has it, there cannot be proxy agency without authorisation (i.e., authorisation is a necessary condition for proxy agency) (see 2014, p. 77, 90). Since I say something more specific about authorisation below (see §6), it should be enough for now to say that not all forms of authorisation are explicit. Depending on the social situation, that is, a principal can informally authorise a proxy (e.g., simply by taking responsibility for the actions performed in her name). However, in institutional contexts, the relevant authorisation (by means of which an institutional proxy power is granted) cannot be informal. Although it would be too liberal to say that Ludwig has an account of authorisation, his understanding of proxy agency makes it clear that this is an essential component: It is the authorisation that gives a proxy the status role to function (or operate) in the relevant social transaction (2014, p. 77). For instance, in relation to the spokesperson case, Ludwig says that “authorization is essential to the group’s making an announcement  Following Leow (2019), we should notice, first, that there is a special kind of power at stake in all these situations, viz., proxy power; and second, that ‘acting for’, ‘acting on behalf of ’, and ‘acting in the name of ’ relations do not (necessarily) express proxy agency, which is better (if not only) captured by ‘acting through’ relations. Although Ludwig (2014, p. 103, n. 23) recognises an important difference between acting ‘in behalf of ’ and ‘on behalf of ’ (where the former, but not the latter, is taken to be in the benefit or in the interests of the principal), he uses the other three relations interchangeably. 4

156 

M. Garcia-Godinez

by way of what its spokesperson says” (2014: 90). The authorisation, then, establishes the spokesperson as a proxy for the announcing group: “The relation the person in question bears to the group is that of being authorized, as we say, to speak for it” (2014, p. 91). To be a proxy agent, then, amounts to being authorised by an (individual or collective) agent to perform actions for it. However, as Ludwig acknowledges, being authorised does not mean to act in the best interest of the authorising agent (2014, p. 103, n 23). A trustee may have an obligation to act in the best interest of the trustor, but the proxy relation is not that of a trust (or fiduciary) relation (see Leow 2019, p. 104). Moreover, a proxy can be authorised to perform a certain action, without the authorisation specifying in much detail how to do it. In the spokesperson situation, for example, Ludwig clarifies this by considering the ‘autonomy’ of the proxy: “The spokesperson, as for proxy agents in general, may be authorised within broad guidelines to represent her principal” (2020a, p. 318). Thus, although it would be incorrect to say, e.g., that the specific words used by the spokesperson were intended by the announcing group when conveying the message, it would also be wrong to say that the announcing group did not convey the message through the spokesperson. In the end, it all depends on the extent of the authorisation granted to the proxy (which can be more or less determined in the relevant description of her status role).

2.3 How Does Proxy Agency Work? Ludwig provides an explanation of proxy agency based on three structural elements: constitutive rules, status function (particularly, status role), and collective intentionality (2014: 77). He starts by making clear that proxy agency can operate both at the individual and group levels. For example, when I sign an estate purchase agreement through my attorney, she acts as a proxy for an individual; however, when a company’s spokesperson conveys a message, she acts as a proxy for a group. Yet, regardless of whether the proxy acts for an individual or a group, as Ludwig emphasises (idem), proxy agency is a capacity that necessarily requires a social context to operate. That is, there must be a social arrangement in place

8  Institutional Proxy Agency: A We-Mode Approach 

157

for an individual or a group to act through a proxy, viz., the kind of arrangement that makes the action of the proxy (partially) count as the action of the principal. Let me explain. Granted the appropriate authorisation, the principal can act through the agency of the proxy. This authorisation, as seen above, establishes the status role of proxy on the authorised (individual or collective) agent. To clarify, a status role is, in Ludwig’s terms, a special case of status function.5 Whereas “[a] status function is a property that an object, event, person, or group has in virtue of, roughly speaking, people so treating it, or being prepared to so treat it, and by virtue of which it plays a certain role in social transactions” (2014, p.  86), a status role is more specific: It can only be attached to an (individual or collective) agent. That is, “a status role is a status function for which being capable of agency is a necessary condition, and which requires the bearer to express her agency in the role in appropriate circumstances for the function to be performed” (2017a, p. 138). This is important for, at least, two reasons. First, to understand that proxies are agents, and so they cannot be conceptualised as mere instruments or tools. For example, one person can perform an action by means of a certain instrument (e.g., redact a message with a pen or communicate a message with a megaphone). However, this is not a case of proxy agency, since the instrument (a pen or a megaphone) is not an agent. Thus, putting metaphors aside, a proxy agent is not a principal’s instrument. Second, the fact that proxy agency requires the proxy’s agency makes it clear that the status role cannot be imposed on the agent. That is, unlike the status function of, e.g., a ten-dollar bill or a hammer, which can simply be assigned to an object, the status role of a proxy agent depends on the agent’s acceptance to participate in (or contribute towards) a certain action (idem). Without such an acceptance, the proxy cannot be taken to exercise the power(s) granted by the authorisation. So, although there follows an obligation on the proxy to act for the principal, if so agreed,  Noticeably, though, Ludwig has not always been clear about what he means by ‘status roles’. Sometimes he appears to equate status roles with status functions (2014, p. 86), and some other times he seems to recognise that they are two different things (2017a, p. 7, 138). As he is now more explicit about the specific nature of status roles, viz., that they are a special case of status functions (see 2020a, p. 308; 2020b, p. 181), this is the characterisation I follow in the text. 5

158 

M. Garcia-Godinez

this cannot be imposed on the former without her corresponding acceptance (which, except in institutional contexts, can very well be implicit). By holding the appropriate status role, the proxy has the capacity to make direct contributions towards the realisation of a constituted action, viz., an action determined by an essentially intentional (individual or collective) action type (2014, p. 78). For example, playing chess or dancing tango are the types of actions that (unlike walking or sleeping) require the intentional following of specific rules for an agent to instantiate them (e.g., moving a chess piece in such-and-such way or taking a step in such-­ and-­such way). These rules, says Ludwig, are nothing special; they are simply the kind of rules that are “constitutive relative to a type of behavior because that type of behavior is defined so that it occurs if the rules are followed intentionally […] and does not occur otherwise” (2014, p. 83). Such rules we call ‘constitutive rules’. Contra Searle, Ludwig submits “a deflationary account of constitutive rules” (idem), according to which they are not the kind of rules that “one has to follow in order to create a social or institutional fact” (idem). Instead, they are the kind of rules that constitute an action type, the instantiation of which admits of a two-layer analysis, viz., “an activity pattern which is describable as occurring in accordance with descriptive rules, and the requirement that it be instantiated intentionally, individually or collectively, depending on whether it involves one or more agents” (idem). So, by intentionally following the appropriate constitutive rules, an (individual or collective) agent can instantiate an essentially intentional (individual or collective) action type. Focusing only on essentially intentional collective action types (e.g., playing chess, playing tic-tac-toe, getting married, etc), Ludwig explains that they can be instantiated only when group members (i.e., the individuals engaging in producing the collective action as a group) execute successfully their corresponding we-­ intentions (i.e., their individual intentions addressed to bringing about a token of the collective action type) (2014, p. 79). As I look more closely at Ludwig’s notion of we-intentions below, I will not elaborate on it here. The important thing to highlight now is that all these elements together provide an explanation of proxy agency: A proxy agent participates in the instantiation of a constituted action by

8  Institutional Proxy Agency: A We-Mode Approach 

159

acting according to the power(s) associated to her role (as specified in the relevant authorisation) when intentionally following the descriptive rules that constitute the action type (e.g., signing a purchase agreement, filing bankruptcy papers, announcing a jury verdict, etc).

2.4 What Is Involved in Proxy Action? For Ludwig, the exercise of proxy agency does not entail that the action performed by the proxy fully constitutes the action attributed to the principal (2014, p. 92). That is, for example, when a group announces something through its proxy, the proxy’s action alone does not constitute the group’s announcement, but only its “most salient part” (idem). Let me unpack this. First, given that an agent can act through a proxy only if the latter has been authorised to act for the former, the authorisation is a constitutive part of the constituted action that both the principal and its proxy bring about.6 In other words, the authorisation in virtue of which an agent can hold the status role of ‘proxy’ is a constitutive element of the action that the principal performs through the proxy. To illustrate this, let us consider again Ludwig’s example of a group announcement: The basic idea is this: all members of the announcing group, in so conferring a status [role] on the spokesperson and thereby on some of her acts, official announcements, are agents of the announcements, though the spokesperson who takes care of the final transaction with another group is, as we might say, its executive agent (2014, p. 89).

With this, contrary to common assumptions, Ludwig wants to demonstrate that proxy action does not involve one, but two stages: In the first stage, the group designates someone to be its spokesperson and provides a message to repeat. To designate someone as a spokesperson is to give that person a certain status [role], given the background agreement between the group and its audience about how to treat that person […]  On partial constitution, see (Ludwig 2014, p. 85).

6

160 

M. Garcia-Godinez

The second stage is performed by the spokesperson doing his or her particular part in the complex action by uttering, e.g., at a press conference, a sentence that expresses the agreed upon message (2014, p. 91).

In the end, by arguing that proxy action involves both the principal’s authorisation and the proxy’s action as constitutive parts of the constituted action, Ludwig is in a position to say, firstly, that the proxy’s action only partially constitutes the whole action; and, secondly, that his account of proxy agency is consistent with a pluralist conception of collective action, according to which all members of the group (and not only the proxy) contribute to its realisation (2014, p. 100).

3 Conditional We-Intentions As suggested above, proxy agency operates both at the individual and group levels. Think, e.g., of my buying a house through an attorney, or a company announcing a message through a spokesperson, respectively. However, in any case, the exercise of proxy agency involves a ‘social transaction’, which Ludwig defines as an “instance of collective intentional behavior” (2014, p.  87). So, to understand proxy agency, we need an account of collective intentionality (i.e., the kind of intentionality that individuals enjoy when performing collective intentional actions, e.g., playing chess, dancing tango, or having a conversation). On this matter, though, Ludwig’s view is quite particular. His theory of proxy agency in collective action is built upon a deflationary conception of collective intentionality, which he calls ‘conditional we-­intentions’.7  Unfortunately, Ludwig does not say here (1) what he means by ‘deflationary’ and (2) what it is meant to be deflationary—indeed, he talks about various things being deflationary, e.g., a “deflationary view of collective action” (2014, p. 75), a “deflationary account of collective agency” (2014, p. 76), and a “deflationary account of constitutive rules” (2014, p. 83). Yet, what I think is deflationary in his project is the conception of collective intentionality (we-intentionality) that underlies his theory of collective action (see Ludwig 2007 and 2016, p. 236). In personal communication, Ludwig has clarified to me that his account is ‘conceptually’ and ‘ontically’ deflationary regarding both collective action and collective agency in the sense that “we can state necessary and sufficient conditions that draw only on concepts already in play in our understanding of individual agency” and that “we do not have to admit group agents per se into our ontology”. As for constitutive rules, his view is deflationary because it takes that “a constitutive 7

8  Institutional Proxy Agency: A We-Mode Approach 

161

Before introducing some criticisms to his view (see §4), I examine here such a conception and explain its role within a general framework of collective agency. According to Ludwig’s pluralist theory of collective action,8 when social groups perform constituted collective actions (i.e., actions governed by descriptive rules that define essentially intentional collective action types), all and only group members participate in its realisation by each intentionally following the corresponding descriptive rules. This, in Ludwig’s terms, means that the constituted collective action (e.g., playing chess or tic-tac-toe) is the “result of the individuals engaging in it having we-intentions to” bring about an action which can be correctly described by the rules defining the action type (e.g., ‘chess’ or ‘tic-tac-toe’) (2014, p. 80). As such, to fully grasp what is going on in constituted collective action, we need to spell out the participants’ we-intentions. For Ludwig, when it comes to collective actions where there are constitutive rules involved (i.e., rules defining an action type, e.g., ‘chess’, and specifying status roles, like ‘black player’, ‘white player’, ‘referee’, and status functions, like ‘chess board’, ‘pawn’, ‘rook’, etc) each participant holds a conditional we-­ intention, which he explains as “a conditional intention whose content is appropriate for a participation in joint intentional action” (2014, p. 88). More specifically, he says: “A conditional we-intention is a commitment to a contingency plan to do something on certain conditions obtaining” (2014, p.  87). For example, if we want to play chess, then we should abide ourselves by the rules of chess—because “[t]o play a game of chess, you have to participate in part by doing things that are constitutive of its play” (2013, p. 114). Yet, as he also clarifies, our we-intentions here are rule is just a rule specifying a form of acting like any other”, and so its “being constitutive is not a property of a rule”, but rather “a relation between a rule and an action type”. 8  His ‘pluralist’ theory of collective action has it that there is nothing special about collective action, except for the fact that more than one individual agent participates in its performance (2013, p.  129). When two individuals carry a table together, for instance, the sentence describing the action quantifies over two individuals, rather than a group or any other collective entity (2013, Sec. 5.5). Moreover, even when those individuals act based on their we-intentions, this does not imply that they constitute anything other than a mere plurality; that is, their we-intentions explain only their intentional participation in the collective action, but not the existence of any group agent (2016, p. 229).

162 

M. Garcia-Godinez

not about making the conditional true (which can happen simply by making the antecedent false), but about committing ourselves to following the chess rules if the condition of our wanting to play chess obtains (2014, p. 102, n 18). While interpreting Searle, Ludwig says that conditional we-intentions are in place when there is collective acceptance regarding the attribution of a certain status function (or role) to an object (or agent) (2014, pp. 86–87). For instance, if we want to play chess or football, we are each to have an intention to treat an object (e.g., a wooden board, a turf ground, etc) as having the relevant status function (viz., ‘chess board’, ‘football pitch’, etc). And all the same is true also of status roles. For example, being a football player (e.g., striker, goalie, etc) amounts to holding a status role, in virtue of which the agent can engage in the constituted collective action. When so engaging, the agent has an intention addressed to acting according to the appropriate constitutive rules, and so treating other objects and agents as having the relevant status functions and status roles, respectively. Now, in regard to proxy agency, something similar is the case. Here, all group members have conditional we-intentions towards treating an (individual or collective) agent as holding the role of ‘proxy’. Indeed, as Ludwig sustains (2014, p. 100), every group member participates in appointing (or authorising) the proxy (e.g., by voting, or taking part in the election committee, or merely agreeing on the appointment process, etc). Since all these participants (group members and proxies) are, according to this view, role holders, which are characterised by holding conditional we-intentions to contribute to the realisation of a collective action (according to and within the normative boundaries of their respective roles) by intentionally following the relevant constitutive rules, then we have here a uniform account of collective intentionality in constituted collective action contexts. This is, in fact, one of the main goals of Ludwig’s project (2014, p. 92). Nevertheless, as I presently argue, his understanding of conditional we-intentions is not robust enough to explain the strong we-­commitment that role holders (including proxy agents) have in institutional action contexts. However, this is not a mere inconvenience for Ludwig’s view as

8  Institutional Proxy Agency: A We-Mode Approach 

163

it challenges his claim that “[i]nstitutional agency ought to be continuous with informal collective intentional activity” (2016, p. x).

4 Strong We-Commitment in Institutional Action Contexts Although we are “pre-eminently social beings”, says Ludwig, “this is not what makes us special: what makes us special is that we make institutions” (2017a, p. vii). And as I have argued elsewhere, institutions, understood as institutional practices, consist in rule-governed group activities performed by formally organised social groups (see Garcia-Godinez 2020). Here, proxy agency is crucial, for, as Ludwig also acknowledges, it is not only a pervasive social phenomenon, but “a common instrument in institutional action” (2014, p. 76): When the Congress passes a Joint Resolution to declare war, the United States thereby declares war. When a corporation’s lawyers file bankruptcy papers, the corporation thereby declares bankruptcy. When a jury foreman announces the verdict at a trial, the jury thereby announces its decision (idem).

Overall, Ludwig’s goal is to give an account of (formal) institutional agency, including institutional proxy agency, that is continuous with his account of (informal) plural agency (2017a, p. 2). However, to achieve this, he must demonstrate that we do not need anything other than a deflationary account of collective intentionality to explain both kinds of agency. Yet, as I aim to show here, there are good reasons to believe that this cannot be so. For one, institutional action requires participants having we-intentions that entail a strong we-commitment to act, which is not (always) required in informal collective action. Let me elaborate. When analysing constituted collective action, Ludwig distinguishes, following Bratman (1979), between ‘conditional we-intentions’ and

164 

M. Garcia-Godinez

‘generalised conditional we-intentions’ (or policies) (2014, p.  88).9 A conditional we-intention is simply an intention to participate in joint intentional action, which involves attributing a status function (or role) to specific objects (or agents), upon certain conditions (e.g., to treat a turf ground as a football pitch if we are to play football, or to treat each other as chess players if we are to play chess).10 Policies, on the other hand, are conditional we-intentions to do something whenever certain circumstances occur. Since these intentions are not addressed to particular objects (or agents), but to types thereof, they help us “sustain the status functions of objects that are not currently being so used” (e.g., US currency) (idem). In general, claims Ludwig, all policies have ‘the same basic form’: if it is appropriate to engage in rule constituted action of type T, then such and such specific objects, or (alternatively) objects with such and such properties, are to play the role R specified in the rules for the action of type T (idem).11

Although in fact this formulation expresses a policy, viz., to treat certain objects (agents) in the way specified by the relevant constitutive  For some objections to Bratman’s view, especially regarding the ‘deconditionalisation’ of conditional we-intentions, see (Tuomela 2007, pp. 70-79). 10  Conditional we-intentions are just the plural form of conditional intentions. A conditional intention, as characterised by Ludwig, is the result of ‘contingency planning’, which creates a commitment to perform an action if a certain condition obtains (2015, p. 31). So, the obtaining of the condition gives the agent a sufficient reason either for (if ‘reason-providing’) or against (if just ‘enabling’) undertaking an action. Unlike conditional intentions, unconditional intentions are commitments (not) to act simpliciter (see Bratman: 1979). That is, an unconditional intention, for Ludwig, is an intention (not) to perform an action for (against) which the agent already has a sufficient reason (2015, p.  33). For example, my deciding ‘to leave the party if it is too loud’ is a conditional intention, the condition of which, if it obtains (i.e., if the party gets too loud), gives me enough reason to leave the party. Thus, by forming this intention, I am committed to leaving the party if it gets too loud (but if it does not get too loud, then I have no commitment to either leaving or not leaving the party). On the contrary, my deciding ‘to study at the library because I want to avoid distractions’ is an unconditional intention: I already have a reason for, and so a commitment to, going to study at the library. 11  For Ludwig, the collective acceptance of something (or someone) as having a certain status function (or role) for the purposes of a constitutive collective action, amounts to a convention—the aim of which is to solve a coordination problem (see 2017a, Sec. 9.1). Ultimately, for him, such conventions are the ground floor of social institutions. 9

8  Institutional Proxy Agency: A We-Mode Approach 

165

rules, it does not yet entail any commitment. It does not say, for example, that we are committed to treating certain objects (agents) as having a certain function (role), but only that ‘if it is appropriate to engage…’, then we are so committed. To be clear, though, the problem here is not (only) Ludwig’s lack of explanation regarding what counts as ‘appropriate to engage’, but (also) the fact that conditional we-intentions, including policies, do not entail themselves any relevant commitment to participate in (or make intentional contributions towards) a joint intentional action. Although in some cases this is just fine (e.g., when you and I want to play chess, and so form a conditional we-intention, according to which, if we want to play chess, then we should abide ourselves by the rules of chess, and treat some objects as having certain statuses, etc), it is not enough in institutional contexts (e.g., legal systems, political parties, international sport leagues, etc). If we each form a conditional we-intention to abide ourselves by the rules of chess if we want to play chess, for instance, our commitment to follow those rules derives only from both our fixing as a condition for such an intention our wanting to play chess and the fact that we do want to play chess. However, if we do not want to play chess anymore, there is no longer such a commitment (cf. Ludwig 2015, p. 44). The source of this commitment, in Ludwig’s words, ‘impinges on’ our own interests (2015, p. 31). So, we can release ourselves from our commitment if we do not want to play chess anymore. But in institutional contexts, things are significantly different. It is not up to us to release ourselves from the commitment we have to act according to specific rules.12 In fact, this is one of the main reasons we create institutions in the first place, viz., to impose stronger (or group-­ dependent) commitments onto the participants. A participant in an  I should note here that Ludwig does not believe that ‘things are significantly different’ in this respect (personal communication). Instead, he thinks that unless people have “more reason to continue in their roles than not”, they can always (and sometimes do) walk away. Yet, my point here is not that people cannot walk away, but that they hold a group-based commitment, qua institutional role holders, not to do so (even when they do it). It is because of this commitment that other participants can rationally plan their contributions to the institutional action without relying on each having appropriate conditional we-intentions (or not having more reason not to continue in their roles, etc). 12

166 

M. Garcia-Godinez

institutional practice (e.g., a judge or a lecturer) cannot decide not to perform a certain action (e.g., make a legal decision or teach a course) simply because it is no longer ‘appropriate to engage’ in it (assuming that ‘appropriate to engage’ has nothing to do with any alethic or ontic possibility). The kind of commitment involved in institutional action contexts, I shall now argue, is not entailed by individual participants’ (generalised) conditional we-intentions, but by their intentions to participate in we-­ mode group actions (i.e., actions performed by group members according to group reasons). To understand the contrast, let me first consider Ludwig’s discussion of role commitments, and then compare it, however briefly, to Tuomela’s collective commitment as entailed by his we-mode approach to collective intentionality.

5 Collective Intentionality and Group Reasons In his analysis of status roles, Ludwig examines the kind of commitment that follows from taking on a role (see 2017a, Sect. 10.2 & 10.3). As mentioned above, to participate in a constituted collective action, one must hold a status role associated with certain normative powers (particularly, rights and duties), based on which one can contribute to the realisation of the joint intentional action that constitutes, at least partially, the collective action. Since holding this role commits the role holder to participate in bringing about the constituted collective action, accepting the role is what gives the agent qua role holder a reason to act in a certain way (as required by the descriptive rules defining the action type). This acceptance, however, is not equivalent to promising to participate in the collective action (and so, does not engender any moral obligation to discharge the role responsibilities), but simply creates a commitment to act according to the role which is itself grounded in the antecedent internal (or motivational) reason the agent has for taking on the role (see

8  Institutional Proxy Agency: A We-Mode Approach 

167

Ludwig 2017a, pp.  144, 153).13 Let me clarify this before continuing with the discussion regarding role commitments. For Ludwig, it is accepting the role (e.g., of judge or lecturer) that grounds the commitment to act as a role holder (for all those roles which require acceptance) (2017a, p. 146). Yet, he wonders whether accepting the role implies having “appropriate conditional we-intentions to perform in that role” (idem). To answer this, he distinguishes between two relevant notions of acceptance. The first is formal acceptance—which just means accepting a role (e.g., by signing a contract or filling in a form) regardless of whether “one retains commitment to perform in accordance with it” (2017a, p.  147). The second is substantial acceptance—which essentially is “having an appropriate conditional commitment to perform in the role” (idem). Now, if one assumes, as Ludwig and I do, that the relevant acceptance in institutional contexts is formal acceptance, then this means that institutional role holders do not need to have ‘appropriate conditional we-­ intentions to perform in that role’. But, as seen above, since having such we-intentions is what guarantees participation in joint intentional action, then it follows that institutional role holders may not have any commitment to participate in the institutional action. This, I think, is questionable. Although Ludwig is right that the agent’s reason to fulfil her role responsibilities derives from the internal reason the agent has for occupying the role (2017a, p. 153), it is not true that “[i]f one loses one’s reasons for participating […] then one ceases to have any reason to fulfil the functions associated with the role” (idem). My point here is not that, by accepting the role, the agent acquires an underived (external) reason for action, but simply that the derived (internal) reason is actually a group reason, which commits the role holder qua role holder to functioning fully as a group member (according to the normative powers associated to

 An internal reason, says Ludwig, is a reason “grounded in one’s existing motivations” (2017a, p. 149). For example, if I want to avoid pain, I have a reason to avoid pain. On the contrary, an external reason (e.g., a legal rule imposing taxes on pain of legal sanctions) is independent of one’s motivations and so, it can only motivate when connected to an internal reason (e.g., that I want to avoid legal sanctions). 13

168 

M. Garcia-Godinez

the role). To see this, let me now introduce Tuomela’s famous we-mode notion of collective intentionality. In many places (e.g., 2002, 2003, 2005, 2007, 2008, 2012, 2013), Tuomela has developed a non-reductivist account of collective intentionality, making clear that there are various ways in which an agent can hold we-intentions in collective action contexts. Although his theory of collective intentional action does not differ (ontologically speaking) from Ludwig’s, as both think that collective intentional action supervenes on (but is not equivalent to) individual intentional actions, it does differ as to the various types of collective intentionality it takes to be involved in their performance. Since there is not enough space here to go into all the details of Tuomela’s account, I shall only focus on the elements that are necessary to explain the strong commitment entailed by the we-­intentions of institutional role holders. In the next section, I shall apply this to the general understanding of the collective intentionality involved in institutional proxy agency. Tuomela starts by distinguishing between I-mode and we-mode collective intentionality. The former, in short, amounts to individual agents thinking and acting for a private (though possibly shared) reason (2013, pp. 70–71). The latter, on the other hand, amounts to individual agents thinking and acting for a group reason (2013, pp. 67–68). If I intend that we play chess and you intend that we play chess, for example, we each have an I-mode we-intention addressed to a joint intentional action (viz., that we play chess). And this is so, even if we share a reason for our intending (e.g., if each of us wants to play chess with each other). However, by contrast, if I intend that our chess team, ‘the Chessy Club’, makes enough money to play the regional chess tournament, then I have a we-mode intention, the content of which is a group reason (viz., that the Chessy Club makes enough money to play the regional chess tournament). This, as Tuomela says, is not a reason shared amongst participants, but a reason collectively constructed for the group, based on which group members qua group members form their corresponding we-intentions to participate in the joint intentional action (2013, p. 68). The we-mode form of collective intentionality, thus, is irreducible to non-group properties (2007, p. vii). Since the group reason necessarily involves a group (e.g., the reason is attributed to the Chessy Club, rather

8  Institutional Proxy Agency: A We-Mode Approach 

169

than to each of the Chessy Club members), the group takes conceptual priority over the individuals (2007, p. 94). And this is possible because the individuals accept such a reason as their reason to act. In Tuomela’s words: We-mode acting is essentially based on group reasons, viz., on what the group decides, orders, or requires (etc.) and where the group members qua group members give some of their “natural” authority to the group (2008, p. 3).

This ‘natural authority’ is simply the collective acceptance of the group member to act as a group member, which in turn entails a collective commitment to perform the activities conducive to the satisfaction of the group reason (2013, pp. 43–46): “Indeed, the members should be collectively committed, viz., committed as group members, to participating in group activities” (2008, p. 5). So, getting back to the example, to the extent that I accept (as member of the Chessy Club) to act for this group reason, then I am committed to contributing towards bringing it about that the Chessy Club makes enough money to play a regional chess tournament (and the same goes for all other members of the Chessy Club). Now, as this commitment cannot be fulfilled without all group members participating accordingly, it implies a collectivity condition, which is the requirement that a group reason be satisfied for a group member if and only if it is satisfied for every other group member (2013, pp. 40–43). When in place, this condition guarantees both coordination and cooperation amongst group members (i.e., there is no other way for we-mode participants to perform a group action than coordinating and cooperating with each other). These three elements (group reason, collective commitment, and collectivity condition) constitute Tuomela’s core notion of we-mode collective intentionality, based on which he formulates his notion of functioning fully as a group member (2003, p. 109; 2008, p. 3; 2013, p. 6; and 2020, pp. 11–12). This notion applies to any group member who (explicitly or implicitly) is recognised by herself and other group members as having a collective commitment to act for a (motivating) group reason (2007, pp. 18–19). This recognition, Tuomela would emphasise, does not imply

170 

M. Garcia-Godinez

any substantial acceptance (in Ludwig’s terms) but only a formal acceptance of the group’s authority to create reasons (i.e., proposition-like entities) that group members can use in their practical reasoning (2007, p. 127). Although Tuomela’s we-mode approach to collective intentionality does not give any support to the claim that group agents exist over and above individual agents (2007, p. 124; and 2013, p. 21), it does provide both a conceptual and ontological analysis of a group action in terms of its individual members functioning properly qua group members, viz., “a person functioning as a group member, e.g., acting jointly with others or performing his institutional duties, should at least in general base his reasons for participatory acting on the group” (2013, p. 11). This analysis, in short, builds on the element of collective commitment. If group members collectively accept to act for a group reason, then they are collectively committed to acting according to that reason (i.e., to bring about an event that counts as the intentional satisfaction of its propositional content) (2013, p.  43).14 Their acceptance, though, especially in institutional contexts, need not be expressed in any other way than, e.g., signing a contract, filling in a form, paying a registration fee, etc. All these ways, in the end, constitute, at least in voluntary groups, the sufficient conditions for an individual (or group) to have an acceptance attitude towards the corresponding role (and the practical or normative implications that follow from it): The participants’ having signed up, and thus agreed, gives each participant a group reason for participating in the agreed-upon action. Furthermore, it also gives a reason for each participant normatively to expect that the other participants indeed will participate. Thus, a participant has the right to expect that the others will perform their parts and is also obliged to respect their analogous rights. In this sense they are normatively socially committed (2013, p. 132).

When the participants’ acceptance attitude is held in the we-mode, it grounds the attribution of a genuine participatory intention (which is  Collective commitment, for Tuomela, is what glues “the members together around an ethos” and gives “joint authority to the group members to pursue ethos-related action” (2007, p. 5). 14

8  Institutional Proxy Agency: A We-Mode Approach 

171

irreducible to, though not fully independent of, the private intention) to act as a group member. As Tuomela says: “In the case of a we-mode intention the important thing is not the specific content of the intention but the mode of having the intention” (2013, p.  67). That is, for example, unlike Ludwig’s view, what makes special a robust collective intention is not its content (e.g., I intend that we-J), but the mode in which it is held (e.g., I we-intend that we-J), which signals having adopted a particular attitude (i.e., a collective commitment) towards a group-dependent reason (2013, p. 106). As a group member in an institutional group (i.e., as an institutional role holder), one adopts a collective commitment to act for a group reason, the very formation of which requires the existence of the group, as an intentional agent (2013, p. 2). Thus, contra Ludwig, one cannot withdraw one’s participatory intentions simply because they no longer fit one’s interests. Although the commitment here is not moral or otherwise substantial, it is (at least partially) independent of the motivating reasons to join the group (i.e., to take on the role). So, while it may be true that one’s interests is the motivating reason to take up the role, once the role has been accepted, one adopts the group’s interests as the motivating reason to act qua institutional role holder. Hence, it is not under the role holder’s control to withdraw her participation in institutional action: Here, there is a collective commitment derived, not from a policy (generalised conditional we-intention), but from the adoption of a group reason.15 As an institutional role holder (e.g., judge, lecturer, bank teller, etc), I already have a reason to participate (coordinate and cooperate) in institutional action (e.g., making legal decisions, teaching a course, recording bank transactions, etc). Maybe I have a conditional intention to accept the role (e.g., I sign up if the salary is attractive, or if the job is not too demanding, etc), but once I have accepted the role, I adopt an unconditional intention to act as a role holder. So, my commitment to act according to the institutional role-description, as Tuomela would say, is not a  Note here, though, that having a group reason is not the same as, nor even entails, having an all-­ things-­considered reason for action. A group reason is, at best, a pro-tanto reason, the (moral, legal, etc) correctness of which must be determined by external factors (moral conventions, legal rules, etc). On this, see (Tuomela 2012 and 2013, p. 277, n 21). 15

172 

M. Garcia-Godinez

matter of contingency planning, but of group reasoning (i.e., what the group decides upon its decision-making procedures) (2013, p. 39, 125).16

6 Institutional Proxy Agency In institutional action contexts, proxy agents are (individual or collective) agents formally authorised to act for the principal. That is, they are authorised through a certain formal process of recognition to carry out the rule-guided activities (e.g., voicing a sentence, signing a document, or filing application forms) that constitute the final stage of the institutional action (e.g., making a legal decision, passing legislation, or hiring qualified personnel). These rule-guided activities, as Ludwig rightly sustains, are not all to the institutional action, but only the concluding or ‘executive’ parts of it (2014, p.  100). Since being authorised by the principal is a necessary, constitutive element of the institutional action, the principal is also a necessary, constitutive agent of it. This authorisation, as seen above, is the means through which the principal can grant the proxy power to an agent. So, if the agent accepts the role of ‘proxy’, then, according to the view sketched above, it becomes an institutional role holder and so, a we-­ mode we-intender. Although having we-mode we-intentions is not a property of all proxy agents (e.g., in informal situations, like a group of friends authorising a group member to organise a birthday party), it is of institutional proxy agents (where there is formal acceptance of fulfilling a role associated to a role-description that specifies institutional rights and duties of the role holder). This is not only consistent with Tuomela’s theory of sociality,17 but also follows from it. Whereas he recognises the variety of modes in which an individual agent can hold intentional states, including intentions to act (2013, p. 6), he affirms as well that we-mode is required for participation in institutional action (2013, pp.  7–8). So, to the extent  For a game-theoretic argument vindicating the preferability of we-mode we-reasoning over I-mode we-reasoning in group decision-making, see (Tuomela 2013: Ch 7 and Hakli et al. 2010). 17  For an overview, see (Preyer & Peter 2018). 16

8  Institutional Proxy Agency: A We-Mode Approach 

173

that institutions are “full-blown group phenomena” created (by groups rather than individuals) with the purpose of producing social order, we-­ mode we-attitudes are essential for their survival (2013, p. 225). As an institutional role holder, the institutional proxy agent does not need to worry much about bringing about an action that benefits the principal. For example, when a judge pardons a criminal, she need not do it because it benefits (even indirectly) the citizenry; or when a spokesperson conveys a discriminatory message, she need not think about how much this might affect the authorising company, etc. Not benefitting the principal, that is, does not imply that the action performed by the proxy cannot be correctly attributed to the former. In the end, all that matters is that the proxy acts as per the authorisation (i.e., according to the role-­ description associated to the role). By accepting the institutional role, nonetheless, the proxy agent accepts and so, is committed to, carrying out the activities conducive to the realisation of a certain institutional action (as determined by the corresponding institutional reason, e.g., to make a legal decision, to pass legislation, or to hire qualified personnel, etc). Whether these actions benefit the principal is an additional burden which is not (necessarily) within the proxy’s responsibilities. Not acting as per the authorisation, however, would result (at least in normal circumstances) in the attribution of a certain degree of shared responsibility: One part corresponding to the principal and the other, to the proxy. As I do not engage here with proxy responsibility, I leave this discussion for another time. What is worth mentioning, though, is, firstly, that proxy responsibility, as Ludwig also emphasises, is not the same as vicarious responsibility (2014, p. 103, n 26); and, secondly, that making a proxy responsible for performing a certain action in the name of the principal is grounded, not (only) in the existence of institutional rules (which can impose appropriate sanctions), but (also) in the collective commitment that she has (which necessarily implies coordination and cooperation) to perform in accordance with the status role. Although the proxy may resist acting for the principal (e.g., based on private considerations), this is not enough to cancel the (unconditional) commitment that she has to act qua institutional role holder. A commitment which, according to the we-mode approach to collective

174 

M. Garcia-Godinez

intentionality, is entailed by taking an institutional role and, thus, accepting to act for a (motivating) group reason. This, as Tuomela would maybe object to Ludwig, signals the irreducibility of (we-mode) institutional proxy agency to an (I-mode) pluralist account of shared agency.18

7 Conclusions In this chapter, I have presented (however schematically) Ludwig’s account of proxy agency. While recognising its importance, I have also argued that it faces a significant challenge, viz., given that it is founded on a deflationary approach to collective intentionality, it cannot explain the strong (group-dependent) commitment that institutional proxy agents hold to act for the principal. To overcome this, I have suggested considering a more robust, we-mode approach. Following Tuomela’s well-known theory of we-mode we-intentionality, I have sketched an alternative account, according to which, by accepting the role, an institutional proxy agent is formally recognised (by herself and other participants) as having a group reason to act, which in turn entails a collective commitment, the fulfilment of which requires the satisfaction of a collectivity condition (i.e., that every other participant also acts in accordance with the group reason). This, in the end, guarantees both coordination and cooperation amongst institutional participants. As mentioned at the outset of this chapter, I intend this proposal to be an extension of Tuomela’s theory of sociality. For, despite being a comprehensive account of social ontology, it does not include a precise view of proxy agency. The closest notion we can find in his large body of work is that of ‘operative members’ (2007, p. 11, 2013: 86). However, this notion is not about the same social phenomenon. First, operative members, unlike proxy agents, are not necessarily authorised to perform actions for the group (think, e.g., of the footballers that play in a match as opposed to those that sit on the bench, or the citizens of a nation-state that  Tuomela (2017) and Ludwig (2017b) have already discussed the irreducibility of we-mode to I-mode intentionality. It should be clear from what I have said here that I do not think the objections to Tuomela’s view are entirely correct; however, I leave to the reader a final assessment upon the merits of each position.

18

8  Institutional Proxy Agency: A We-Mode Approach 

175

exercise their right to vote as opposed to those who abstain from voting, etc). Second, proxy agents, unlike operative members, need not be group members (e.g., I can hire a lawyer to seal a contract for me, without this making us any kind of group; or a company can hire a spokesperson to deliver a message, without conferring any membership on her, etc). Thus, I believe that submitting this proposal is a necessary step towards completing Tuomela’s analysis of social reality. Acknowledgements  Many thanks to Kirk Ludwig for his comments on an earlier version of this chapter. Thanks also to the audiences at both the ISOS 2022 Conference and the Themes from Raimo Tuomela’s Social Ontology Workshop for helpful discussion.

References Bratman, Michael. 1979. Simple Intention. Philosophical Studies 36 (3): 245–259. Cordelli, Chiara. 2020. The Privatized State. New Jersey: Princeton University Press. Hakli, Raul, et al. 2010. Two Kinds of We-Reasoning. Economics and Philosophy 26: 291–320. Garcia-Godinez, Miguel. 2020. What Are Institutional Groups? In Social Ontology, Normativity and Law, ed. Miguel Garcia-Godinez, Rachael Mellin, and Raimo Tuomela, 39–62. Berlin: De Gruyter. ———. 2022. The Institutionalisation of the Basic Validity Rule. Law & Philosophy. Leow, Rachel. 2019. Understanding Agency: A Proxy Power Definition. Cambridge Law Journal 78 (1): 99–123. Ludwig, Kirk. 2007. Collective Intentional Behavior from the Standpoint of Semantics. Noûs 41 (3): 355–393. ———. 2013. The Ontology of Collective Action. In From Individual to Collective Intentionality, ed. Sara Chant, Frank Hindriks, and Gerard Preyer, 112–133. Oxford: Oxford University Press. ———. 2014. Proxy Agency in Collective Action. Noûs 48 (1): 75–105. ———. 2015. What are Conditional Intentions? Methode: Analytic Perspectives 4 (6): 30–60.

176 

M. Garcia-Godinez

———. 2016. From Individual to Plural Agency. New  York: Oxford University Press. ———. 2017a. From Plural to Institutional Agency. New  York: Oxford University Press. ———. 2017b. Methodological Individualism, The We-Mode, and Team Reasoning. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with his Responses, ed. Gerhard Preyer and Georg Peter, 3–18. Springer. ———. 2020a. Proxy Assertion. In The Oxford Handbook of Assertion, ed. Sanford Goldberg, 307–327. New York: Oxford University Press. ———. 2020b. The Social Construction of Legal Norms. In Social Ontology, Normativity and Law, ed. Miguel Garcia-Godinez, Rachael Mellin, and Raimo Tuomela, 179–207. Berlin: De Gruyter. Preyer, Gerhard, and Georg Peter. 2018. Raimo Tuomela’s Philosophy of Sociality. International Journal of Advances in Philosophy 2 (1): 1–14. Roth, Abraham. 2017. Shared Agency. In The Stanford Encyclopedia of Philosophy, ed. Edward Zalta, https://plato.stanford.edu/entries/shared-­agency/. Schlosser, Markus. 2019. Agency. In The Stanford Encyclopedia of Philosophy, ed. Edward Zalta https://plato.stanford.edu/entries/agency/. Tuomela, Raimo. 2002. The Philosophy of Social Practices. A Collective Acceptance View. Cambridge: Cambridge University Press. ———. 2003. The We-Mode and the I-Mode. In Socializing Metaphysics. The Nature of Social Reality, ed. Frederick Schmitt, 93–128. Maryland: Rowman & Littlefield Publishers. ———. 2005. We-Intentions Revisited. Philosophical Studies 125: 327–369. ———. 2007. The Philosophy of Sociality: The Shared Point of View. New York: Oxford University Press. ———. 2008. Collective Intentionality and Group Reasons. In Concepts of Sharedness. Essays on Collective Intentionality, ed. Hans Bernhard Schmid, Katinka Schulte-Ostermann, and Nikos Psarros, 3–19. Frankfurt: Ontos Verlag. ———. 2012. Group Reasons. Philosophical Issues 22 (1): 402–418. ———. 2013. Social Ontology: Collective Intentionality and Group Agents. New York: Oxford University Press. ———. 2017. Response to Kirk Ludwig. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with his Responses, ed. Gerhard Preyer and Georg Peter, 19–35. Springer.

9 From We-Mode to Role-Mode Michael Schmitz

1 Introduction Nobody has done more to develop a we-mode account of collective intentionality than Raimo Tuomela (e.g., Tuomela 1995, 2007, 2013). In this paper, I want to pay tribute to Raimo by extending the mode-account of collective intentionality from we-mode to role-mode—the mode in which individuals and groups function in roles in organizations and institutions and take positions as occupants of these roles, as when, for example, an official announces economic data, a committee advises a certain course of action, or a parliament passes a law. The account I will sketch is inspired by Tuomela’s account of the we-mode, and as he noted (2017, 77), there are also similarities with what he says about “normatively structured groups with positions and a division of tasks and rights” and the “positional mode” in his 2013 book Social ontology: Collective Intentionality and group agents.

M. Schmitz (*) University of Vienna, Vienna, Austria e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_9

177

178 

M. Schmitz

I prefer to use the term “role” though because I use “position” to refer to the kind of theoretical or practical position associated with mode/force in the sense of what distinguishes beliefs and intentions and assertions and directions. I call we-mode and role-mode “subject modes” to distinguish them from mode/force in the more established sense, which I call “attitude mode” or “position mode” (Schmitz 2017, 2018). While my approach is inspired by Raimo and I have learned much from him, my outlook also differs from his in some important ways that I should mention here at the outset—to set the stage and to help avoid misunderstandings. The first of these consists in the fact that—as Raimo noted in what was sadly to remain our only exchange in print (2017, 74, 76)—I’m much more focused than he was on the phenomenology of the we-mode, on the we-mode as a mode of consciousness. I’m therefore also happy to embrace the terminological kinship to the use of “mode” in consciousness studies to refer to global states of consciousness such as depression or even wakefulness. Though we-mode and role-mode states are of course more specific states, they still pervade consciousness in a way that is comparable to these states. For example, the experience of acting in one’s professional role is often described in terms of being—and having to be—“on”: a peculiar state of heightened alertness and attention. A second respect in which my outlook differs from Tuomela’s is that I attach more importance to more basic forms of intentionality and consciousness such as sensory-motor-emotional consciousness. For example, I believe that fellow feelings of affiliation, alignment, empathy and identification as well as experiences of joint attention and joint action are absolutely central for understanding collective intentionality. A third and related respect is that I think we should think about intentionality in general and collective intentionality in particular in terms of a layered picture according to which higher-level forms of collective intentionality such as role intentionality in institutional and organizational contexts can only function against the lower-level collective intentionality of shared beliefs and intentions as well as that of joint attention and joint bodily action. An important part of this picture is also the idea that the intentionality on these different levels can be distinguished in terms of the kinds of representations and the kinds of content involved.

9  From We-Mode to Role-Mode 

179

(I don’t mean to suggest that Tuomela would have disagreed with all of this. He did, for example, accept a role for “bond-creating non-­conceptual elements” in his response to me (2017, 76). But he neither centered nonconceptual content in his own work nor layers.) Fourth, and finally, one of the reasons for embracing the notion of role-mode for me is that it can help us to leave behind the fictionalist elements Tuomela felt compelled to introduce into his account of the intentionality of group agents such as organizations and institutions. On Tuomela’s picture, such a group agent is “partly fictitious” (2017, 71; his italics) because it has “no intrinsic, biologically based features”, but only extrinsic intentionality, which, however, may be “internalized by the agents (in a sense resembling actors internalizing role attitudes in a play)” (all 2017, 71). There are certainly important similarities and connections between the capacities to play a role in a fictional context and to fill an institutional role such as being a government official. However, there is also a difference which is as important as it is elementary. The official does not merely pretend to make determinations and decisions in his role, but actually makes them. It therefore seems to me that it is important to develop an account of group agents such as corporations or governments that does not just treat their intentionality as merely fictitious and extrinsic. Identifying with our functions e.g., as government officials and as citizens and accepting the powers and obligations of officials in our role as citizens (to the extent that we do!) and seeing the world from the perspective of these roles in role-mode, we really become officials and citizens. Taking positions, perceiving, acting, believing, planning, asserting and directing in these roles, we display a role intentionality that is not merely extrinsic or derived. Given these differences between Raimo’s outlook and mine, I was very moved that one of the last times we met, he said to me after a talk of mine: “I think you are on the right track”. I often think back to this kind comment, especially in difficult times, and feel encouraged by it to present this account in the context of this volume. I will now first motivate the mode account by comparing it to its principal rivals. I will then sketch its subject mode version, introduce the

180 

M. Schmitz

notion of role-mode, and explore it further by discussing 10 theses about it and its relation to we-mode.

2 Content vs. Mode vs. Subject Approaches to Collective Intentionality The philosophy of collective intentionality in the twentieth century has been haunted by the specter of the group mind. It has seemed to many that to take groups seriously ontologically and particularly to take them seriously as agents and as subjects of mental states would be to commit to that “abominable” and “perfectly dreadful metaphysical excrescence” (Searle 1998, 150). To avoid this has been a motivation not only for the most strongly reductionist view of collective intentionality, the content view most closely associated with Michael Bratman (2014) and Kirk Ludwig (2016, 2017), which assumes that at least small-scale collective intentionality can be entirely understood in terms of the content of individual mental states and speech acts—where content is understood in the traditional sense as propositional and as what is intended or believed, in contradistinction to the subject and the mode. It has also been a motivation for the mode account as defended by Tuomela1 and also by John Searle—though Searle never uses the term “mode” in this connection, but only speaks of we-intentionality. The attraction of the mode account is precisely that it seems to allow us to adopt a more liberal stance regarding the kind of mind involved in collective intentionality than the content account, while still avoiding the notion of a group subject which is taken to be “abominable” because, given certain plausible-seeming assumptions, it is tied to that of a group mind.  However, in his response to me he writes: “In my view, the subject of an attitude or action to which the mode is attributed accordingly may be either an individual or a collective (or a collection of people). Thus my we-mode notion can be regarded as primarily a subject account” (2017, 74). I may thus have misrepresented what Tuomela wrote earlier, but I am not entirely convinced for the reasons that I stated in my original piece and don’t want to repeat here, as I focus on developing the positive picture. And though we should not give too much importance to such labels, others (e.g. Schweikard and Schmid (2013)) have also taken Tuomela to be the paradigmatic representative of a mode account taken to be distinct from a subject account. 1

9  From We-Mode to Role-Mode 

181

This, I believe, is at least part of what compelled Tuomela to think of the intentionality of a group agent only as a fiction entertained by the members of the group. But, as noted already, the determinations and decisions of a university, a corporation, or an association like ISOS, the International Social Ontology Society, are not merely fictitious. Nor is the intentionality of the people making these determinations and decisions merely extrinsic. The thoughts they have in their roles are just as intrinsically intentional (and as biological) as the thoughts they have as private persons. A fortiori, corporations, associations and organizations are certainly not fictitious either. It is true, the law has often used the legal fiction that corporations are persons to make it possible to hold them legally responsible. But this does not mean that the corporation is a fiction. The fiction here is that the corporation is a person. I think it is quite revealing that the law would choose this conceptualization: the individual person is so much our paradigm of a locus of agency and responsibility that lawyers find the fiction that the corporation is a person the most expedient way of integrating it into the law. Now a philosopher pondering this idea of a corporation conceived of as a person is very likely to be puzzled, even perturbed, by this somewhat in the way that Wittgenstein imagines philosophers to be puzzled by his fictitious example of a language where instead of “Nobody came into the room” one says something like “Mr. Nobody came into the room” (1958, 69). (As Wittgenstein showed so convincingly, when we philosophize, we tend to operate with a much cruder and more primitive picture of what it would mean for expressions of our language—and nouns in particular—to succeed in referring to or representing something in the world than the one that is inherent in our actual linguistic practice.) Two characteristic reactions of philosophers in such a situation would be to either try to make sense of the existence of such a person because they feel—correctly—that this expression serves an essential function in our fictitious language, or conversely, to explain away the apparent existence of such a person by giving a fictionalist interpretation of this kind of locution.

182 

M. Schmitz

I think that a similar set of responses can be found in debates about groups and collective intentionality. Philosophers may be tempted to make sense of the notion of a group mind, or they may try to do such things as eliminating all references to “we”—at least when it occurs as the subject of attitudes like intention—or to give a fictionalist interpretation of the notion of a group agent because they are captivated by a mistaken idea of what it would mean for such a subject to really exist. What is that idea? It’s hard to even express it except by saying something like that the group would have to be just like another person or that its mind would have to be like the mind of another person, something over and above the individual minds of the group members. Such an idea is indeed an abomination, but does anybody really believe it? This philosopher wants to suggest that we should not leave the idea of a group subject to mysterians trying to make sense of this idea—assuming they even exist—and should instead rather take seriously the notion of a plural subject. A group no more has a single mind than it has a single body. It rather has minds—and bodies.2 What we have to understand is what it is about these minds—and bodies—and the persons or animals whose minds and bodies they are that ties them together as group members. We should think of a group as a higher-level unity, a unity of people (or animals), of minds and bodies. We are always tempted to think of a group in such a way that the difference between individuals is dissolved, that they disappear into the group as it were. (Strikingly, evocations of this fantasy are also a hallmark of totalitarian collectivist ideologies.) But if this were to happen, the group would cease to exist. The group is essentially a higher-level unity of individuals. To put this into a slogan: no we without Is. So what we have to understand are the modes of mind, or, more broadly, the forms of group mindedness, that tie individuals together so that they form higher-level plural subjects. This is not meant to suggest though that the connections between group members are only mental or spiritual. A group is not a communion of disembodied spirits, but of flesh-and-blood individuals. At its  Tuomela wrote: “a group agent necessarily lacks an intrinsic mind qua group” (2017, 72). Certainly, but my question is: why should we even expect it to have one intrinsic mind? 2

9  From We-Mode to Role-Mode 

183

most basic, it is typically held together through physical interaction, through touch, joint attention, joint bodily actions, joint skills, feelings of affiliation and identification, and so on. The emphasis on mind is only meant to indicate that all these interactions and relations have an essential mental component, not that they are purely mental. The mind-body dualism that we are accustomed to does us a disservice here and makes it hard to understand how we relate to others at the most basic, sensory-motor-emotional level, because it is inadequate either to say that we relate to others merely as bodies, nor only as minds. It seems to me that there is a way of understanding others that is prior to the mind-body differentiation as when e.g., we perceive somebody as doing something. We perceive the movement as purposive, as the kind of action that it is, but this does not mean that we separately perceive bodily movement and purposes or intentions. Such a differentiation or analysis of the action into physical and mental components is only made at a higher, reflective and conceptual layer of intentionality. Still, the focus of this paper will be on the mental component of collective intentionality. But it is important to keep in mind that, like other forms of intentionality, collective intentionality is essentially embodied and enacted. I propose to think of modes like we-mode and role-mode as explaining collective subjects rather than as explaining them away. This is one reason I call the account that I will present in the next section the “subject mode” account of collective intentionality: it indicates that I attempt to marry a mode account in the tradition of Tuomela and John Searle with a subject account as it has been defended by Margaret Gilbert (2013) and Bernhard Schmid (2014) among others. One important difference between the subject mode account and Gilbert’s and Schmid’s subject accounts is that I acknowledge a greater variety of collective subject-constituting entities. For Gilbert, it is joint commitment alone that constitutes collective subjects, while for Schmid it is plural pre-reflective self-awareness. I believe that the basic sort of collectivity constituting intentionality is pre-reflective and pre-conceptual— the sensory-motor-emotional intentionality of joint attention, action and affiliation—but that against the background of this lower-level intentionality, the higher-level intentionality of joint commitments, of shared beliefs, intentions and values, as well as that of contracts, laws and other

184 

M. Schmitz

forms of institutional intentionality, creates corresponding higher-level unities from plural subjects of a commitment to take a walk together to corporations and countries. Now why can’t we account for all this collectivity-constituting intentionality in terms of propositional content as traditionally conceived, as proposed by the content account? I can’t argue this point at length here, but I believe that there are two related reasons for this. The first is that “we” and its cognates, at least in their basic uses, on which others depend, directly latch onto pre-conceptual, sensory-motor-emotional forms of collective intentionality. In the central cases, I use “we” to refer to us as related in one or several such pre-conceptual ways—e.g. as standing or walking together and/or attending to something jointly—not by first forming a conception of these relations (in terms of which it could then be analyzed), but directly, similarly to how it has been argued that uses of “I” are also based on pre-conceptual forms of self-consciousness (Bermúdez 1998, 2011). The second reason is that using “we” also centrally involves the capacity to see the world from our shared perspective and therefore us as a subject with a (partially) shared point of view. That is the reason we-­ intentionality cannot be reduced to what I-subjects intend, believe, or otherwise represent as an object of their intentionality. This brings me to the second reason why the “subject mode” account is so-called and a crucial respect in which it differs from the standard versions of the subject-, mode- and content-accounts (see Schweikard and Schmid 2013). These all accept the received understanding of intentional states and speech acts—postures—as propositional attitudes, which I believe makes it hard to understand how subjects can experience and represent themselves as subjects of attitudes towards facts and goals in the world, including as collective subjects who share goals or are jointly aware of facts. The basic idea of the subject mode account is that such awareness of collective subjects as subjects can be understood in terms of representing others as co-subjects of postures towards the world, ranging from joint attention and action over shared beliefs and intentions to the positions we take as occupants of roles in institutional contexts. If subjects mutually represent each other as subjects of such postures, various kinds of groups ranging from joint attention dyads over subjects of shared

9  From We-Mode to Role-Mode 

185

beliefs, plans and values, to corporations and other institutional, legal units, come into existence.

3 The Subject Mode Account To pave the way for the subject mode account, I need to at least briefly address the received understanding of propositions and propositional attitudes, which stands in its way and, in spite of some recent criticisms, still has a dominant role in contemporary thinking about mind and language. It embodies three biases in the philosophy of mind and language: 1. For propositional and conceptual forms of representation over nonpropositional and nonconceptual ones 2. For theoretical, mind-to-world direction of fit representation of the world over practical, world-to-mind direction of fit representation 3. For individual over collective intentionality First, on the received understanding of intentional states as propositional attitudes, all content is propositional and conceptual. But the content of basic individual and collective intentionality—the sensory-motor-emotional intentionality of joint attention, action and affiliation—is nonpropositional and nonconceptual. Second, the content of a posture is taken to be identical to its propositional content. That is, neither subject nor mode/force makes a contribution to content on this model. For example, when we believe that it rains, neither we nor our believing as such makes a contribution to content—though of course they would be represented in the content of a report of this attitude. And the state of affairs that the individual or collective subject is directed at is always implicitly taken as represented from a theoretical position towards the world—as truth is representational success from a theoretical position towards the world—regardless of whether the attitude is theoretical like belief or practical like intention. That is, even a practical attitude like intention is taken to have a proposition as its content, in spite of the fact that ordinarily we only ascribe truth values to theoretical attitudes like

186 

M. Schmitz

belief. Third, as we have discussed already, collective subjects are often seen as an abomination, and the default assumption is always that the subject is an individual. The alternative picture I want to present is inspired by recent criticisms of the notion of a proposition and the force-content dichotomy to which this notion is closely linked (Hanks 2015, 2019; Recanati 2019, 2022), as well as by a tradition of thinkers ranging from Immanuel Kant, over Jean Piaget, Peter Strawson and Gareth Evans to contemporary thinkers like José Luis Bermúdez that treat self- and object-consciousness as inextricably linked, as two sides of the same coin. On this kind of view, we are always aware of objects and states of affairs (SOAs) in the world in relation to ourselves. This is true for a multitude of such relations, e.g., spatial or temporal ones, as well as the theoretical, epistemic ones and practical ones that are manifest in belief and intention as well as in theoretical and practical knowledge. And, again, it is true on various layers or levels of intentionality which are characterized through the different modes and formats of representation that are involved. These layers can be distinguished in a fine-grained way through various criteria (Schmitz 2020), but for purposes of rough orientation it is sufficient to distinguish three broad layers: 1. The preconceptual and nonpropositional layer of perceptual, actional and emotional experience. This layer corresponds to sensory-motor-­ emotional experience. 2. The conceptual and propositional layer of belief, intention, theoretical and practical knowledge. This layer corresponds to spoken language. 3. The documental level of shared contracts, constitutions, legal frameworks, cultural heritage etc. This layer corresponds to written language. On all these levels subjects experience and represent their position towards the world. In perception they experience the world as acting on them, in action they experience themselves as acting on the world. In asserting something like that the door is closed the subject presents itself as knowing this. In directing somebody to close it, a subject presents itself as knowing this as the thing to do, as possessing knowledge of what to do. Basic force indicators such as intonation, word order and grammatical

9  From We-Mode to Role-Mode 

187

mood nonconceptually indicate this theoretical or practical position towards the world and express a sense of this position (Schmitz 2022). This also illustrates how the different levels of intentionality work together: when a subject asserts that the door is closed by means of uttering a sentence like “The door is closed”, it indicates its position nonconceptually, while the SOA it presents itself as knowing is represented conceptually. The subject of the posture does not only have a sense of its position and presents itself as occupying this position, but also experiences and presents itself as the subject of this position. In our example, this also happens nonconceptually. The subject has a nonconceptual actional experience of performing a speech act and also typically nonconceptually experiences itself bodily and in yet other ways as well. Contrast this with a case where a subject performs an assertion through a performative use of a speech act verb as in “I assert that the door is closed”. In this case, the subject conceptually represents itself and the position it takes. The content representing the relevant SOA can be called “SOA-­ content”, that representing the subject’s position “position-content” and that representing the subject “subject-content”. Now the basic idea of the subject mode account is that there is a difference between representing others as mere objects of one’s attitudes and as co-subjects of such attitudes. In representing others as co-subjects of positions towards the world, we share a perspective on the world and see it from the perspective of the group. Seeing the world from a we-­ perspective expands the subject slot of individual intentionality, as it were. The subject now takes a position in a mode of identification with others. Again, this happens on different levels of collective intentionality, as when one attends and acts with others and jointly commits to plans and beliefs with them. And again, the different levels function together. For example, think about congregants in a church saying a prayer or singing a song together. They nonconceptually experience themselves as singing or praying together and thus as jointly affirming what they express even if conceptually it is an SOA I-content like e.g., “I am a sinner”. This content is still nonconceptually experienced and presented as something that is known and known jointly.

188 

M. Schmitz

On the traditional view of propositions and propositional attitudes there is a strong tendency to interpret sentences like the following so that subject and speech act verb are part of propositional content (as traditionally conceived): (1) We assert that no nation can long endure half republic and half empire. (2) We order you to learn Dutch.

Now, it is true that such interpretations are possible if we think of them as answers to the question of what we would or should do under certain circumstances. For example, (1)—part of the of 1900 platform of the US Democratic Party—might be uttered in a group or individual deliberation on how to respond to any foreign policy initiatives that smack of imperialism (for a similar example, see Green (2022, 211 fn 5)). Under such an interpretation, that we assert or order may indeed be the SOA-content of something that a subject asserts or advises. (Whether it is an assertion or a piece of advice, and who the subject of that advice is may again be determined nonconceptually and/or through the broader context.) However, this is certainly not the most common or even a mandatory interpretation. The most common and most natural uses of sentences like (1) and (2) are the “performative” ones, where the speech act verbs are actually used to indicate the speech act being performed and the first-­ person pronoun indicates the subject of the speech act. Attempts to in turn derive this performative interpretation from a truth-conditional reading of sentences of this kind have to my mind not been successful, nor are they well-motivated (compare (Schmitz 2019, 119–120) for more discussion). I think we should just accept that such sentences are ambiguous and that the most natural reading is the one where the speech act verbs function as force indicators. On the subject mode account, groups are constituted through the intentionality of the group members, through their group mindedness. And this mindedness is (at least also) representational, it consists in experiencing and representing others as co-subjects of positions towards the world. Though this has often been denied by philosophers of collective intentionality, including by Tuomela, I think this also means that

9  From We-Mode to Role-Mode 

189

misrepresentation is possible. I might be in a we-mode state of jointly intending with others, who have, unbeknownst to me, already abandoned what had been our shared plan. In this case, I misrepresent a subject of joint intention that does not really exist. My state is in one sense still an instance of collective intentionality because it is an instance of the kind of group mindedness through which collectives are constituted. In another sense it isn’t, because it is not the intentionality of a group. I believe both senses have to be recognized because the group mindedness of individuals constitutes the intentionality of groups. But it is tempting to ask at this point: who is really the subject of collective intentionality, the individuals or the groups of which they are members? It seems to me the only possible answer is that the individuals are jointly the subject. They jointly constitute it by mutually representing themselves as co-subjects of the positions the group takes. So, fittingly, the task of representing the group’s position is also a shared one.

4 Introducing Role-Mode We are now in a position to begin addressing how the notion of we-mode can and should be extended to include role-mode as another subject mode. Let me begin by discussing the relation between roles and identities. Identities may seem to be more pervasive and more deeply anchored than roles. For example, being a man or a woman is a deeper and more pervasive aspect of one’s being than one’s professional role. One is always, or almost always, perceived as belonging to a gender, but not nearly so often as having a professional role. At the same time, gender is both conceived of as an identity and as a role. But when we think of it as an identity, we think of it more as coming from an internal sense of what a person is, while if we think of it as a role, we think of it more as something that comes from the social environment of the person. Many definitions of roles highlight their social character. For example, Merriam-Webster’s online dictionary defines one meaning of “role” as “a socially expected behavior pattern usually determined by an individual’s status in a particular society”. Similarly, philosophers have often emphasized an understanding of role as a social status assigned and imposed

190 

M. Schmitz

from outside. But it seems we also need to take into account subjects’ identification with them. Individuals do not just passively accept the roles others assign to them. They also strive for roles and even create them for themselves. And they develop their own interpretations of roles assigned to them and adapt them to their own preferences and purposes. I think we therefore cannot understand roles just as conferred from the outside. We also have to understand the subject that identifies with them and that takes position towards the world as a role bearer. Roles therefore cannot be distinguished from identities purely on the basis of an internal/external opposition and the distinction in terms of pervasiveness, while suggestive, is certainly not sharp either. And, as noted, gender is conceptualized as a role as well as an identity, and this is also true of other relevant concepts. Luckily, for present purposes it is not really important to sharpen the distinction between identities and roles. In what follows I will focus on professional and other institutional roles though, since I am primarily interested in the role subjects have in fairly tightly organized and quite strongly cooperative contexts, the kind of contexts Tuomela also had in mind when thinking about group agents. While we-mode is more egalitarian—without claiming that completely egalitarian relations ever exist in human groups—roles have to do with power and competence differentiation within groups and therefore also with specialization. A role is defined in relation to other group members, but also in relation to the world. A role bearer has a sphere of power, of influence, responsibility and—at least hopefully—of competence. That is, role has not only to do with what I am proposing to understand in terms of subject mode, but also in terms of position mode. The role bearer will have particular actional and perceptual skills and particular theoretical and practical knowledge regarding the role domain. Role-mode is a mode of identification with the role, with the group in the context of which a role is defined, with its powers and obligations, and with the practical and theoretical skills and the domains of knowledge it requires. Of central importance for role-mode are the theoretical and practical positions subjects take in their roles and the powers and obligations they have as role bearers. Canonical linguistic expressions of role intentionality are therefore “As [role] …” and “In my role as [role]…” as in examples like the following:

9  From We-Mode to Role-Mode 

191

(1) As chancellor, I believe… (2) As your supervisor, I order you… (3) As certified yoga instructors, we know… (4) As members of the committee, we have decided… (5) In my role as your student, I have the right…

Of particular significance is the fact, which we will return to, that people sometimes take different positions as role bearers than they would privately or in different roles. Before I begin characterizing role-mode through a series of theses, I want to consider a question which has been raised by Bernhard Schmid (2017) primarily with regard to the we-mode, but which a fortiori can also be raised regarding role-mode: why should we think of these as species of mode at all? What is the justification for calling both subject-­ modes and position-modes “modes”? It seems to me it is natural to extend the concept of mode in this way. It is natural to distinguish position-mode from SOA-content because the same SOA can be represented from different positions. For example, I may direct somebody to close the door and then go on to assert that this somebody has closed it. Analogously, I may take up the same position in different modes. For example, I may intend to support the election of a candidate both as a private person and as a member of the same party as the candidate. But, as just noted, one can also take different positions in different roles, as when one supports a candidate as a party official, but not as a private person, or conversely. The deeper motivation for Schmid’s question is of course the traditional understanding of speech acts and propositional attitudes, where subject, mode/force and propositional content are taken to be three fundamentally different constituents or aspects of these postures. The present approach does try to give a more unified account. Force/position-mode and content as reconceptualized as SOA-content are still different, but not as different as in their earlier conceptualization, as now content is also ascribed to force/position-mode. Both individual and collective subjects also remain distinct from subject mode, but they are still importantly connected, as a subject can be transformed into a group member by

192 

M. Schmitz

experiencing and understanding itself in relation to co-subjects and by being in turn experienced and understood by these co-subjects. From this perspective, position-mode and subject-mode naturally belong together because they are both forms of self-consciousness. Subjects take positions in self-aware ways and in modes of identification with others and/or with their roles, experiencing and representing themselves as co-subjects of such positions.

5 10 Theses About Role-Mode and We-Mode I will now explore role-mode further by discussing 10 theses about role-­ mode and its relation to we-mode. 1. We-mode and role-mode are further modifications of self- or I-consciousness Just like there is no we without Is, there cannot be a we-consciousness that is not a modification (or a mode-ification if you will) of an I-consciousness. It must always be possible to expand the “we” as e.g., “Maria and Tony and I” or “the other Germans and I”. (Note that who we refer to by “I” when expanding is also what marks my and your we-­ mode or role-mode state as different even when we mutually include ourselves in the same “we” and represent the same position.) Likewise, there cannot be a role consciousness that is not a modification of a we- or I-consciousness. It must be possible for “As committee members” to be followed by “we”, which in turn it again must be possible to expand in the way just described. And it must be possible for e.g. “As CEO” to be followed by “I” in corresponding fashion. 2. We-mode and role-mode consciousness are both bigger and smaller than I-consciousness A graphic way to state an important relation between role-, we- and I-consciousness is to say that we- and role-consciousness are both smaller

9  From We-Mode to Role-Mode 

193

and bigger than I-consciousness. They are smaller because any particular form of we- or role-consciousness is just one of the we- or role-modes a given individual participates in. And they are bigger in so far as we- and role-mode are modes of identification with something larger. 3. Role-mode and we-mode representation can both misrepresent Just like we-mode representation, role-mode representation can misrepresent. I may represent myself as CEO and take some positions towards the world from the CEO vantage point, while still failing to be CEO because relevant others do not recognize and therefore do not represent me as such. So just like in the we-mode case successful role-mode representation is a shared labor of the subjects involved in the relevant institutional context. 4. Roles are not merely belief- or observer-dependent Roles are essentially a matter of the intentionality of subjects. However, they cannot be simply reduced to being belief- or observer dependent, as it is attempted at least in John Searle’s early work on institutional reality (Searle 1995). On Searle’s view of institutional reality, a subject has certain powers and obligations because a collective jointly believes that it has these powers and obligations. But, for example, for somebody to be a CEO, it is not enough that a group of outside observers believes this person to be CEO. This is not sufficient for two reasons: first, because the attitudes of outsiders are not enough. It is the attitudes of the co-subjects, in this case, the co-workers and other members of the relevant corporation, that are crucial. They must accept that this person is CEO, and they must accept it from the vantage points of their respective positions, for example, as employees and as relevant government officials. Second, the relevant stances are not merely theoretical / observational, but they are essentially practical stances. They are practical stances of recognizing and accepting the powers and the authority of the role bearer. (To think that only beliefs and observation are relevant is an example of the bias for the theoretical discussed above.)

194 

M. Schmitz

5. Roles are closely connected to theoretical and practical powers and procedures In an institutional context, there are rules that determine which theoretical and practical powers a subject as acting in their role has and which procedures have to be followed in certain situations. There is again a familiar structure in that these powers and procedures connected to the role may both extend and restrict the theoretical and practical reach of the subject acting in the role—they make it bigger in some ways and smaller in others. For example, as a judge or juror I may have access to knowledge I wouldn’t have outside of my role; but I may also be unable to make use of knowledge I have, for example, evidence that is legally inadmissible. The same is true for practical powers: I have access to the power of the organization I am a part of, but its procedures may also bar me from using some of my personal powers. 6. We-mode and role-mode self-consciousness is the consciousness of taking a position, not of introspecting it Both we-mode and role-mode consciousness are forms of self-­ consciousness. A subject is aware of itself as the member of a group, or as the occupant of a role in that group, and through such episodes of self-­ consciousness the group is also conscious of itself. But neither we-mode nor role-mode consciousness is introspective self-consciousness. That is, I do not look into myself to determine what my positions as a group member or as a role occupant are. Nor does the group as a whole take an introspective stance towards its attitudes. When we jointly take positions, our minds are directed towards the world. The relevant self-consciousness of the individuals and the groups they constitute is a backgrounded consciousness of taking positions in a self-aware way. Likewise, when making up my mind as a role occupant, I am not directed at my role as an object. I rather look at the world from the point of view of the role and take positions in light of the theoretical and practical knowledge, the values, tasks, obligations, powers and so on, that the role affords.

9  From We-Mode to Role-Mode 

195

7. In both we-mode and role-mode one takes positions one would not take merely as an individual or as a private person For example, in we-mode identification with Germany and / or in my role as a German citizen, I may accept responsibility for crimes committed by Germans acting in their roles in the German government, even if I don’t accept any personal responsibility for these crimes. (I may still be personally affected by these crimes as being German is part of who I am.) Politicians in particular often take role-specific positions. For example, when Angela Merkel was both chancellor of Germany and leader of the CDU, she sometimes took a position as party leader that she did not take as chancellor, or at least not yet. Similarly, a member of the police may feel duty-bound to arrest somebody for the use of marijuana even if privately or as a citizen they do not believe that marijuana use should be legally sanctioned. 8. We-mode and role-mode help to explain the structure of atti tude conflicts While taking different positions in a role than one takes in one’s personal life may lead to a life that is inauthentic, we must also acknowledge that any life within a society or group is to some extent based on compromise and accepting certain rules and ways of acting one may not agree with, at least not initially and in some cases never. So conflicts between the requirements of different roles, or between one’s personal values and attitudes and the values and attitudes required by certain roles or groups, are an essential part of life, in particular of modern life in a large and diverse society. It may be tempting to think that the notion of role-mode is meant to sanitize this kind of conflict by compartmentalization, or at least could be used for this purpose: we would just say that the relevant attitudes are held in different modes and the conflict would be gone. But that is not intended at all: the role-mode account is meant to help explain the structure of the conflict rather than to explain it away. Role-modes do constitute a certain degree of compartmentalization, but the compartments are not completely sealed off from one another and so conflicts still arise. The

196 

M. Schmitz

point is just that a conflict within a role is different than one between different roles. If, for example, as a government official I have to do something with which I disagree as a private citizen as in our police example above, this is a different kind of conflict than is a clash of attitudes within a given role. To resolve the former kind of conflict, I may have to work to change the requirements of the role. Or I may just learn to live with the fact that I represent values and rules that are not always the ones I personally favor. I may even strongly identify with doing this because I believe it to be essential to the functioning of a diverse and liberal society. But I could also decide that this conflict is unbearable and that the only right thing for me to do is to resign my role, or to try to start a revolution. 9. Switching from and to modes works similarly for we- and role-mode How do we decide a conflict between the requirements of different roles, or between a role and my values and attitudes as a private person? Do we need a special mode here like an “overall mode” (see Laitinen (2017)), in which conflicts between different modes can be negotiated and resolved? It seems to me that we do not. As all we-modes and role-­ modes are modifications of the individual consciousness or I-mode, it is simply in I-mode that the conflicts will be negotiated and resolved. The I-mode already is the “overall mode”. How then can we switch between different modes as between I-mode and we-mode, or between different role-modes? The switch between I-mode and we-mode is at least sometimes arational, like when we perform a switch based just on emotional identification, and we can certainly also switch into a role or between different roles in this way. What kinds of switching are there? A temporal switch can just be triggered by being immersed in a group context—e.g., joining my friend group for a hiking excursion, or simply starting work in my professional role. When we talk about joining a group or starting a job, this will often and, in the latter case typically, be based on deliberation. Such deliberation occurs in I-mode—by which I essentially mean here: not in the mode of the group I am thinking about joining or that of the role I am about to adopt, as my reasoning may well be, partially or entirely,

9  From We-Mode to Role-Mode 

197

conducted from the point of view of modes of identification with one or several other groups or roles. A crucial point is that this kind of deliberation of whether to join a group or to adopt a role is essentially different from reasoning and acting from the point of view of a group or role. When engaging in the former kind of deliberation, I may try on, as it were, the attitudes and responsibilities of the new role to see if they fit me; but when I really adopt them, when I really grow into this role and make it mine, which of course will typically take time—even after I have already legally adopted it—I see the world from the vantage point of the role. I think this means that there must be a presumption in favor of at least some of the positions, the attitudes, values and rules defining the role when reasoning from the point of view of the role. Otherwise I haven’t really identified with the role and my reasoning collapses into I-mode reasoning. One could say that this is how the I-reductionist envisages things. Again, it seems to me that what reductionism misses is how we often reason, think, act and perceive from the point of view of the role or group, how our group memberships and the roles we occupy structure our intentionality. 10. Role-mode reaches all the way down So far, we have focused on conceptual level positions and on reasoning. But in accordance with the layered model I also want to emphasize that role-mode reaches all the way down, is embodied and enacted. For example, the role structure in a group of animals may be entirely realized in their sensory-motor-emotional interaction, in how they experience the other group members and the world at that level. But professional and other institutional roles also reach all the way down in that they are reflected in characteristic perceptual and actional experiences and skills and in the structure of attention. For example, members of the police on the beat perceive the world differently, attend to other things, and carry themselves differently than, say, construction workers in the same environment.

198 

M. Schmitz

6 Conclusion Is the subject mode account collectivist or individualist? Let me conclude this paper by reviewing its argument in the light of this fundamental question. In the spirit of Raimo Tuomela I have argued that collective intentionality is conceptually—or intentionally, as we might say more broadly—irreducible. That is, humans and some other animals display irreducible forms of group and role mindedness. I have gone beyond Tuomela in arguing that this intentionality is also constitutive of groups. When group members mutually experience and otherwise represent themselves as co-subjects of positions they jointly take up towards the world and perceive the world and act in it in modes of identification with the group, group subjects do come into being. I have further begun to extend the concept of the we-mode made popular by Tuomela to role-mode in order to account for the intentionality of group agents such as institutions and organizations. It seems to me we can then be equally realist about such subjects. Again, there is no reason to leave the concept of a group subject to mysterians, or to turn fictionalist in any sense regarding group agents. This is because an antireductionist and fully realist view about collective intentionality which emphasizes that we take positions towards the world in irreducible modes of identification with groups and roles, in no way commits us to group subjects or group minds that float free of individual subjects and minds. Our account can and should still be individualist in the sense that it holds that all we- and role-consciousness is a modification of I-consciousness and that a group can only take a position through at least one individual taking it as a group member and / or as the occupant of a role within the group. At the same time, that one individual takes such a position is never sufficient for the group taking it because representing the group’s positions is always essentially a task shared by all group members. Acknowledgements  This paper is dedicated to the memory of Raimo Tuomela, who was not only an intellectual inspiration for me, but also a friend. For some personal remembrances by myself and others, visit https://isosonline.org/ In-­Memoriam-­Raimo-­Tuomela

9  From We-Mode to Role-Mode 

199

I thank audiences in Helsinki and Rome, Manchester and San Diego (respectively online), especially Gunnar Björnsson, Abe Roth and Randell E. Westgren, and the editors Miguel Garcia and Rachael Mellin for questions and comments, and Zuzana Toth for her support.

References Bermúdez, José Luis. 1998. The Paradox of Self-Consciousness. MIT Press. ———. 2011. Bodily Awareness and Self-Consciousness. MIT Press. Bratman, Michael E. 2014. Shared Agency: A Planning Theory of Acting Together. Oxford: Oxford University Press. Gilbert, Margaret. 2013. Joint Commitment: How We Make the Social World. Oxford: Oxford University Press. Green, Mitchell. 2022. Force, Content and Translucent Self-Ascriptions. In Force, Content and the Unity of the Proposition, ed. Gabriele M.  Mras and Michael Schmitz, 195–214. Routledge. Hanks, Peter. 2015. Propositional Content. Oxford University Press. ———. 2019. On Cancellation. Synthese 196 (4): 1385–1402. Laitinen, Arto. 2017. We-Mode Collective Intentionality and Its Place in Social Reality. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. Gerhard Preyer and Georg Peter, 147–167. Springer. Ludwig, Kirk. 2016. From Individual to Plural Agency: Collective Action I. Oxford University Press. ———. 2017. From Plural to Institutional Agency: Collective Action II. Oxford University Press. Recanati, François. 2019. Force Cancellation. Synthese 196 (4): 1403–1424. ———. 2022. Entertaining as Simulation. In Force, Content and the Unity of the Proposition, ed. Gabriele M. Mras and Michael Schmitz, 112–135. Routledge. Schmid, Hans Bernhard. 2014. Plural Self-Awareness. Phenomenology and the Cognitive Sciences 13 (1): 7–24. ———. 2017. What Kind of Mode Is the We-Mode? In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. Gerhard Preyer and Georg Peter, 79–94. Springer. Schmitz, Michael. 2017. What Is a Mode Account of Collective Intentionality? In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy

200 

M. Schmitz

of Raimo Tuomela with His Responses, ed. Gerhard Preyer and Georg Peter, 37–70. Springer. ———. 2018. Co-Subjective Consciousness Constitutes Collectives. Journal of Social Philosophy 49 (1): 137–160. ———. 2019. Force, Content and the Varieties of Subject. Language & Communication 69: 115–129. https://doi.org/10.1016/j. langcom.2019.10.001. ———. 2020. Of Layers and Lawyers. In Social Ontology, Normativity and Philosophy of Law, ed. Miguel Garcia, Rachael Mellin, and Raimo Tuomela, 221–240. De Gruyter. ———. 2022. Force, Content and the Varieties of Unity. In Force, Content and the Unity of the Proposition, ed. Gabriele Mras and Michael Schmitz, 71–90. Routledge. Schweikard, David P., and Hans Bernhard Schmid. 2013. Collective Intentionality. In Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta. http://plato.stanford.edu/archives/sum2013/entries/collective-­intentionality/. Searle, John R. 1995. The Construction of Social Reality. New York: Free Press. ———. 1998. Social Ontology and the Philosophy of Society. Analyse & Kritik 20 (2): 143–158. Tuomela, Raimo. 1995. The Importance of Us: A Philosophical Study of Basic Social Notions. Stanford University Press. ———. 2007. The Philosophy of Sociality: The Shared Point of View. Oxford University Press. ———. 2013. Social Ontology: Collective Intentionality and Group Agents. Oxford: Oxford University Press. ———. 2017. Response to Michael Schmitz. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. Gerhard Preyer and Georg Peter, 71–78. Springer. Wittgenstein, Ludwig. 1958. The Blue and Brown Books. Oxford: Blackwell.

10 Group Morality and Moral Groups: Ethical Aspects of the Tuomelian We-Mode Björn Petersson

1 Introduction Raimo Tuomela’s we-mode groups are partly characterized by norms. Some norms may be characteristic of all we-mode groups, like the norm restricting a member’s right to leave the group without permission. They may also be group-specific and dependent on the goals of the particular group. These group norms, Tuomela stresses, are not to be conflated with moral norms. A group of criminals may follow their “group morality” without behaving morally in any proper sense (Tuomela 2013, p. 295). Some have thought that this aspect of Tuomela’s theory has implausible ethical implications concerning the rights and autonomy of members in we-mode groups. (Corlett and Lyons Strobel 2017) This worry about a substantive ethical issue vanishes, I argue, on a plausible interpretation of Tuomela’s notion of social normativity and a reasonable precisification of the notion of autonomy in this context.

B. Petersson (*) Department of Philosophy, Lund University, Lund, Sweden e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_10

201

202 

B. Petersson

On the other hand, Tuomela’s general silence on the nature of moral normativity makes it unclear how his distinction between social and moral normativity should be drawn more precisely. Is this a difference in kind, or merely a difference in the scope or social basis of the norms in question? Perhaps moral normativity is just a species of social normativity, but applied to, and/or grounded in, different kinds of social entities? I find the possibility worth exploring briefly with the aid of resources available within a broadly Tuomelian framework. The point of departure for such speculations, I suggest, should be Tuomela’s views on the role of we-intentions for resolving potential social dilemmas, and his views on social institutions and collective acceptance. Like Jeremy Koons (2019), I believe that Tuomela’s framework could be used to refine Wilfrid Sellars theory of ethical judgments as expressions of we-intentions. My preferred interpretation of Tuomela differs from Koons’ though, and so does the resulting reconstruction of Sellars, which is more Humean than Kantian.

2 Autonomy in Tuomela’s We-mode Groups? According to Tuomela, a “collective g consisting of some persons […] is a we-mode social group (if and) only if (1) g has accepted a certain ethos, E, as a group for itself and is committed to it; (2) every member of g group-normatively ought to accept E as a group member (and accordingly to be committed to it as a group member), at least in part because the group has accepted E as its ethos; (3) it is a mutual belief in the group that (1) and (2). (Tuomela 2013, p. 27).

So, by definition, members of we-mode social groups are subject to group norms related to the group ethos. Furthermore, we-mode groups have norms restricting members’ right to leave the group without permission.

10  Group Morality and Moral Groups: Ethical Aspects… 

203

A paradigmatic we-mode group is not only democratic in that it itself can determine its ethos and try to promote it, but is also autonomous in the sense that the members may freely enter or exit the group conditionally on their participating (or ceasing to participate) in the collective acceptance and promotion of the ethos. Thus a paradigmatic we-mode group is open relative to its actual and potential members […](Tuomela 2013, p. 30).

In non-paradigmatic we-mode groups, the group ethos may in part be externally determined, and members may be less free to enter or exit the group. But even in paradigmatic we-mode groups, special conditions may concern individuals who have made explicit agreements or contracts in their capacity as group members. Yet the members typically need the permission of the other members to rescind their commitment once more specific goals and intentions have been accepted for the group, since the feature of group-normativity […] binds the members strongly together around a shared ethos (Tuomela 2013, p. 30).

This requirement for permission to leave has worried some. The worry is that the analysis does not grant members in we-mode groups personal autonomy, at least not sufficiently. [T]o the extent that we-mode group membership requires the consent of even anyone else in the group for one to either, say, resign membership or in some other way exit the group, in what sense is genuine personal autonomy preserved in Tuomela’s analysis? And how reasonable is it to suppose that anyone not lacking in self-respect would even consider joining such a group if this crucial property of personal autonomy is either absent or significantly diminished? It amounts to group membership without a right to either defect or resign membership on one’s own. And this seems more akin to membership in a street or prison gang or the Mafia than it does a group that values personal autonomy and rights that exist beyond the social rights that are granted by the group. (Corlett and Lyons Strobel 2017, p. 564)

In the present context, one may distinguish roughly between a non-­ normative concept of autonomy as an actual ability to exercise one’s

204 

B. Petersson

capacity for self-determination without external obstacles, and a normative notion of autonomy as a moral right to exercise one’s capacity for self-determination without interventions from others. It seems rather obvious that when you enter a joint project in the we-­ mode, your autonomy in the first, non-normative, sense becomes restricted. If you join a group and start interacting with people, this will eliminate some options that would otherwise have been open for you, and the mere presence of others and their behaviour may constitute rather concrete obstacles that you have to navigate, and which affect your action alternatives. Corlett and Lyons Strobel do not question this, and they do not claim that this type of restriction would pose a problem for Tuomela. However, they do question whether his account does justice to respect for autonomy in the second, normative, sense. “It is the second ethical sense of personal autonomy that we argue poses problems for Tuomela’s analysis insofar as a social ontology must at some point be made congruent with, or at least be made not to be incongruent with, plausible ethical theory” (Corlett and Lyons Strobel 2017, p. 568). To begin with, I find the idea that your moral right to exercise your capacity for self-determination becomes affected when you enter a joint project rather obvious and acceptable as well. When you enter a group committed to a common cause, you make commitments to others, at least implicitly. Typically, the other members have invested in the project and rely on you doing your part and not suddenly leaving them without any explanation. So, you have taken on new obligations, which must be balanced against your general right to self-determination. In that sense, your autonomy in the normative sense—your moral right to exercise your capacity for self-determination without interventions from others— becomes restricted to some extent, due to other moral considerations created simply by your interaction with group members. In his response to Corlett and Lyons Strobel, Tuomela admits that the critics are to some extent right in the sense that in his “account the members of a well-­ functioning we-mode group can be said somewhat figuratively to have given up part of their autonomy and self-determination to their we-mode group” (2017, p. 29). To me, this admission seems unproblematic from an ethical point of view.

10  Group Morality and Moral Groups: Ethical Aspects… 

205

Nevertheless, Tuomela is concerned about the complaint from Corlett and Lyons Strobel. In his response, he weakens his claim about the requirement for permission to leave we-mode groups. Or rather, he tries to make clear that the claim is weaker than it may have sounded in his previous work. Tuomela says that his phrase “permission to leave the group” may have been misleading, and he points out that he has been talking about “cases where leaving the group harms the other members and propose[d] that sometimes rather mere informing the members might be appropriate” (2017, p. 29). I think that another and more straightforward line of defence would have been natural. The norms restricting a member’s right to leave or to abandon the group ethos in a Tuomelian we-mode group are social norms. The definition states that a member group-normatively ought to accept the group ethos etc. Social norms, according to a fairly standard view, are norms that exist within a certain social group and apply to the members of that group. Such norms involve “shared expectations about what should or should not be done in different types of social situations” (Bicchieri et al. 2018). Social norms typically involve a sense of accountability and the possibility of sanctions. Crucially in the present context, they can be harmful, inefficient or immoral. This rough characterization of social norms is in line with Tuomela’s views. “Group reasons in the present sense of participation […] involve some strong normative bonds between the members, which minimally include mutual expectations with significant normative entailments” (Tuomela 2013, p. 41) “Of course the members may exit the group but then they have to be prepared to suffer the (possibly) agreed-upon sanctions for quitting” (Tuomela 2017, p. 29). From Tuomela’s claim that we-mode groups have group norms restricting a member’s right to leave or to abandon the group ethos, nothing follows about the existence, non-existence or justification of rights and obligations that exist beyond the rights and obligations created within the group. So, it could be the case that while the member is not group-­ normatively allowed to leave the group without permission, she may be legally or morally allowed, or even obliged, to do so. In other words, since Tuomela’s theory is silent on which moral rights we have, and has no

206 

B. Petersson

implications about which moral norms are plausible,1 it cannot be “incongruent with plausible ethical theory”. In a footnote, Corlett and Lyons Strobel notes that “[i]ronically, Tuomela mentions ‘Mafia groups’ in the context of groups having a morality absent a universalized ethic (Tuomela 2013, p.  295, note 2)” And they suggest that “he seems not to notice that his analysis of we-­ mode groups bears a striking similarity to Mafia groups in the manner in which is suggested herein” (Corlett and Lyons Strobel 2017, p. 571). I think that this points to a strength of Tuomela’s account rather than to a problem. An account of strong collectivity should capture all kinds of tightly knitted groups, not just the honorable ones. Group morality as a system of social norms can obviously clash with morality proper.

3 Tuomelian Meta-ethics? At some point in their discussion of Tuomela, Corlett and Lyons Strobel seem to turn their ethical challenge—that Tuomela’s views on we-mode groups clash with substantive moral norms aimed at protecting personal autonomy—into a metaethical challenge. They request, from Tuomela, an analysis of social groups that can accommodate moral facts and make sense of ethical normativity. Part of our criticism is that his social ontology is incapable of accommodating some basic moral facts, such as moral rights to unilaterally exit a we-­ mode group one has joined in concert with others (Corlett and Lyons Strobel 2017, p.563). But this points to the glaring problem with his social ontology: It fails to provide an analysis of social groups that can make sense of at least some of  It may be a slight exaggeration to say that Tuomela´s theory has no ethical implications, or that Tuomela completely refrains from claims about ethical, or for that matter political, norms. For instance, he says the following. “A liberal communitarianism (which the present theory is compatible with) enriched by the conceptual tools and theses of this book might result in a normative political doctrine contributing to ‘saving the world’” (Tuomela 2013, p. 17). But I think that it is fair to say that while Tuomela sometimes makes tentative ethical claims from the perspective of his social ontology, he mostly leaves it to others to develop them. 1

10  Group Morality and Moral Groups: Ethical Aspects… 

207

the vital concepts of ethical normativity (Corlett and Lyons Strobel 2017, p. 568).

Can Tuomela’s theory “accommodate” ethical normativity? It certainly makes room for ethical normativity and allows for its possibility. The very contrast between group norms and ethical norms in the standard sense makes it clear that Tuomela acknowledges ethical normativity. But does his theory provide us with the conceptual tools to explicate ethical normativity, even if this is beyond Tuomela’s ambitions? Corlett and Lyons Strobel “assume for the sake of [their] paper that some versions of moral realism and moral rights realism are sound” (2017, p. 568). If that assumption is made for the sake of this specific discussion, I find it slightly unexpected. Given the nature of Tuomela’s project, the complaint that his social ontology fails to provide us with an account of ethical normativity would seem more natural if one assumes that moral norms are fundamentally expressions of social attitudes rather than descriptions of independent moral facts, i.e., if one assumes that ethical normativity is a species of the kind of normativity that Tuomela does try to explain. Accounts of that kind have been proposed. Wilfrid Sellars, whose work has several affinities with Tuomela’s, proposes what one might call a “we-­ expressivist” theory of moral judgments. Moral judgments are expressions of we-intentions, where the scope of the ‘we’ is “rational beings generally” (Sellars 1980, p. 101). Sellars’ “intention” is a technical term “for any cognitive state intrinsically geared towards action” (Loeffler 2020, p. 116), i.e., desires, volitions, action intentions in the ordinary sense, etc. While Sellars is an expressivist about moral judgments in claiming that they are expressions of intentions rather than expressions of beliefs, he thinks that such judgments presuppose some interpersonal and categorically valid we-intention. The Kantian appeal to ‘all rational beings’ means that there are universally valid standards of correctness for moral judgments. And he thinks that his meta-ethical explanation of moral judgments means that the speaker in the end will be committed to accepting a specific substantive moral view—welfare utilitarianism.

208 

B. Petersson

Relatively little seems to have been written on this part of Sellars’ work, and it is sometimes complained that it has been unfairly neglected.2 However, in The Ethics of Wilfrid Sellars (2019), Jeremy Koons actually suggests that Sellars’ rather underdeveloped notion of we-intentions at least in this moral context should be understood in Tuomelian (and Gilbertian) terms (Koons 2019, ch.4). So, could the group-normative “I don’t have the right to leave this group without permission” and the moral “I have the right (or obligation) to leave the group without permission” be mutually consistent expressions of we-intentions differing merely by their individuations of “we”? There are some difficulties with combining Sellars and Tuomela in this way. First, Tuomela’s definitions of we-mode intentions require quite complex beliefs about other group members, their beliefs, including beliefs about what the we-intending agent believes, and motivations. It seems unlikely that these conditions will apply to how an individual thinks of large collectives, let alone of humanity, or of “all rational beings”. Second, intuitively there is a difference in epistemic status between social and moral norms, not merely a difference in scope and base. Sellars’ response to this lies in his Kantian appeal to “all rational beings”—which is supposed to explain why there are interpersonal categorical standards of correctness for moral norms, making them differ from other social norms. How could this be captured in a Tuomelian framework? Third, Tuomela’s explicit definitions of the term “we-intention” do not fit well with Sellars’ characterizations of we-intentions. Tuomela’s analysis of the “we-mode we intention”, if we take his explicit definitions seriously, is arguably unlike Sellars’ a “content” account in terms of the standard subject/mode/content taxonomy of theories of collective intentionality, despite his label (Schmitz 2017; Petersson 2017; and Ludwig 2016, p. 234).

 “Given this prominence of practical philosophical concerns and the rich body of work Sellars produced on practical philosophical issues, it is surprising that the steep uptick in Sellars scholarship in the last two decades did until now almost completely bypass Sellars’ practical philosophy” (Loeffler 2020, p. 114). 2

10  Group Morality and Moral Groups: Ethical Aspects… 

209

My analysis of we-mode we-intention can be summarily formulated as follows. (WI) A member Ai of a collective g we-intends to do X if and only if (i) Ai intends to do his part of X (as his part of X); (ii) Ai has a belief to the effect that the joint action opportunities for an intentional performance of X will obtain (or at least probably will obtain), especially that a right number of the full-fledged and adequately informed members of g, as required for the performance of X, will (or at least probably will) perform their parts of X, which under normal conditions will result in an intentional joint performance of X by the participants; (iii) Ai believes that there is (or will be) a mutual belief among the participating members of g (or at least among those participants who perform their parts of X intentionally as their parts of X there is or will be a mutual belief ) to the effect that the joint action opportunities for an intentional performance of X will obtain (or at least probably will obtain); (iv) (i) in part because of (ii) and (iii) (Tuomela 2007, pp. 93–94; but see also 2000, p. 64).

Roughly, this definition explicates “we-mode we-intends” in terms of the member’s intending to do her part of a collective action, along with beliefs about joint performance. So, what makes the attitude collective is that a notion of collectivity figures in its content, rather than that the attitude is held in a special mode, or from a special perspective.3  In the conventional use of the term “mode” about attitudes, a mode is what makes an intentional state the kind of state that it is. Hoping that P, wondering whether P, believing that P are different modes of attitudes with the same content. The term mode is therefore a bit unfortunate in the present context, because the we-mode/I-mode distinction is supposed to cut across the modes in the conventional sense. The we-mode is not supposed to be a distinct attitudinal mode on a par with beliefs and desires. (Schmid 2017) I think that the term “perspective” would be less likely to create misunderstandings. My preferred version of this idea of a we-perspective starts from the assumption that some intentional states are self- and agent-referential. That is, if we want to give an analysis, in terms of success conditions, of the kind of state that it is, this analysis must include a reference to the state itself and its bearer. My intention to raise my arm is only successful if I raise my arm by way of this very intention, and “I perceive this flower” is only veridical if the flower causes my perception. But arguably (John Searle, Kirk Ludwig and others think otherwise) the flower’s causal relation to me is not part of the content of my perception—what I perceive is the flower and what I intend is to raise my arm. The intentional subject of the perception is instead a perspectival feature of the intentional state. The idea is that the intentional subject, the perspective from which a certain content is conceived, can be a 1st person plural perspective, although the ontological subject—the bearer of the intentional state—is a single individual. (Just as the perspective of some types of intentional states need not be “here and now”). (Petersson 2014, 2017) 3

210 

B. Petersson

On the other hand, in his applications of the concept of we-mode we-­ intentions, Tuomela treats the we-mode as a perspective or a form of intending rather than as intending where various conceptions of the group and its members merely figure in the content. So, Tuomela’s use of the term “we-mode” is ambiguous. Elsewhere I have argued that interpreting Tuomela’s we-mode in the perspectival sense, despite his explicit definitions, fits better with the role he assigns to the we-mode in certain potential social dilemmas, where dissolving the potential dilemma requires a shift of agential perspective—a shift from thinking about what I should do for us to thinking “What should we do?” (Petersson 2017; Blomberg and Petersson 2023) According to the standard reading, Sellars’ account is a “mode” account in terms of the subject/mode/content taxonomy of theories of collective intentionality. (See e.g. the Sellars paragraph in Schweikard and Schmid 2021.) Sellarsian we-intentions appear to be intentions held from a collective point of view, distinguished from I-intentions by their form rather than by their content. Admittedly, there is room for both interpretations of Sellars position. However, his typical expressions for the form of possibly shareable intentions do not place any necessary references to the collective or individual subject in their content. “Thus, consider the intention ‘Shall be [whooping cranes survive]’” (Sellars 1980, p. 98). And when he describes how individual action intentions can be derived from we-intentions, he says, e.g., that “we can say that Jones intends to do A sub specie “one of us” and flag our representation of his intention with a subscript “we”, thus, Jones intends “Shallwe [I do A]” (1980, p. 99), which seems to be a natural way of expressing the thought that even when my individual action is what figures in the content of my intention, the intention can be held from my group’s perspective. Moreover, he contrasts “private” intentions, which may concern the group of which I am a member (in his example the Whooping Crane Society) with intentions held from the point of view of a member of the group. In his explication of Sellars’ metaethics, Koons relies on Tuomela’s definitions of we-intention, and admits that he offers a Tuomelian (and Gilbertian) “reconstruction” of Sellars’ metaethics, with “amendments” and “additions”, rather than a mere interpretation. In doing so, he reads

10  Group Morality and Moral Groups: Ethical Aspects… 

211

Tuomela’s analysis, and thereby Sellars’, as a “content” account. According to Loeffler, “[T]his Gilbertian and Tuomelian blend is ill-suited to do the work Sellars assigns to his concept of a moral we-intention” (Loeffler 2020). In my view, a “perspectival” understanding of Tuomela’s we-mode would do that work better, and less reconstruction would be needed. The perspectival understanding of we-intentions has merits on its own, but what difference does it make in the present context? My tentative and somewhat vague answer—and here I can mainly appeal to intuitions—is that the perspectival understanding of we-intentions provides us with a better understanding of the sense of mind-independent standards of correctness, the sense of bindingness. The (moral) sense that what I must do is not up to me. It is something that I infer from what everyone must do. If I merely think about what I should do for us, like when Jones in Sellars’ example is concerned about the Whooping Crane Society, there is an element of up-to-me-ness in the decision process, as opposed to when I infer what I should do from what I think that we should do, from the 1st person plural perspective. There is a sense of bindingness in my conclusion about what I ought to do as part of what all of us ought to do, as opposed to what I ought to do given my beliefs about you and my commitments to you. Consider the difference between thinking “I intend that we keep our promises” and thinking “we intend that promises are kept”. While I can feel that it is up to me whether I shall respect others property or keep my promises in each specific case, I don’t feel that it is up to me whether I morally ought to do so, not because there are moral truths independently of human attitudes, but because my pro-attitude towards these conventions is held from the community’s perspective, and not only from my individual perspective, or even from my family’s or some other small social group perspective. There is another feature of the perspectival understanding of we-­ intentions that might further strengthen the hypothesis that we-­ intentionality is part of what explains the sense of bindingness of moral norms. It seems plausible to think that to some extent I can voluntarily choose whether to worry about myself or whether to be concerned about my group. I can actively direct my attention to my interests and problems, or to our interests and problems. However, I cannot intentionally

212 

B. Petersson

choose the perspective from which I am worrying. Like Michael Bacharach, whose theory of team reasoning Tuomela interprets in terms of the Tuomelian we-mode, I think that shifts of agential perspectives—what Bacharach calls “agency transformations”—are prompted rather than chosen (Bacharach 2006; Petersson 2017; Blomberg and Petersson 2023). In this matter, I dissent from Tuomela, who takes the “common-sense view that not only can the mode be intentionally selected by an agent but in some cases it can also be rationally intentionally selected”. (Tuomela 2013, p. 195) So, in the choice between paying full attention to every detail in Tuomela’s rather complex definitions of the we-mode, and doing justice to his interesting applications of the concept, especially in relation to game-theoretical puzzles, I advocate the latter strategy.4 This means that we should understand “we-mode” basically as a 1st person plural perspective. This will also make the notion of a we-mode better suited for a way of understanding ethical normativity, to some extent along the lines proposed by Sellars. Like Koons, I propose some amount of reconstruction of Sellars, but in another direction. First, abandon Sellars’ solution to what Michael Smith later labelled “the moral problem”. We should give up his appeal to “all rational beings” as a way of combining internalism about moral judgments with the admission that there seems to be mind-independent standards of ethical correctness. It is unlikely that agents ever identify with all rational beings and view things from this perspective. Sellars own normative conclusion—that a person who does this will end up being a welfare utilitarian—also shows that it is difficult to avoid question-begging understandings of practical rationality once we go beyond purely instrumental conceptions. Very roughly, the result of this reading of Tuomela and this modification of Sellars would be a view which says that moral judgments express pro-attitudes “in the heads of individuals” but held from the perspective of the community with which the individual identifies.  I believe I got some support for this strategy when I questioned some of these analyses in a panel debate with Tuomela, where one of his responses to me was that the passages that I had read as strict conceptual analyses were not meant to be read in that strict way. (ENSO IV, Palermo 2015) 4

10  Group Morality and Moral Groups: Ethical Aspects… 

213

In places where Tuomela refers to morality proper, as opposed to the group morality that he attempts to explain and that even Mafia groups may possess, he sometimes labels it “Kantian” or “universalizable”. So, could the idea of binding universalizable moral norms be explained in terms of we-intentions, understood in the perspectival sense? My suggestion is that this requires us to bring in social conventions or institutions, in a a basically Humean account of the social basis of universalizable moral norms. As Tuomela says, “what is thus up to us clearly includes social institutions /…/. Here, for example, various kinds of social positions and roles are included and so are social rules and norms. Accordingly, law is included here, and a case can be made for morality to be included as well” (Tuomela 2002, p. 147). More specifically, I am thinking of the part of morality that Hume calls “the artificial virtues”. Virtues like justice, respect for other’s property, promise-keeping and law-abidingness. Unlike Hume’s “natural virtues”, artificial virtues are socially constructed but the norms governing them are not relative to small specific social groups, they are rigid and apply to friends and enemies alike. They don’t vary with the agent’s feelings for the patient, and they rest on general conventions, adopted in response to potential social dilemmas, like the prisoner’s dilemma type situations that Hume uses in his examples. Tuomela actually suggests that morality in the form of social rules and norms are social institutions. And like Hume, he stresses that such institutions evolve in response to social dilemmas and coordination problems. Another similarity concerns how social patterns can develop without explicit agreements, just like, as Hume says, two men who start rowing about together gradually can find the optimal pattern of coordination. Social institutions typically have as their general goal or at least function to create order in society by solving coordination problems and collective action problems involving conflict between individual and collective rationality. (Tuomela 2013, p. 215)

These conventions, in turn, rest on what Hume calls “a general sense of common interest”. And he even says that a convention “is only” a general sense of common interest (Hume 1739, 3:2:2). The latter concept has an

214 

B. Petersson

important explanatory role in Hume’s theory of the artificial virtues but has been given relatively little attention (unlike, for instance, his notion of convention). So, my suggestion is that an understanding of Hume’s “general sense of common interest” in terms of a perspectival understanding of we-intentions would be consistent with the role Hume assigns to the phenomenon. So, a reasonable Tuomelian account of universalizable moral norms would be an account where we-intentions play an indirect role, as they do in Tuomela’s as well as Sellars’ (and later Searle’s) account of the role of collective acceptance for the creation of social institutions, and as the general sense of common interest does for Hume’s account of how artificial virtues come into existence and persist. Individual moral judgments of the universalizable kind express collective approval—approval from the 1st person plural point of view—of social institutions in the form of certain types of rules and social practices, developed in response to social dilemmas and coordination problems. In Sellars’ original account there is a normative element built into the analysis of what a moral judgment is: the appeal to the perspective of all rational beings. That enables him to draw normative conclusions from his metaethics—welfare utilitarianism. By contrast, the suggested approach here has no direct normative implications, and the we-­ perspective is the perspective of the larger community with which the agent actually identifies rather than an abstract community of all rational beings. So, on this view, the difference between the prison gang morality and general morality is after all a matter of base and scope, where the we in the moral case is the larger community with which the speaker identifies. The perspectival understanding of we-intentions does not presuppose that we-intending agents have reflexive beliefs about how others are motivated, although such beliefs may be among the triggering causes prompting a perspectival change, along with less conscious and more primitive social cues. So, it is in a sense less demanding than the Tuomelian we-­ mode as it appears in his explicit definitions, where quite complex beliefs about the others’ beliefs and motivations are among the necessary conditions. This means, also, that the question of how to individuate the ‘we’ in the we-perspective characteristic of morality, will be of a different kind.

10  Group Morality and Moral Groups: Ethical Aspects… 

215

The ‘we’, on this perspectival understanding, is a community with which the agent identifies. How, exactly, this community must be delimited to give rise to the sense of universality and bindingness that is characteristic of ethical norms can be left open here. One might ask whether this would lead to ethical relativism. In an uncontroversial descriptive sense, an account of the kind suggested here would be relativistic. What rules and practices are approved from the collective point of view will vary with the communities that the speaker identifies with. But relativism in the controversial and more interesting sense is a normative position, implying that the validity or justifiability of moral judgments is relative to the speaker or to the society in which they are uttered. While this may be a legitimate worry about Sellars’ original position (Koons 2019, pp.  93–103), no normative claims of this kind would follow from the sketched account.5

4 Summary Some have argued that the norms assigned to we-mode groups by Tuomela clash with reasonable substantive ethical norms and rights, like the moral right to leave a group without permission. However, Tuomela makes clear that the norms which characterize we-mode groups belong to another category than ethical norms. So, although a person may be torn between the social commitments she has taken on as a member of a we-­ mode group, and what she thinks that morality requires of her, there is no formal inconsistency between these sets of requirements. Moreover, this kind of practical dilemma seems to be a rather common predicament in real life, which Tuomela’s theory provides a plausible explanation of. Tuomela does not say much about how the distinction between these kinds of norms should be drawn more precisely though, or about what characterizes ethical normativity. Like Jeremy Koons (2019), I find it fruitful to bring together Tuomela’s theory of we-mode intentions with  A related issue is whether the account presupposes consensus, or even requires conformism, about moral evaluations. It would reach too far to deal with this issue here, but as Sellars says, “An individual can have an intention of intersubjective form even if no one else in point of fact shares it” (Sellars 1967, quoted by Koons 2019, p. 93). 5

216 

B. Petersson

Sellars theory of ethical judgments as expressions of we-intentions, in order to think about what a Tuomelian theory of ethical normativity might look like. One desideratum for a theory of ethical norms of this kind is that it should give a plausible explanation of the sense of bindingness and interpersonal validity that seems to distinguish ethical norms from other social norms. Both Tuomela’s “we-mode” and Sellars “we-intentions” are open for interpretations in two different directions. Roughly, they can be read as claiming that the ‘we’ figures in the content of these attitudes, or that the ‘we’ is rather a perspectival feature of them. The latter interpretation might help us develop a theory of ethical normativity that meets the mentioned desideratum in a more convincing way than the first interpretation would. One reason for this is that there is an element of “up-to-­ me-ness” in thinking about what I should do for us or for me or for my community, as opposed to when I am concerned from our perspective and infer what I should do from what we should do. Moreover, the we-­ perspective itself is not up to me—agential perspectives are prompted rather than chosen (or so I claim, pace Raimo Tuomela). Acknowledgements  One of the last times I saw Raimo Tuomela was at Social Ontology 2018, where he noticed me across the room at a big reception and rapidly approached me, shadow boxing towards me because of something I had written that he disagreed with. He was in a good mood and wanted me to come to Helsinki to discuss the matter but sadly we never found the time. I am very grateful to Raimo, for his works, for fun and useful conversations at various ISOS events over the years, and for taking the initiative in the International Social Ontology Society to let me host ENSO V in Lund 2017. Versions of this paper have been presented at Social Ontology 2021, San Diego (online) and at the Higher Seminar in Practical Philosophy in Lund. I am grateful to both audiences for helpful comments and criticisms. Special thanks to Olle Blomberg, Mattias Gunnemyr, and Arto Laitinen.

References Bacharach, Michael. 2006. Beyond Individual Choice: Teams and Frames in Game Theory. Princeton: Princeton University Press.

10  Group Morality and Moral Groups: Ethical Aspects… 

217

Bicchieri, Cristina, Ryan Muldoon, and Alessandro Sontuoso. 2018. Social Norms. In The Stanford Encyclopedia of Philosophy (Winter 2018 Edition), ed. Edward N.  Zalta. Accessed June 30, 2022. https://plato.stanford.edu/ archives/win2018/entries/social-­norms/. Blomberg, Olle, and Björn Petersson. 2023. Team Reasoning and Collective Moral Obligation. Social Theory and Practice. Online first January 27 2023. https://doi.org/10.5840/soctheorpract2023120177 Corlett, J.  Angelo, and Julia Lyons Strobel. 2017. Raimo Tuomela’s Social Ontology. Social Epistemology 31 (6): 557–571. https://doi.org/10.108 0/02691728.2017.1346724. Hume, David. 1739. A Treatise of Human Nature. Koons, Jeremy R. 2019. The Ethics of Wilfrid Sellars. New  York: Routledge Studies in American Philosophy. Loeffler, Ronald. 2020. Sellars’ Theory of We-Intentions and Gilbert’s Theory of Joint Commitment: A critical notice of Jeremy R.  Koons, The Ethics of Wilfrid Sellars. International Journal of Philosophical Studies 28 (1): 114–127. Ludwig, Kirk. 2016. From Individual to Plural Agency, Collective Action: Volume 1. New York: Oxford University Press. Petersson, Björn. 2014. Bratman, Searle, and Simplicity: Comments on Bratman, Shared Agency, A Planning Theory of Acting Together. Journal of Social Ontology 1 (1): 27–37. ———. 2017. Team Reasoning and Collective Intentionality. Review of Philosophy and Psychology 8 (2): 199–218. Schmid, Hans Bernhard. 2017. What Kind of Mode is the We-Mode? In Social Ontology and Collective Intentionality. Studies in the Philosophy of Sociality, ed. Gerhard Preyer and Georg Peter, 79–93. Springer. Schmitz, Michael. 2017. What is a Mode Account of Collective Intentionality? In Social Ontology and Collective Intentionality. Studies in the Philosophy of Sociality, ed. Gerhard Preyer and Georg Peter, 37–70. Springer. Schweikard, David P., and Hans Bernhard Schmid. 2021. Collective Intentionality. In The Stanford Encyclopedia of Philosophy (Fall 2021 Edition), ed. Edward N.  Zalta. Accessed June 30, 2022. https://plato.stanford.edu/ archives/fall2021/entries/collective-­intentionality/. Sellars, Wilfred. 1980. On Reasoning About Values. American Philosophical Quarterly 17 (2): 81–101. Tuomela, Raimo. 2000. Cooperation: A Philosophical Study. Dordrecht: Kluwer. ———. 2002. The Philosophy of Social Practices: A Collective Acceptance View. Cambridge: Cambridge University Press.

218 

B. Petersson

———. 2007. The Philosophy of Sociality: The Shared Point of View. Oxford: Oxford U P. ———. 2013. Social Ontology: The Shared Point of View. New  York: Oxford University Press. ———. 2017. The Limits of Groups. Social Epistemology Review and Reply Collective 6 (11): 28–33. https://wp.me/p1Bfg0-­3QM.

11 Tuomela on Social Norms and Group-­Social Normativity Olle Blomberg

1 Introduction Imagine that we are at a semi-formal dinner with friends and acquaintances. During the dinner, you start chewing loudly with your mouth wide open. In the circles in which I socialize and am familiar with, your behaviour would violate a social norm of dinner etiquette. Assuming that we and our friends and acquaintances are members of such a circle, they and I could, and probably would, hold you accountable with respect to your violation of a social norm requiring, other things being equal, that one chews quietly with one’s mouth closed at a social dinner. It is unacceptable for you to chew loudly and with your mouth wide open, unless there is some special reason for you to do so. Without such a special reason, I and the other dinner guests are likely to at least frown and distance ourselves from you, but we may also challenge you and demand that you

O. Blomberg (*) Department of Philosophy, Linguistics and Theory of Science, University of Gothenburg, Gothenburg, Sweden e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_11

219

220 

O. Blomberg

behave (this is to be expected, in particular, from your close friends or a partner). That is, we would hold you accountable for the breach and demand that you conform.1 On the face it, we would not be holding you morally accountable here, since the norm of chewing quietly with your mouth closed is a not a moral norm that ought to be observed irrespectively of what the local culture and expectations are. Suppose that you responded to our attempts to hold you responsible by claiming that you had done nothing wrong, or if you with puzzlement in your voice asked us why you ought not chew loudly and open-mouthed. In answer to this, we would not justify our criticism by appeal to moral considerations. Rather, we would simply point out that you were violating our—and your—social norm. We might exclaim: “We are having dinner together for goodness sake!”, “You just don’t do that!”, or “That is not the way to eat!”. We might answer by in some way referring to respect—“You are behaving disrespectfully!”; “Show some respect!”—and this may seem like a moral complaint (see Southwood 2011, 787; Valentini 2021). However, if you then asked us why the particular norm-violating behaviour is disrespectful, the explanation would have to appeal to the local social norm, that is, to the sort of behaviour to be expected in the semi-formal dinner setting. At any rate, it is not clear that disrespect for other individuals can fully justify demands on you to conform. Suppose that the other dinner guests have temporarily left the table to smoke at the balcony. Here, you arguably still have a pro tanto reason not to chew loudly and open-­ mouthed, even if no one besides yourself would notice (cf. Valentini 2021, 388–389). I also find it plausible that others may justifiably demand that you act in accordance with “our” social norm even if that social norm itself is morally problematic or unjustified. Consider, for example, a norm “that men ought to open doors for women” (Tuomela 1995, 13). Such a norm, previously widespread, at least partly reflects a morally problematic inegalitarian social order. Nevertheless, if I identify as a member of a group where this social norm holds sway—say, the group whose members consider themselves “gentlemen”—then  other  Such responses of holding each other accountable in response to breaches of what is considered normal behaviour is often the result of so-called “breaching experiments” (see Garfinkel 1964). 1

11  Tuomela on Social Norms and Group-Social Normativity 

221

group members could arguably nevertheless justifiably demand that I conform to this norm.2 This could arguably be the case even if I and other group members are rightly criticizing the social norm or the unjust social order it reflects. This is a reason for thinking that appropriate demands to conform to social norms are not primarily justified by moral considerations. These reflections suggest that the normativity of social norms is a distinct “group-social normativity” (Tuomela 2007, 27, 2013, 248). In this chapter, I will examine Raimo Tuomela’s earlier and later views of social norms and group-social normativity. Such an examination is interesting for several reasons. First, Tuomela’s work on group-social normativity has not been given much attention (but see Corlett and Strobel 2017; Schweikard 2017). Secondly, his view of social norms and their normativity has made an interesting shift from a reductionistic and individualistic analysis (in 1995) to an irreducibly collectivistic analysis (in 2007). While I can only speculate about why Tuomela’s view shifted in this way, a reasonable hypothesis is that he concluded that a reductionistic and individualistic account could not make sense of our practice of holding each other to account for social norm violations (as well as make sense of the demands and obligations present in acting together and other social phenomena). Examining Tuomela’s early and later views can thus highlight some of the strengths and weaknesses of a non-reductionistic and collectivistic approach to social norms and related phenomena such as shared intention and acting together. In reflecting on the kind of examples of social-norm breaches that I have brought up, I find the idea that there is a distinct kind of group-­ social normativity appealing. At any rate, I will take for granted here that there is a kind of group-social normativity involved at least in social norms, and then examine how Tuomela’s individualistic and collectivistic accounts can make sense of it. I leave aside whether the same or a similar kind of normativity is involved in mundane cases of acting together. I believe that an individualistic account gives the right socio-psychological story of shared intention and joint intentional action, where I take the  Admittedly, I find such an intuition hard to sustain when it comes to more morally outrageous social norms, such as a social norm proscribing female genital mutilation. 2

222 

O. Blomberg

latter to be equivalent to intentional cooperation toward a shared a goal. The account I have defended is primarily a modified version of Michael Bratman’s account (see Blomberg 2016a, b), according to which a shared intention is a pattern of ordinary individual intentions and beliefs that is supposed to appropriately cause and coordinate the joint activity. On this sort of account, shared intention and acting together do not constitutively involve a distinct kind of group-social normativity, nor need it give rise to moral entitlements and obligations among participants. However, it is not obvious that reductionist accounts and irreducibly collectivistic accounts such as Tuomela’s are targeting the same phenomenon. Tuomela himself acknowledged this since he believed that there is both I-mode and We-mode shared intention and acting together. The plan for this chapter is as follows: In Sect. 2, I examine Tuomela’s early individualistic account of social norms and compare it with a similar account presented by Geoffrey Brennan et al. (2013). In light of the problems that these accounts face in making sense of how group members can be justified in holding each other accountable with respect to social norms, I go on, in Sect. 3, to discuss the recent intention-based accounts of social norms and their normativity offered by Laura Valentini (2021) and Michael Bratman (2022). I contend that while these accounts are promising in some ways, they cannot make sense of group-social normativity. Individualistic accounts cannot explain why group members who are not personally committed to their community’s social norm would nevertheless have a pro tanto normative reason to comply with the norm. In Sect. 4, I then go on to present Tuomela’s more recent irreducibly collectivistic account of social norms, and explain how it has the resources to make sense of group-social normativity. Finally, I wrap things up in the conclusions and briefly discuss the benefits and costs of such a collectivistic account.

11  Tuomela on Social Norms and Group-Social Normativity 

223

2 Tuomela’s Early I-mode Normative Attitude Account Tuomela starts out his book The Importance of Us (1995) by giving an account of social norms, where he distinguishes between what he calls “proper” social norms and social rules created by formal or informal agreements, such as “laws, statutes, regulations, charters, by-laws, and so on.” (1995, 17). Proper social norms—from now on, simply “social norms”—are not agreement-based though, but rather mutual belief-­ based (ibid., 20). Indeed, they “are mutual beliefs of a certain kind.” (ibid., 27). However, they are not simply mutual beliefs about some behavioural regularity. The account is supposed to rule out as social norms “social practices such as baking ham for Christmas or going to sauna every Saturday night” (ibid., 22). Rather, it seems fair to say that on Tuomela’s view, to co-opt a phrase from Brennan et al., social norms are mutual beliefs about “clusters of normative attitudes” (2013, 29, emphasis in original). On Tuomela’s early view, social norms “are centrally based on mutually shared beliefs about what one is normatively expected—by one’s group members—to do, on the penalty of social sanctions (approval, disapproval)” (1995, 18). I think Tuomela would agree with the contention by Brennan et  al. that the core function of norms is “to make us accountable to one another” (2013, 36, emphasis in original).3 Tuomela does not give a wealth of examples of social norms. Besides the example of the norm that men open the doors for women that I have already mentioned, Tuomela also gives the example of a social norm specific to a certain religious sect that members of this sect ought to kneel when praying (1995, 13).4 According to Tuomela (1995), the following sort of mutual beliefs constitute social norms:

 And: “What accountability involves is others having a recognized right or entitlement to determine how one is to behave.” (Brennan et al. 2013, 36) 4  Both Tuomela (1995, 14), Brennan et  al. (2013, 29), as well as Gilbert (1999) and Bratman (2022, ch. 3), refer to H. L. A. Hart’s (1961) idea of a social rule is being (more or less) the target phenomenon. 3

224 

O. Blomberg

(SN) A norm N of the form ‘Everyone in G ought to perform task T when in situation S’ is a proper social ought-to-do norm (or an ought-to-do s-norm) motivationally in force in G if and only if [Acceptance condition:] (a) there is a mutual belief in G to the effect that the members of G ought to perform T in situation S; [Pervasiveness condition:] (b) (i) many members of G perform T when in S (or at least are so disposed), and [Normative motivational condition: b)] (ii) at least some of them sometimes perform T at least in part because of their believing that they ought to perform T in S, and are, in accordance with (a), expected by other members of G to perform T in S; and there is a mutual belief approximately to the effect that (i) and (ii) in G; [Recognized sanction condition:] (c) there is in G some pressure, at least in part due to social sanctions, against deviating from performing T in S; and there is mutual belief to this effect in (and this mutually recognized pressure tends in part to account for (b)). (Tuomela 1995, 23–24, the labels for the conditions are Tuomela’s own, see p. 24)

The acceptance condition requires that every member of G believes that they and the other G-members ought to perform an action T in situation S, but this is arguably too strong as it would rule out the possibility that a social norm can be motivationally in force in a group even if there are some dissenting group members who do not share this belief (see Tuomela 2007, 205). In line with this, Tuomela (1995, 25–26) suggests that perhaps it is better to think of (SN) as giving graded conditions that determine the extent to which N is motivationally in force in G. Indeed, the account allows for the limiting case of “norms obeyed by nobody” (Tuomela 1995, 26). When it comes to the pervasiveness condition, Tuomela does not think that a social norm necessarily has to be followed—a norm is thus not centrally a behavioural regularity—even if most social norms do correspond to behavioural regularities. A case offered by Brennan et al. (2013) as a counterexample against the view that a norm is a kind of behavioural regularity is illustrative: Consider a group in which it is mutually believed that one ought to avoid urinating in public swimming pools, but contrary to common belief, everyone does it, despite having a (weak) disposition to avoid it. If it happens to come to light that someone has urinated

11  Tuomela on Social Norms and Group-Social Normativity 

225

in a swimming pool, people are indignant and the perpetrator will feel guilty, will tend to apologize, and so on. The Normative Motivational condition is supposed to impose a requirement that individuals perform T “for the right reason, viz., because ‘that’s the norm.’” (Tuomela 1995, 24; see also 2007, 214) Each individual’s belief that “one ought to perform T in S is, at least in part, grounded in the agent’s belief that he is normatively expected (required) by the others to act thus, that there is mutual belief about this, and that s[ocial]-sanctions will (otherwise) be applied (about which there is mutual belief ) […]. (1995, 192). As I interpret Tuomela, it is thus the group members’ view of the justification of the belief that one ought to perform T in S which makes the principle a social norm rather than, say, a moral norm. Tuomela’s account thus embodies what Brennan et al. (2013, ch. 4) call “the Grounds view” regarding what distinguishes social and moral norms (see also Southwood 2011). This appropriately captures the kind of responses that one can expect when challenging why a social norm ought to be followed, as I illustrated in the Introduction. As Brennan et al. (2013, 70) argue, it is important that in order for N to be a social norm (rather than a moral norm), the G-members must think that their belief that they ought to perform T in S is at least in part justified non-derivatively by the presumed shared disposition or behavioural regularity to perform T in S. If the G-members believe that, ultimately, one ought not to chew loudly with an open mouth because the noise risks distracting or damaging the hearing of one’s fellow diners, then chewing quietly with a closed mouth is not a social norm in G but rather a moral norm. This also seems to be what Tuomela has in mind: the right reason for complying is supposed, after all, to be “because ‘that’s the norm’” (Tuomela 1995, 24). Because of the qualification “at least in part” in the normative motivational condition, Tuomela’s account does not explicitly exclude that some social norms are (also) moral norms. The norm of the form ‘Everyone in Sweden ought not to lie under oath in court’ might (hopefully) be both a social norm and a moral norm. People might believe that they ought not to lie, after all, partly because it is the morally right thing to do, partly because they believe that others expect them not to lie and that they will be socially (as well as legally) sanctioned.

226 

O. Blomberg

The conditions in Tuomela’s account are restricted to concern members of the group G. Hence, the account is meant to explain why it makes sense for group members to hold each other, but not outsiders, accountable with respect to deviations from the norm. If you are a foreigner from a culture where it is customary to chew loudly and with one’s mouth open, then it would not be appropriate for me or the other dinner guests to demand of you that you conform to our—but not your—social norm. It may even be impolite to point out and inform you about the social norm that your behaviour is not in conformity with, although whether this is so will depend on circumstances. With respect to capturing this difference between insiders and outsiders, let me note an advantage of Tuomela’s account compared to the account offered by Brennan et  al., who agree that group members are normatively bound by social norms in a way that outsiders are not (see Brennan et al. 2013, 77–78; Southwood 2011, 787). Brennan et al. give the following account of social norms: A normative principle P is a [social] norm within a group G if and only if: (i) A significant proportion of the members of G have P-corresponding normative attitudes [e.g., a judgement that individuals must not loudly chew with open mouths]; and (ii) A significant proportion of the members of G know that a significant proportion of the members of G have P-corresponding normative attitudes. [(iii) A significant proportion of the members of G think that the P-corresponding normative attitudes are non-derivatively justified partly by a (presumed) P-corresponding behaviour regularity in G] (Brennan et al. 2013, 29).5

Conditions (i) and (ii) are together similar to Tuomela’s acceptance condition, and (iii) partly overlaps with Tuomela’s normative motivational condition and recognized sanctions conditions. However, because (i) and (ii) are formulated to cover norms in general, including moral norms, they do not quite get things right, because the normative  Brennan et al. (2013) only explicitly set out conditions (i) and (ii) for norms in general, where these include both social and moral norms. I have added condition (iii), distilled from chapter 4 of their book, to get their view of social norms in terms of necessary and sufficient conditions. 5

11  Tuomela on Social Norms and Group-Social Normativity 

227

attitudes are not required to be about what the G-members ought to do, unlike in Tuomela’s acceptance condition. Furthermore, it is not clear that the conditions require the G-members to take the normative attitudes to be justified in part by a presumed behavioural regularity that is explained by their memberships in G. Consider the case where Adam is a member of family A, Beth a member of family B, and Clara is a member of family C. In each family, there is a social norm to put the toilet seat lid down after use. Suppose that this is common knowledge among Adam, Beth and Clara. If Adam, Beth and Clara go on vacation together, renting an apartment in a ski resort. Is there a social norm among them to put the toilet seat lid down after use? Arguably, the answer is “no”. That they ought to put the toilet seat lid down after use in the apartment is not a normative principle that they can hold each other accountable with respect to (unless there are moral reasons to put the toilet seat lid down after use, but then the norm would be a moral norm). However, this is not the verdict delivered by conditions (i)–(iii) in Brennan et al. account. Tuomela’s account does not have this problem since it requires the G-members to believe that they, the G-members, ought to perform T in S.

2.1 The Problem of Justification While I think Tuomela’s account from The Importance of Us (1995) sheds light on and seems to be able to explain much of the behavioural facts of the phenomenon of social norms and our practice of holding each other accountable with respect to them, it faces problems in making this practice of accountability intelligible and robust enough to withstand scrutiny of rational and reflective group members. How, after all, does a presumed shared disposition or behavioural regularity in the group non-­ derivatively justify the members’ beliefs that they ought to act in accordance with it? Why would the fact that the members of my group believe that we all ought to chew our food at social dinners in a certain way in itself give me a normative reason to act accordingly? Given capacities for rational reflection and critical reasoning, it cannot just be a brute psychological fact that it will appear to group members as if there is such a justificatory relation. Without a plausible justificatory story, the normative

228 

O. Blomberg

beliefs required by the acceptance condition would arguably be very unstable and the social norms therefore very fragile. Of course, social norms do occasionally disappear or change rapidly, and perhaps a purely psychological “error theory” of the normativity of social norms will turn out the best we can get in the end. However, it would be premature to embrace such an error theory before vindicatory theories have been considered properly. In trying to respond to this problem of justification, Brennan et  al. consider the possibility of a distinctive kind of agent-relative value at play in social practices, the proper acknowledgement of which requires moulding one’s conduct in light of practices that play a certain role in one’s life. […] Acknowledging the value of social practices […] is a matter of recognizing and acting on reasons we have to honour the practices from the inside. (2013, 79–80) (see also Southwood 2011, 790; cf. Manne 2013).6

This might work as part of a story of how fully committed group members can be bound by social norms that (are presumed to) regulate social practices. However, it is not clear how it can make sense of how a presumed shared disposition or behavioural regularity could justify the belief that even a dissenting group member, someone who happens not to themselves share the belief that G-members should perform T in S, ought to perform T in S. Consider the man who identifies as a gentleman, but who is at the same time critical of the patriarchal social order that the social norm that men ought to hold open the door for women is at least partly a reflection of. How can we make sense of other group members sanctioning him for violating this norm, even though they hardly think that the norm is backed up by moral considerations? (Cf. Helm 2017, 216, 233–234). If the feminist gentleman does not share the belief that men ought to hold doors open for women, why should he “honour the practice from the inside”? The individualistic accounts of social norms that are on the table arguably lack the resources to explain this. Brennan 6  Gilbert (1999) refers to the justification problem as the “the grounding problem” (157) in a readable and insightful discussion of issues similar to those I discuss here, but in relation to her own account of social norms rather than Tuomela’s.

11  Tuomela on Social Norms and Group-Social Normativity 

229

et al. refer to how “practices may come to represent for one aspects of a valuable identity” for group members (2013, 70), but arguably, this means that the valuable identity cannot simply consist in various clusters of normative attitudes (and mutual beliefs about them).

3 I-mode Intention-based Accounts One way of partly avoiding the problem of justification is to give a different kind of reductionist account, according to which social norms are not clusters of normative attitudes but rather clusters of commitments or intentions. Such views have recently been defended by Valentini (2021) and Bratman (2022). A commitment or an intention to perform T does not generally give me a reason to perform T; it does not in itself provide the agent with an ‘ought’, and is thus arguably not a normative attitude. If it did, it would generate bootstrapping problems (see Bratman 1987). If I plump for buying one of two equally good detergents in the grocery store, thus forming an intention to buy it, then I do not create a new reason that favours buying that detergent rather than the other. Similarly, I do not create a new pro tanto reason in favour of torturing someone simply by deciding to torture them. According to one influential view, an intention rather imposes a rational requirement on how the agent’s attitudes can be combined (see Broome 2013). If I intend to perform T, then I am criticizably irrational unless I either go on to perform T (sometime in the future) given the means and opportunity to do so, or else (at some point) drop my intention to perform T. Valentini suggests that social norms are underpinned by individuals’ robust intentions that a requirement “function[s] as a standard of behaviour” in a certain type of situation (2021, 386), where the requirement might be, for example, that one should chew quietly with one’s mouth closed in the context of a social dinner. Indeed, she suggests that social norms may (but need not) be underpinned by what Bratman (2014) calls shared intentions (Valentini 2021, 398). This would mean, among other things, that everyone in the group G share an intention that chewing quietly with a closed mouth functions as a standard of behaviour, or to put it differently, that “we” conform to this standard. Furthermore,

230 

O. Blomberg

simplifying greatly, for there to be such a shared intention-based norm that ‘Everyone in G ought to perform T when in S’, there must be common knowledge in a group G that the members each intend that they perform T in S, each intends that they perform T in S by way of these intentions of each being effective, and each intends that they do this by way of sub-plans that are co-realizable (see Bratman 2014, 2022, ch. 3). If the social norm is underpinned by such a shared intention, there would thus be a shared rational requirement that can justify a kind of criticism between group members. There will be rational pressure among them to stick to their intentions that make up the norm, and they can be criticized for being irrational when they fail to conform to the norm while continuing to partake in the shared intention. Furthermore, members who drop their intention and thereby abandon their participation in the shared intention are also likely be deemed unreliable and may suffer from bad reputation (Bratman 2022, 51–55; cf. Tuomela 1995, 14). However, as Bratman points out, this falls short of justifying demands for compliance (see 2022, 55–56). In addition, it seems that not even criticism directed at a group member’s rationality or reliability will be appropriate when it comes to dissenting group members who perhaps never fully participated in the social norm. It is unclear what to say, for example, about the criticism that fellow group members direct against the feminist gentlemen who does not hold the door open to a woman. Valentini does not propose that her intention-based model captures the normativity of social norms on its own, but only when combined with a “pro tanto duty to give agency respect to others—to respect their permissible and genuine commitments” (2021, 394). She argues that in order to appropriately respect persons and their moral value, we must respect them as particular individuals who exercise their agency in concrete ways through their freely made decisions and commitments. Such agency respect can require of us to constrain our behaviour to accommodate the commitments of others. For example, I might have a pro tanto duty to “schedule a meeting with you one hour later than usual, so that you can attend a church service and thereby honour your religious commitments” (ibid., 391). If social norms are constituted by individual intentions that a certain requirement functions as a standard of behaviour, then it follows that others, irrespectively of whether they also have

11  Tuomela on Social Norms and Group-Social Normativity 

231

an intention with this content, have a pro tanto duty to comply with the social norm, at least given that  it is a  morally permissible  norm. You would thus have a pro tanto duty to comply with the requirement that you chew silently with your mouth closed at the dinner, and the feminist gentleman would have a pro tanto duty to comply with the norm of holding doors open to women. The reasons favouring violation of the norm may win out, but there would then be remedial duties remaining when a social norm is violated (the norm-violator may have a duty to acknowledge and explain why the norm was violated, for example). Valentini’s account is interesting, but it does not solve the justification problem. It is one thing to claim that I have a pro tanto duty to adapt my behaviour in accordance with your self-regarding commitment (such as your commitment to attend church or kneel when praying); it is another thing to claim that I have a pro tanto duty to adapt to your commitment that we (including me) comply with a standard of behaviour. In an attempt to respond to something like this worry, Valentini observes that “[t]here appears to be nothing intrinsically objectionable about the other-­ regarding nature of queuing and tipping norms. Therefore, respect for the commitments of the agents who support them requires respect for the norms themselves” (2021, 395). While I agree that these norms are not objectionable, the challenge is explaining why they are not objectionable, and how even those who are not personally committed to queuing and tipping functioning as a behavioural standard nevertheless seem to be bound by them. I fail to see how respect for others’ agency helps justify demands of compliance nor do I see how it helps explain why even a dissenting group member would have a pro tanto reason to respect the others’ intentions that everyone conforms to a certain standard of behaviour. It is also worth noting that Valentini’s account does not distinguish between how a social norm binds group members and how it binds outsiders. While Valentini thinks that this is an advantage of her account (see 2021, 400–401), it does not help to make sense of something like Tuomela’s group-social normativity. Perhaps a modified version of Valentini’s account can explain why individuals in general have a pro tanto reason to comply with morally permissible social norms, but it does not explain the intuitively special bind that social norms have on group members in particular.

232 

O. Blomberg

If there are pro tanto reasons to comply with social norms which are tied to group membership then the individualistic accounts of Tuomela (1995), Brennan et  al. (2013), Valentini (2021) and Bratman (2022) cannot fully make sense of such reasons and of our practice of holding each other accountable with respect to our compliance with our social norms. This, it seems to me, is the best reason for exploring whether a non-reductionistic collectivist account might do better. Perhaps some holistic medicine can help.

4 Tuomela’s Irreducibly Collectivistic We-Mode Account A few years after The Importance of Us (1995), Tuomela’s views of human sociality changed significantly.7 He introduced a distinction between acting in the “I-mode” and acting in the “we-mode”, where the latter is associated with individuals acting as group members as opposed to acting as private individuals. When acting as group members, individuals are together collectively committed to acting as if they were one agent, and they are acting for the group’s reasons. The concepts of collective commitment and group reason should here be understood in an irreducible collective way according to Tuomela—that is, they cannot be analysed reductively in terms of interrelated individual commitments and ordinary individual reasons for belief and action. When it comes to some social phenomena, such as acting together, the phenomenon exists both in an I-mode version and in a we-mode version. Agents can act on I-mode shared intentions or on we-mode joint intentions (see e.g. Tuomela 2007, 72–73), where the former but not the latter can be understood solely in terms of concepts already needed for making sense of individual agents and their strategic interaction. However, as far as I can tell, this is not Tuomela’s later view of social norms. Social norms exclusively appear as we-mode phenomena it seems. When he presents an account of social norms in The Philosophy of Sociality (2007), he takes a  As far as I can tell, the change occurred shortly before the turn of the millennium. The I-mode/we-­ mode distinction is at least present in (Tuomela 2000). 7

11  Tuomela on Social Norms and Group-Social Normativity 

233

social norm to be a we-mode phenomenon that requires collective acceptance of the norm among group members. While I do not know why he changed his mind about this (the shift, as far as I can tell, is not something he comments on), I think the best argument for such a change of mind would make reference to the difficulties of making sense of our practice of social-norm accountability that I have highlighted in the previous two sections. Others who have argued for collectivistic accounts of the group-social normativity of social norms—such as Gilbert (1999) and Helm (2017)—have more explicitly appealed to such difficulties. However, my focus here is on Tuomela’s views. Here is Tuomela’s later collectivistic account, which is deceptively similar to his earlier account, but is in fact quite different (I have bolded phrases that indicate the substantive differences between this account and the earlier one): (SN) A norm N expressible by ‘Everyone in G ought to perform task T when in circumstances S’ is a proper social ought-to-do norm (or an ought-to­do s-norm) motivationally in force in social group G if and only if [Acceptance condition:] (1) the members of G implicitly collectively accept or tend so to accept—and are disposed to mutually believe to the effect— that the members of G ought to perform T in S; [Pervasiveness condition:] (2i) many members of G perform T when in S (or at least are so disposed), and [Normative motivational condition:] (2ii) at least some of them sometimes perform T, at least in part, because of their believing that they ought to perform T in S, because it is their group’s norm and also in part because they are, in accordance with (1), normatively and factually expected by other members of G to perform T in S; and (2iii) there is a mutual belief approximately to the effect that (2i) and (2ii) in G; [Recognized sanction condition:] (3) there is in G some group-social pressure, at least in part due to social sanctions (members’ disapproval for deviation and approval for obedience), against deviating from performing T in S; and there is a mutual belief to this effect in G (and this mutually recognized pressure may in part account for both (2i) and (2ii)). (Tuomela 2007, 214, I have changed the variable letters to align them with the earlier account)

234 

O. Blomberg

The most important difference here is in the acceptance condition. Tuomela now requires that the G-members, “or at least a substantive part of them” (214), collectively accept that they ought to perform T in S— otherwise the norm is not motivationally in force in the group. What is it (for a substantive part of ) them to collectively accept that they ought to perform T in S? Roughly, it is for them to agree that all G-members act as if they were one (group) agent that believed that they, all its members, ought to perform T in S.8 This collective acceptance need only be implicit or tacit, where individuals’ participation in such an acceptance may perhaps simply be signalled by them behaving in accordance with the norm or its associated sanctioning behaviour. For Tuomela, this entails that the G-members have a we-mode joint intention and a collective commitment to “satisfying and maintaining the norm” (2007, 214). This we-­ mode joint intention gives them, as group members, a group reason to act in accordance with it by doing their parts of satisfying and maintaining the norm. Just like the concepts of group reason and collective commitment, the concept of joint intention is, according to Tuomela, an irreducibly collectivistic concept associated with individuals thinking and acting as group members. So, on Tuomela’s later account of social norms: A proper social norm in this sense is a we-mode norm due to the fact of having been collectively accepted by the members (or at least a substantive part of them) as their norm. […] The relevant group reason for the members’ doing T in S in the case of both kinds of norms is that it is a group-­ social norm in G (“our norm”) to act so. (Tuomela 2007, 214, variable letters changed)  Gilbert (1999) provides a similar collectivistic account of social norms that she succinctly summarizes as follows: “There is a social rule [norm] if and only if the members of some [group G] are jointly committed to accepting as a body a requirement of the following form: members of [G] are to do A in C.” (163; see also Gilbert 1989, ch. 6) There are many similarities between Tuomela’s account of collective acceptance and Gilbert’s account of the psychological process that leads to her joint commitments. The main difference, it seems to me, is that on Tuomela’s account, a collective acceptance entails a joint intention, which constitutively involves each member being in a certain mental state, namely being in a state of we-intending to do one’s part. For Gilbert, a joint commitment has normative implications for the parties (that is, certain directed obligations and entitlements), but once the joint commitment has been made, nothing in particular is entailed about the mental states of the parties. (See Tuomela 2007, 121–122, 257–258 endnote 35, and 260 endnote 53 for comments on Gilbert’s work; see Gilbert 1998, for comments on Tuomela’s [1995] view of the normativity of acting together). 8

11  Tuomela on Social Norms and Group-Social Normativity 

235

In other words, Tuomela now explicitly sees joint intention and social norms as part of the same phenomenon, whereas he earlier (in 1995) explicitly distinguished social norms, which were mutual belief-based, from joint intention and joint intentional action, which were agreement-­ based. Note though, that Tuomela allows that individuals perform T in S for private I-mode reasons, in which case they conform to the norm without “obeying” it (see 2007, 205–206). Nevertheless, even the members who act on private I-mode reasons still have, I take it, group reasons to comply with the norm by virtue of their implicit collective acceptance, and these group reasons could at least potentially also be their motivating reasons for conforming (and thus, “obeying”). This account, if otherwise acceptable, can help address the problem of justification that individualistic accounts face. In particular, it can help with the problem of justifying why dissenting group members can have a pro tanto reason to comply with a social norm. The key here is that there is a discontinuity between group-level (“the group’s norm”, “group-social pressure”) and the individual level. On Tuomela’s view, each of the members of a group who together have a we-mode joint intention has a “we-­ intention”, where a we-intention is each individual’s “slice” of the joint intention (2007, 93, 2013, 63, 78). However, this we-intention is conceptually and ontologically dependent on the joint intention. Thus, I cannot alone have a we-intention. At most, I can falsely believe that I have a we-intention. This means that a we-intention cannot be identified with some cluster or pattern of ordinary individual intentions and beliefs (Tuomela 2013, 79). As I have pointed out, a member can have a we-­ intention but nevertheless act on an individual intention and be motivated by private I-mode reasons. In light of this, I take it to be possible on Tuomela’s view that an individual can also personally intend to act contrary to what is proscribed by a social norm, even though she has a (standing) we-intention to do her part in maintaining norm-compliance among group members, including herself (see Tuomela 2013, 72–73). The feminist gentleman who has implicitly taken part in collectively accepting that gentlemen ought to hold doors open for women (in certain situations) can thus have a we-intention to do his part of ensuring that this social norm is satisfied and maintained, while being privately committed to violating this social norm (perhaps he is also privately

236 

O. Blomberg

committed to change the group’s norm). This, I think, is a genuine possibility. By identifying as a gentleman, by largely behaving as one might expect gentlemen to behave, and by refraining from sanctioning other gentlemen for conforming to the norm, it seems plausible that he would take part in the implicit collective acceptance of the social norm. The norm is an aspect of a valuable identity for him, although he does not value that particular aspect itself. But why would the collective acceptance (including the feminist gentleman’s slice of it) have a bearing on his conduct related to holding doors open to women, given that he has personally decided not to conform to this particular social norm? Here, there seem to be two different lines of thought in Tuomela’s writings. On the one hand, there is the suggestion that the we-mode joint intention or collective commitment has authority over the members because they are prepared to defer to an authoritative group reason, that is, to treat the joint intention as giving the members reason to exclude acting on other reasons than the group’s reasons (see Tuomela 2007, 130, 2013, 120). However, this arguably just pushes the puzzle to a new location. Perhaps it is right that a group reason gives members reason to exclude acting on other conflicting private reasons, but why think that there is a group reason at all? After all, as was noted in the previous section, an intention to Φ does not generally give me a reason to Φ, so why would a joint intention to Φ give the group a group reason to Φ? Well, the idea is that while a joint intention to Φ does not give the group as such a reason to Φ, it can in many cases give individual group members reasons to do their part of the group’s Φ-ing (see Tuomela 2007, 32–33, 130–131; cf. Gilbert 2018, ch. 8). We can put the relevant group-level rationality requirement associated with a joint intention in something like the following form: Group-level rationality requires [if the G-members jointly intend to Φ and the G-members mutually believe that it is necessary for each G-member to do her part of the G-members’ joint Φ-ing for the G-members to jointly Φ, then the G-members jointly intend that each G-member does her part of G’s Φ-ing].

11  Tuomela on Social Norms and Group-Social Normativity 

237

This requirement is a “wide scope” requirement. It does not require that the G-members jointly intend that each G-member does her part of their Φ-ing, such as her part of conforming to and maintaining a group norm, for example. Instead, it requires either that the G-members jointly intend this, or that they drop their joint intention that the group Φ. In other words, they are together required to either ensure that they act in conformity with the social norm or ensure that the social norm stops being motivationally in force among them. However, from the point of view of each individual group member, doing one’s part of satisfying this group-level rationality requirement will typically require that one we-­ intends to do one’s part of the group’s Φ-ing rather than that one tries to influence the group to drop the joint intention to Φ. Given that the pervasiveness condition is satisfied so that there is relatively widespread conformity with the social norm, the only possible way in which the individual group member can now contribute to satisfying the group-­ level rationality requirement will be by further contributing to the pervasiveness of the conformity, and this will be a matter of mutual belief among the group members. This means that there is a kind of “top-down” bootstrapping where a joint intention will give individual group members a joint intention-based reason, given the satisfaction of the pervasiveness condition, to contribute to the joint intention being satisfied.9 To illustrate, given the widespread conformity among gentlemen to the norm that they ought to hold doors open for women (the pervasiveness condition is to a large extent satisfied), the joint intention implied by the collective acceptance among gentlemen that they ought to hold doors open for women will give the dissenting feminist gentleman a pro tanto group-based reason to behave in conformity with the norm, despite his personal dissent and intention to act in violation of the norm. The account would also justify why other group members could demand that he comply in light of the relatively widespread conformity to the norm in the group. This seems to me to be a plausible interpretation of what Tuomela has in mind when he talks of group-social normativity.

 Tuomela talks of bootstrapping in the joint case which is not “harmful” (2007, 131). Perhaps this is what he had in mind. 9

238 

O. Blomberg

More importantly, the account provides an interesting solution to the problem of justification. However, the solution does create another philosophical challenge. This challenge can be put in terms of explaining what the metaphysical status of irreducibly collective we-mode joint intentions is. Such joint intentions are supposed to be “nonaggregative and emergent (‘creative’) relative to the members I-mode attitudes.” (Tuomela 2007, 131) How can we make sense of this group-level emergence given that the joint intention is supposed to be psychologically manifested in the individual group members’ mental states (their we-­ intentions)? What is essentially the same challenge (I suspect), can also be put in normative terms. How can we make sense of the idea that a group agent’s rational requirements put normative constraints on the members’ thinking and acting? Tuomela does not have much to say about this. The normative story is essentially the following: “The ‘oughts’ and ‘mays’ involved in contexts of acting as a group member are at bottom based on the concept of togetherness (‘groupness’). Thus, when acting together with others, one ought to do one’s part and one has the right to expect that others do theirs, and this normativity has a partial basis in the framework of group concepts—it belongs to the ‘conceptual logic’ of group concepts.” (Tuomela 2007, 27–28) This is suggestive but, at least if the metaphysical version of the challenge cannot be overcome, then a deeper explanation of how these group concepts give rise the group-social normativity would be needed.

5 Conclusion My aim in this chapter has been to take what appears to be an everyday phenomenon at face value—our practice of holding each other accountable for acting in accordance with “our” social norms—and then examine how Tuomela’s earlier individualistic and later collectivistic account of social norms could explain the “group-social normativity” presupposed by this practice. Tuomela’s individualistic account as well as intention-­ based individualistic accounts fail to make sense of how demands for social norm-compliance directed at dissenting group members could ever be justified. The possibility of making sense of this is arguably the main

11  Tuomela on Social Norms and Group-Social Normativity 

239

reason for considering an irreducibly collectivistic account of social norms, and perhaps also of other social phenomena. This is not an original line of thought. After all, the basis for Margaret Gilbert’s accounts of various social phenomena in terms of irreducibly collectivistic “joint commitments” between group members is the idea that these members have obligations and rights vis-à-vis each other that are grounded in a distinct kind of social normativity (see e.g. Gilbert 1989, 1999, 2018). The normative aspects of Tuomela’s work have not been given much attention, and I have tried to give an interpretation focusing on these normative aspects that might explain the transition from his earlier individualistic account of social norms to his later irreducibly collectivistic account. I have argued that Tuomela’s later collectivistic account makes better sense of our practice of holding each other accountable for acting in accordance with “our” social norms, when the judgements and demands that are part of this practice are accepted as, by and large, justified. I have not, however, argued that this makes his collectivistic account all-things-­ considered the best account of social norms and their normativity. Such an argument would require finding a wide reflective equilibrium that takes into account not only our considered normative judgements regarding cases involving the violation of social norms and our particular moral judgements and general moral principles, but also “relevant background theories” (Daniels 1996, 22), where these include our best account of the nature of our sociality and its socio-psychological underpinnings. Nevertheless, Tuomela’s collectivistic account of social norms can help us articulate what needs to be either explained or explained away by an account of social norms. Acknowledgements  Material in this chapter was presented at the research seminar in practical philosophy at the University of Gothenburg and at the Social Ontology & Collective Intentionality 2022 conference. I am grateful for helpful questions and comments from the audiences. My research was funded by the Lund Gothenburg Responsibility Project (PI: Paul Russell), which is in turn funded by the Swedish Research Council.

240 

O. Blomberg

References Blomberg, Olle. 2016a. Common Knowledge and Reductionism about Shared Agency. Australasian Journal of Philosophy 94 (2): 315–326. ———. 2016b. Shared Intention and the Doxastic Single End Condition. Philosophical Studies 173 (2): 351–372. Bratman, Michael. 1987. Intentions, Plans, and Practical Reason. Cambridge: Harvard University Press. ———. 2014. Shared Agency: A Planning Theory of Acting Together. Oxford: Oxford University Press. ———. 2022. Shared and Institutional Agency: Toward a Planning Theory of Human Practical Organization. Oxford: Oxford University Press. Brennan, Geoffrey, Lina Eriksson, Robert E. Goodin, and Nicholas Southwood. 2013. Explaining Norms. Oxford: Oxford University Press. Broome, John. 2013. Rationality Through Reasoning. Oxford: Wiley-Blackwell. Corlett, J.  Angelo, and Julia Lyons Strobel. 2017. Raimo Tuomela’s Social Ontology. Social Epistemology 31 (6): 557–571. Daniels, Norman. 1996. Justice and Justification: Reflective Equilibrium in Theory and Practice. Cambridge: Cambridge University Press. Garfinkel, Harold. 1964. Studies of the Routine Grounds of Everyday Activities. Social Problems 11 (3): 225–250. Gilbert, Margaret. 1989. On Social Facts. New York: Routledge. ———. 1998. Review of The Importance of Us: A Philosophical Study of Basic Social Notions by Raimo Tuomela. Ethics 108 (4): 811–812. ———. 1999. Social Rules: Some Problems for Hart’s Account, and an Alternative Proposal. Law and Philosophy 18 (2): 141–171. ———. 2018. Rights and Demands: A Foundational Inquiry. Oxford: Oxford University Press. Hart, H.L.A. 1961. The Concept of Law. Oxford: Oxford University Press. Helm, Bennett W. 2017. Communities of Respect: Grounding Responsibility, Authority and Dignity. Oxford: Oxford University Press. Manne, Kate. 2013. Being Social in Metaethics. In Oxford Studies in Metaethics: Volume 8, ed. Russ Shafer-Landau, 50–73. Oxford: Oxford University Press. Schweikard, David. 2017. Voluntary Groups, Noncompliance, and Conflicts of Reason: Tuomela on Acting as a Group-Member. In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with his Responses, ed. Gerhard Preyer and Georg Peter, 97–111. Cham: Springer.

11  Tuomela on Social Norms and Group-Social Normativity 

241

Southwood, Nicholas. 2011. The Moral/Conventional Distinction. Mind 120 (479): 761–802. Tuomela, Raimo. 1995. The Importance of Us: A Philosophical Study of Basic Social Notions. Stanford: Stanford University Press. ———. 2000. Cooperation: A Philosophical Study. Dordrecht: Kluwer. ———. 2007. The Philosophy of Sociality: The Shared Point of View. Oxford: Oxford University Press. ———. 2013. Social Ontology: Collective Intentionality and Group Agents. Oxford: Oxford University Press. Valentini, Laura. 2021. Respect for Persons and the Moral Force of Socially Constructed Norms. Noûs 55 (2): 385–408.

12 Cooperation Rests on Trust—But What Is Trust and How Do We get There? Maj Tuomela

1 Introduction Having worked with Raimo Tuomela since the late 1980s on his research in my spare time, I got an offer to start my own work on trust in 1995, when he had a five-year research period as a Finnish Academy Professor. I chose my topic based on what I deemed he would need next in his own project. Beliefs, commitment, and responsibility were already on the table in our group. However, trust is also relevant, not only in Acting together, but also in Sociality in general. I mostly wrote beside a full-time job and progressed slowly. I remember saying: “First we need to know what trust is, before we can use it”, and later: “don’t ask me to say when to trust people”. I only realized afterwards that my trust account actually answers the question. In 2004 I finally gave in to his wish to start the joint article he now eagerly awaited: “Cooperation and Trust in Group Context” (2005). I am pleased that we did not use the trust account before it was ready and that our joint article became much cited. Raimo’s way of working was to write many versions and continuously ask for comments. He

M. Tuomela (*) University of Helsinki, Helsinki, Finland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7_12

243

244 

M. Tuomela

was a true we-moder. As I was writing my own dissertation, I felt obliged to avoid his influence. What I owe him instead, is the careful, pedantic way of considering a text. Our work on Social Norms has the same features: It was my Master’s thesis that became a joint article as soon as I wrote the last sentence. That topic was his choice for me in the late 1980s. Our cooperation was joyful—he supplied the skill and knowledge, and my part was to check with the real world and ask basic questions, and then even more basic ones. I am deeply grateful for all that he taught me. Social norms as well as trust are truly at the very center of social life. I will turn to our joint work on Cooperation and Trust in Part II. Modern city-life is supposed to function guided by institutional rules and unwritten social norms. However, some people cheat when they can get away with it and do not care about social sanctions. Trusting people in urban communities is problematic, even in social-democratic societies of free education and social and health care. In times when masses of people are leaving their homes due to war and famine, strangers come to live and work among the natives. Mutual confusion about habits and norms arises. However, trusting refuges for being friendly and honest would at least be furthered by informal joint activities. Here charitable organizations backed up by governments could help with shared activities and meeting points. To better understand how to further trust, it would be good to know what social trust is. The notion of social trust as a topic of analysis and investigation is important in Philosophy and in the Social Sciences, as it plays a crucial role in social life and especially in cooperative activities. An existence without trust is almost impossible to imagine. In Part I, I will present an analysis of Social Trust that distinguishes two main types of trust, to be called Rational Social Normative Trust (RSNTR), in short, Genuine trust; and Rational Predictive Trust that is often considered to be real trust but is based on our own prediction. Trust will be seen from a subjective, first—person point of view, and concerns a person’s subjectively rational trust in another person vis-à-vis his performance of an action. The theory, and a required Trust context (Y), both consisting of necessary and sufficient conditions, are defended, as to their crucial parts. The account of Genuine trust is compared to the notion of Predictive “trust,” the latter claimed to be like a prediction that one “falls into,” but being far from Genuine trust that arises from a relationship of

12  Cooperation Rests on Trust—But What Is Trust and How… 

245

mutual respect for the other party’s right to be accommodated in this specific case. The expectation of having this right respected is grounded in mutual experiences, through time, of what each party is granted by the other. Themes like “subjective rationality” and “a basic moral We-perspective” are accounted for, while topics like “a collective’s trust in a collective”, “thick and thin genuine trust”, and “general trust” are just touched upon. In Part II, “Cooperation and Trust”, as of Raimo Tuomela and Maj Tuomela (2005) is briefly presented.

2 Part I: What Is Trust and How Does It Come About? 2.1 Social Normative Trust (Genuine Trust) versus Predictive “Trust” First, I will show how the two types of trust, Genuine trust and Predictive “trust,” may be distinguished from each other in practice. When trust is betrayed, the trustor had Genuine trust—her trust was justified, but the trustee did not respect her rights, although she had subjectively rational reasons to believe he would (they had a relationship of mutual respect for specific rights). On the other hand, when the trustor is mistaken in her “trust” on the grounds that her own predictions were poor, she is to blame. She had Predictive trust, based on her own incorrect beliefs about the trustee. In Genuine trust the betrayed trustor rightfully blames the trustee, unless the trustor was mistaken about there being a mutual understanding about respect for certain rights. In Predictive trust the disappointed trustor makes the wrong predictions and has herself to blame. Of course, she could accuse the trustee of deception, in case he deliberately gave her the false impression. Misplaced trust is an ambiguous term as it can mean betrayed justified Genuine trust, mistaken (unjustified) Genuine trust, or mistaken Predictive trust. I could end the presentation here, as this was one way of saying what trust is and what it is not. My best example will shed further light on the difference between Genuine trust and Predictive trust. I have Predictive trust in my partner to present me with a necklace for my birthday, because I believe he loves me, but I have Genuine trust in him to tell me the truth, as this is based on our

246 

M. Tuomela

relationship of mutual respect for each other’s rights, and to be told the truth is a mutual right. The belief of having such a right is based on our mutual experience of having such a right, but it need not be verbalized. Genuine trust may be betrayed by the trustee, while Predictive trust cannot be betrayed—it is the trustor who made the wrong prediction, leading him into having a trusting attitude, and perhaps into the state of relying on another person or to decide to rely on him. This latter kind of “trust” is further stretched in the context of Economics into a mere Prediction. The trustor calculates the probability for an action to take place, he decides to “trust,” and may have various degrees of “trust.” The trusted person might not even be aware that someone is counting on his action. In Predictive trust the attitude of trust is had without a decision, just as in Genuine trust, and the trustor either trusts, due to his calculations, or he does not trust. In Genuine trust there are no calculations or predictions, but a relationship of mutual respect for specific rights, based on shared experiences of what the rights are.

2.2 Rational Social Normative Trust, viz. Rational Genuine Trust Below, I will present my account of Rational genuine trust in the form of five jointly necessary and sufficient components of a conceptually and ontologically irreducible holistic notion. I will argue that this is what trust is, and how it can be theoretically analyzed, or elucidated. Despite the claimed irreducibility of the notion, these “components” are clearly discernable in it. If I had all the beliefs of the trust account, I would trust if I were a robot, but I claim that a human being “falls into” trust as she “falls into” love. This means she might not “fall into” trust, after all, which should make us suspect that there is more to trust than what my account will offer. My hunch is that if the “falling in trust” does not occur, despite all the conditions being seemingly fulfilled, the expected trustor has reason-based doubts, maybe on a subconscious level, affecting his Predictive trust, leading to expectations of “danger.” Thus, “this may be the day when the respected rights of yesterday” stop being valid. My account will stand as “a good as can be” elucidation of trust, even though genuine trust can rightfully come to an end (or never lift from the ground), despite

12  Cooperation Rests on Trust—But What Is Trust and How… 

247

a seemingly solid “charter of rights”. This “falling out of trust” (or never “falling in trust”) is a new observation concerning my account—intuition may take over before conscious beliefs are present. Next, I will present the three jointly necessary and sufficient context components, called “the Trust context (Y).” Real life, examples, practical reasoning, and scientific testing may contest the theory, but at present this is the offer. The account concerns trust vis-à-vis performance of an action, and not general trust. The validity of the theory of Rational genuine trust, as presented below, and of the criteria for distinguishing Genuine trust from Predictive “trust,” is argued by common sense reasoning, based on the author’s informed understanding of these notions. An evaluation of the present theory of trust must focus on the quality of reasoning used in the defense of the analysis of trust, and thus for the underlying view of trust. The present conception of trust goes against a view of trust as a kind of prediction, but purports to say more than just to define trust as a righteous demand that should be met.

2.3 The Theory of (RSNTR) Rational Social Normative Trust (Rational Genuine Trust) The Trust context (Y): The following three conditions are necessary and jointly sufficient for a context (Y), “the Trust context,” in which A can rationally trust B that he will perform a specific action X: a) A wants B to perform X, an action that A does not intend to perform himself. A thus believes that he is concerned with B’s performing X, and believes that he is in that sense dependent on B’s action; b) A has a belief about B’s belief of the dependence of A; and c) A believes that his dependence is “relevant” to B. That the dependence is relevant to B means here that B is free to let A’s dependence play a role in his decision-making, and that B will take A’s dependence into account in his actions. A’s beliefs in a)—c) may be dispositions to have such beliefs and they are based on (at least) subjectively rational reasons (or dispositions to have such).

248 

M. Tuomela

(RSNTR): For a person A to have Rational social normative trust (Rational genuine trust) in another person B that he will perform a specific action X, in a situation in which context Y is present, the following necessary and jointly sufficient conditions must hold: 1. Intention condition: A expects (or has the disposition to expect) that B intends to perform X. The expectation is based on (at least) subjectively rational reasons or dispositions1 to have such. 2. Ability condition: A expects (or has the disposition to expect) that it is possible for B to perform X (viz., that the internal and external opportunities for B’s action obtain). The expectation is based on (at least) subjectively rational reasons or dispositions to have such. 3. Goodwill condition: A expects (or has the disposition to expect) that B will intentionally act with goodwill, including goodwill towards A, when performing X. A believes (or has the disposition to believe) that due to their relationship of mutual respect for each other’s rights, he is entitled to expect this of B, on social normative or (quasi-) moral grounds, and that B acts at least in part because A has this right.2 These expectations (or dispositions) are based on (at least) subjectively rational reasons or dispositions to have such. The reasons for A’s expectation in 1. may be (partly) the same as here. 4. On the grounds of 1, 2 and 3, A has a “positive” feeling. This feeling is or, at least, includes the feeling of being comfortable about his dependence on B for performing X. 5. On the grounds of 1, 2, 3, and 4, A has an accepting attitude vis-à-vis his position of dependence on B for performing X.  This attitude is non-intentionally acquired. It is not difficult to argue for the context beliefs of the account, nor for the first two, or the last two clauses of the account.3 To save space, I will  See (R. Audi 1993) on unreflective beliefs and dispositions to form beliefs.  People have their own “charter of rights” that they mutually recognize and, this charter is a result of confirming experiences through time. Thus, they trust when they expect what they have a right to expect. Other expectations are based on trust as a prediction that they are convinced of, and do not belong here. 3  In M. Tuomela (2006b) I have some basic arguments for the necessity and joint sufficiency of the conditions of (RSNTR) and the trust context (Y) for a person’s Rational social normative trust in another person concerning a specific action. 1 2

12  Cooperation Rests on Trust—But What Is Trust and How… 

249

thus focus on the third belief that I call the goodwill belief.4 The first part of it is quite evident: we expect those we trust to act with goodwill towards us vis-à-vis the action in question. I claim that this expectation is grounded in the parties’ relationship of mutual respect for each other’s rights.5 6 7 8  Cf. (A. Baier 1995, p. 136; P. Pettit 1995, p. 206; D. Gambetta 1988, p. 217; K. Jones 1996; J. Dunn 1984, pp. 73-74). 5  R. Solomon (2000, p. 238) stresses the “dynamics” of the relationship, and here his view is close to mine. I have discussed the dynamics of the relationship in terms of mutual respect of rights, where the parties have to agree with each other about the rules of the game, viz. What are the norms that specify their mutual rights. 6  The notion of “right” is crucial in Genuine trust. We may expect gratification from others, having Predictive “trust” in them, concerning matters that we have no right to ask for. Then we just depend on features of the person and the situation. However, expecting to have one’s rights accommodated by another’s intentional action is yet not sufficient for Genuine trust. The respect of another’s right must be grounded in a relationship of mutual respect or else we just have a situation where rights are anticipated to be respected for some predictable reason, e.g., prudential reasons or moral standards. Then we are still in the area of Predictive trust. Many scholars define trust to involve rights. E.g., O. Lagerspetz (1998, p. 82) notes that “to describe a relationship as one of trust is to make a statement about what a person has a right to expect, even require from us” and L. Hertzberg (1988; 1994, p. 319) “[-] … unlike reliance, the grammar of trust involves a perspective of justice: trust can only concern that which one person can rightfully demand of another.” 7  P. Aykens (2002, pp. 3, 15) discusses the mutual commitment and obligation that develop when some persons interact over time. He holds that these are needed for affective trust. He notes that violations of well understood rules of the game can lead to mistrust. A. Wolfe (1989, p. 215) argues that people create their moral rules—that is, mutual obligations—through the social interactions they experience with others. B. Misztal (1996, p. 21) stresses that social relations and the obligations inherent in them are mainly responsible for the production of trust. She holds that people’s commonsense knowledge about what to expect from various types of social relations and their understanding of the motives of others is central in discussing motivations underlying trust relationships. 8  V. McGeer (2002, p. 37) argues convincingly that an evidence-based paradigm of rationality in the context of trust should give place for a more appropriate paradigm: “[-] … he (a mature agent) uses reason to respond well to particular others, conducting himself towards them in ways that invite them into a trusting relationship. He shows his awareness that their responsiveness is partly, though not entirely, contingent upon his ability to understand and respect their interests and desires as features of their agency, rather than merely as potential blocks to his own. And he looks for signs in their responses that they are willing to do the same towards him. Only in such concrete exchanges can an agent determine how far his trust is warranted, for these will show him to what extent he can use reason constructively, not just to negotiate the difficulties that inevitably arise within mature relationships, but also to expand his self-understanding as he strives to understand and move forward with particular others.” In McGeer’s rational trust the rationality does not lie in disengaged evaluation, but in the willingness to engage in a relationship, using reason constructively to respond well to the other. Her idea of mutual respect for interests and desires are close to my “mutual respect of rights” in that these rights are relationship-specific, viz. the parties have their own “rules of the game,” which are often a result of negotiation. Thus, my desire for certain (mutual) rights may be in conflict with yours and thus peaceful trusting co-existence requires that we find a way to come to terms with our differences. Internal norms that are reasonable to both parties make for mutual rights that they may reasonably expect to be respected. 4

250 

M. Tuomela

This is the (causal) reason “why” the trustor expects the good-willed action, not the reason “for.” “He will do this to gratify me,” should not be based on reasons, like “he loves me,” “he promised,” “he is a good person” in Genuine trust. That is what I call “Predictive trust,” not Genuine trust. This is what I do in Predictive trust: I trust someone to help me for some reason, e.g., the reason that he is a good person. I also believe that the beliefs in Context Y are in place (i.e., I depend on his help, he is aware of my predicament, and that he is going to consider the situation, and that he is free to help), and I believe he intends to help, that he is able to, and that he will act with goodwill towards me. This could be a situation where the trustor just counts on the goodness of somebody’s character. Another reason for Predictive trust could be social prestige, e.g., people will get to know, or social pressure, if not helping would go against social norms.9,10 In Rational genuine trust the trustee is expected to act with goodwill, not for the reason of this or that, but with a goodwill that is grounded in a relationship of mutual respect for rights with the trustee. This relationship may be new or long-term, private or business, but both parties seem to believe that this kind of relationship is theirs. The rights may be specific to each party, e.g., one party receives a certain kind of favor, and the other party some other favor. The knowledge of what this “charter of rights” contains is based on mutual experience, through time, of what I grant you and you grant me. It may be that we never even discussed the matter, we just know. Maybe this could be something like “a plural pre-­ reflective awareness” (see Schmid 2017). The fair distribution of the granted favors would be interesting to study. Another interesting question is how this kind of scenario works in situations where the power relations are unequal. The “charter of rights” may be in balance,  J. Baker (1987, p. 8) thinks that a minimum of trust in others, in their good will or in their desire for recognition or respect is required for there to be respect. In my framework this could be viewed as our need for at least predictive “trust” in others for us to be able to enter a relationship with them where mutual respect can flourish, and genuine trust can develop. (Predictive trust involves the trustor’s belief that the trustee will act with goodwill towards the trustor, e.g., out of his good will or his desire for respect). 10  “Social norms” refer to “proper social norms” that are (unarticulated) society-wide conventions or group-specific norms; Cf. (R. Tuomela and M. Bonnevier-Tuomela 1992 and 1998). 9

12  Cooperation Rests on Trust—But What Is Trust and How… 

251

imbalanced, openly negotiated, or causing problems that are never openly discussed. In the present context, we just conclude that the “charter of rights” exists as a basis for Rational genuine trust. The parties expect to be met with goodwill that is grounded in mutual respect for specific rights. If something I want is not my right, I cannot expect you to act with that kind of goodwill towards me, but maybe just hope for it, and try to predict my chances. The rights in this context may be just quasi-moral or social rights. They may not follow any conventional views on morality, although they are fully accepted in this dyad.

2.4 Terminology Used for Trust—Examples and Discussion  o Trust, to Predict, to Depend/Rely on Someone, T to Entrust Someone Social trust is trust between people or groups that can function as agents.11 Trust is an attitude (and/or feeling) that people have when they are at peace and have an accepting attitude about being dependent on another person for an action. The trustee should be aware of the dependence and have a choice to either accommodate the trustor or let him down. The trustor may have no other option than to trust, but he may still have Genuine trust in the other person. In everyday language to “trust someone for buying the tickets” may mean to have Genuine trust or Predictive trust in him, or it may mean to (decide to) depend/rely on him, based on trust or prediction. In everyday speech the distinction between the meaning of words often become blurred. To depend/rely on someone for something or to decide to do so, based on a prediction of his future favorable behavior  A collective’s rational genuine trust in another collective could be analyzed according to the same theory as trust between individuals. However, although a certain kind of collective can act as a group and have attitudes as a group, the positive feeling in the trust account need to be thought of as a positive attitude. Two groups could have a history of various kinds of interaction and deals that pave the road to a “charter of mutual rights” based on experiences. An extra feature of group agents is the internal cohesion around its ethos. For collective trust the group needs to be in control of possible dissidents’ actions. Groups that “keep together” well, could be trusted to do what they should do, and function as trustors, as well. (See M. Tuomela 2003 and M. and R. Tuomela 2005). 11

252 

M. Tuomela

deserve to be kept apart from trusting (having an attitude/feeling of trust). “Entrusting someone with buying the tickets and after that depending/ relying on him” does not need to but may entail trust. The person is usually made aware of what is expected of him, and that people depend on him. Entrusting usually involves decision just like depending on and relying on may do.

An Example of Genuine Trust When John finds himself relying on Mary or decides to rely on her for acting on a matter of his concern, without fear of being disappointed, he may be said to genuinely trust Mary and rely on her vis-à-vis the action. He believes that Mary is aware of his dependence and that she has a choice to let him down. John may have no options, but still trusts her. Alternatively, he may just trust her for potentially acting on the matter and still decide not to rely on her but choose to act himself. The trust he has for Mary is based on his belief that he has a right to be accommodated by her in this matter, based on their relationship of mutual respect for certain rights, this being such a right. If Mary will let him down, John will feel rightfully betrayed in his trust, if he indeed was right about their “charter of rights.”

 e Relationship of Mutual Respect for Rights in Rational Th genuine Trust Genuine trust does not only involve that the trustor is at peace with being dependent on the trustee and all the beliefs of predictive trust, but it also involves the trustor’s belief of a mutually acknowledged right to be accommodated in a specific case. In the example, John’s belief of such an acknowledged right is not the reason for his expectation to be accommodated by Mary, but this is why he expects this of Mary and this is why he feels betrayed in his trust if Mary does not perform the action. More specifically the belief of the right goes as follows. John believes that Mary and he have a relationship of mutual respect for certain rights, a belief that

12  Cooperation Rests on Trust—But What Is Trust and How… 

253

is based on his experience of having these rights and her having certain (other) rights that are respected. If the present expected action is not among the actions that he has a right to expect, he is not justified in having Genuine trust in Mary. When John’s Genuine trust is justified, and he is not mistaken about what his respected rights are, Mary betrays his trust, if she does not accommodate him. The trustor’s belief of the relationship of mutual respect for rights is not just another reason for predictive trust. It is why he trusts (a causal reason)—the backbone of Genuine trust.

Predictive Trust Predictive trust is an attitude that John arrives at after believing that Mary can and will perform the action in a way that furthers John’s welfare. He thus has an accepting attitude of being dependent on Mary and has a positive feeling about it. This takes place in a context where John wants Mary to perform the action, he believes she knows about this, and that his want is relevant to her—she is free to and will take it into account in her acting. John’s Predictive Trust is based on his reason-based beliefs and if Mary does not deliver, John blames himself for his mistaken beliefs (if Mary did not mislead him). When people have Predictive trust in each other, they just count on each other to behave in the desired way, having good reasons to expect this. E.g., Mary trusts that John will buy her a necklace, because she believes that he is in love with her. She trusts the priest to give her wallet back to her because she believes he is honest. Furthermore, when people have good reasons to believe that they have rights, but they lack the belief that these are acknowledged in a relationship of mutual respect for certain rights, they may do no more than count on the trustee to act in accordance with the rights, e.g., based on his moral character or some expected sanctions. Mary has Predictive trust in John concerning his non-violent behavior, believing that he follows the law. Later she has Genuine trust in him vis-à-vis this matter, when she believes that they have a relationship of mutual respect for certain rights, including the right to be treated without violence.

254 

M. Tuomela

 rust as a Gift that Needs to Be Accepted; and Thick and Thin T Genuine Trust Genuine trust is like a gift that needs to be accepted by the trustee. It is an honor to be offered someone’s trust, but as the gift of trust involves a demand on the trustee, his acceptance is required. If Mary trusts John to come home early, she may have Predictive trust in him, based on good reasons. For her to have Genuine trust in John she must believe that they have a relationship of mutual respect for rights.12 Only if it is her right (she rightfully believes that he has accepted this), can she blame him for betrayal if he comes home late. If trust does not involve the option of betrayal, it is not Genuine trust. Of course, Mary may be wrong in her beliefs (based on her subjective experiences) about their mutual understanding of their “charter of rights.” In that case she has misplaced Genuine trust in the sense of unjustified Genuine Trust. Only justified Genuine trust may be betrayed. Genuine trust is not restricted to personal relationships. Two businessmen may, at first, have Predictive trust in each other that will develop into Genuine trust through time. When people respect each other’s rights for instrumental reasons one could call it thin Genuine trust. In thick Genuine trust a person’s rights are respected for intrinsic reasons— he ought to have what is his right to have.

General Trust, Mutuality, and Power General trust in a person may be either Predictive trust or Genuine trust. It is Predictive trust concerning a wide range of matters, when rightful demands are expected to be accommodated, e.g., due to the trustee’s fairness. People may have general Predictive trust in a priest, based on his position and assumed moral standards. It may be Genuine trust when it is based on a relationship of mutual respect for a wide range of rights. The relationship of mutual respect for rights in Genuine trust could in extreme  See (O. Lagerspetz 1998, pp. 49, 64, 103)

12

12  Cooperation Rests on Trust—But What Is Trust and How… 

255

cases be a relationship of one-sided respect for rights, as in a situation where one party is in no position to accommodate the other party. The effect of unequal power, e.g., in the context of cultural habits of inequality between the sexes, would be an exciting area for study.

 rediction, Predictive Reliance, Trustworthiness, Risk, Degree, P and Decision Trust has been studied as a commitment, a relationship,13 an attitude, an emotion, or a mixed case, e.g., a belief with an affective component. Two notions that are weaker than Predictive trust, “Predictive reliance” on, and a plain “Prediction” of a person’s behavior, allow for depending on side-­ effects of his action. Trusting requires a context where the trustee is believed to be aware of the trustor’s wish to be (potentially) dependent on the trustee’s action and to be free to choose to take the trustor’s wish into account or let him down. The belief of a person’s trustworthiness is a result of Predictive trust or Genuine trust or a judgment preceding trust. In the context of prediction, it is the result of an evaluation before decision to depend on the person. Trusting involves risks, but only seen from a third-­ person point of view. When a person trusts, he may be aware of the third person point of view, but he does not deem it risky, or else he would not trust. This holds both for Predictive trust and Social normative (genuine) trust—two main kinds of trust that one does not decide to have, but one just “falls into,” like falling in love. Risk calculations belong to predictions and so do degrees of “trust.” People may decide to depend/rely on others based on their predictions.

 O.Lagerspetz (1998, Chapter 2) on a view that trust is not primarily a psychological state, a particular disposition, but something that is applicable to human relations, and (Chapter 3) on trust as something we typically think of only when we have reason not to trust. 13

256 

M. Tuomela

2.5 On Subjective and Objective Rationality in the Context of Trust14 In “(subjectively) rational Predictive (intuitive) trust” the trustor feels secure, but he knows that, from “a third person point of view,” his trust may be based on his own optimistic evaluation of the situation, and thus he should be aware of the risks, (seen from an outsider’s point of view). If his reasons for having the goodwill belief of the trust account above, are mainly due to his own basic sociality, he is not objectively rational, and the ensuing cooperation is not objectively rational. However, in a broader sense of objective rationality he is rational, when he thinks that people are rational enough to see what is good for them, and that it is good for people to strive for happiness through virtue, as man is basically social— thus, if possible, refraining from letting others down. Objective rationality  J. Baker’s (1987) view on the rationality of the trustor’s commitment to certain beliefs involved in trusting also stresses the dynamics of the relationship between trustor and trustee. She thinks that the rationality of the commitment can be judged in the same way as that of decisions, although this commitment is only indirectly the result of the trustor’s own intentional behavior—the control he exerts in the process in which the relationship develops. She considers a decision rational if it serves the goals of the deciding person in an optimal way. In my view this idea of rationality can be seen as two-fold in this context. The trustor’s commitment to certain beliefs about the trustee is rational if it results in good guidance (e.g., it pays off in warranted trust) or if the commitment itself produces that which was believed to exist (viz. trust as a good). The subjective rationality of my trustor becomes the more “objectively” (viz. inter-subjectively) rational, the more the parties share views about how to interpret their shared experiences, what their norms are, and when these apply. When people need to get along, the wisdom of “how” must be jointly produced. Expectations are rational if they are the product of fair “negotiation.” When “we” have found out how we function well together we can trust each other in an objectively (inter-subjectively) rational way. B. Lahno’s (2002, pp. 426-429) suggestion for an explication of rational trust as an emotional attitude is based on a somewhat similar view that R. de Sousa (1980, 1987) purports concerning the rationality of emotions. According to Lahno an emotion could be seen as rational if it guides our theoretical reflection towards true knowledge and  our practical reasoning towards right decisions. Lahno thinks that trust can be considered rational in a broad sense when it gives a rational understanding of the various circumstances in such a situation and when it motivates to trusting decisions that are reasonable in the light of short- and long-term goals. However, he thinks that trust is not in principle rational in any sense. An account of rational trust is just an explication of what we mean by appropriate trust. My notion of subjectively rational trust expresses the idea that trust is a person’s attitude and as such subjective. If a trustor could have objectively justified evidence for his beliefs there would be no place for trust. What we have is a person’s subjective view of some evidence, and he is the expert in his own trust. The best he can do is to be sensitive to his partner’s desires, views, and norms in the process of building their relationship. A shared understanding and shared norms are the  best building-blocks for  inter-subjectively rational trust, the  closest one can get to “objectively” rational trust. 14

12  Cooperation Rests on Trust—But What Is Trust and How… 

257

in a narrow sense, counts on people being social mainly for instrumental reasons, or in relation to in-group persons. This would grant contacts with, e.g., business partners, friends, and family. Research gives some support for the idea that a cooperative predisposition could be in the human genes. Human beings are social, in the sense that they live in groups as their pre-human ancestors. A gene-based basic sociality could provide the basis, not only for the predisposition to cooperate and to trust, but also for a predisposition to feel sympathy for others. The emerging field of cultural coevolution may have something to offer in these matters. Virtue is life-preserving, i.e., has a positive impact on individuals and groups. In a specific case of Genuine trust, the trustor may, of course, be mistaken about having a relationship of mutual respect for the other’s rights, with the trustee. E.g., the trustee has given false signals, he may be unsocial, and thus insensitive to a bad reputation, or then he is non-rational. These kinds of situations do not prove that the trustor’s subjectively rational trust was objectively non-rational, but they show that we sometimes misjudge our relationships, or encounter people who are non-rational (viz. they want to harm themselves, or they are shortsighted), or just plain non-social. To sum up: Naturally dependent persons cannot fare well alone, and long-term relationships require virtuous behavior. Thus, it is objectively rational (in a broad sense) to be fair. For naturally dependent persons cooperation is necessary and requires mutual trust. Trust does not arise where virtuous behavior is lacking. Genuine trust is based on a relationship of mutual respect for the other’s right, and subjectively rational trust is thus objectively rational in a broad sense—in the most basic sense of human rationality. This kind of rationality rests on criteria for what makes for a good life for people living in interdependence with each other. In simplified terms, one could say that it is rational to believe that people are rational, viz. that they do what is good for themselves. However, people are also naturally egoistic, aggressive, and competing for resources inside their groups. Thus, we could claim that it is equally objectively rational to live by the premise that people want to cheat to avoid costs. However, the cooperative predisposition is suggested to be predominant in a group of the size of a village. In the outer layers of the network there is room for competition, as long as relevant rights of the other are respected. If we step on a trustor’s rights in the fringes of the network, our smeared

258 

M. Tuomela

reputation may catch up with us in the next turn, but even worse—we smear our character. A truly good life is not only to thrive in the fellowship of others, but to thrive in the company of oneself, as well. Sociality (viz. to seek cooperation and fellowship) is part of man’s biological make-up and above it has been argued to be a basis for a predisposition to respect others’ rights for instrumental reasons and thus for cooperation based on thin Genuine trust. Sociality has also been argued to be the basis for sympathy, the development of morality, and thus for genuine respect and cooperation based on thick Genuine trust.15

2.6 A Basic Moral We-perspective The relationship of mutual respect for rights in Genuine trust can be considered a relationship of a basic moral “we-perspective.” The parties believe that they have such a relationship, that when one of them performs a relevant action (viz. an action that belongs to the realm of mutual rights), he ought to and will consider the point of view of both parties. The most basic level of we-thinking is to grant others what you think you may rightfully expect to have from them, on the grounds of seeing all men as equal members of mankind.16 We-thinking is learned from early childhood on in our dealings with in-group persons, where our egocentric wants are confronted with theirs. The sympathy and ­interdependence between close friends and kin compel us to see the equal rights of our fellows.17 A we-perspective of this basic kind is deeply social or moral.18  Cf. (L. Hertzberg 1988, p. 317) on trust as a default position. The article is part of the author’s later work (Acta Philosophica Fennica The Limits of Experience, 1994). 16  Cf. (O.  O’Neill 2002, 33) on Kantian fundamental duties and rights; principles of justice as principles fit for all. 17  See the survey by van Lange et. al (W. Liebrand et al. 1992, Chapter 1.) The experiments of P. Kollock (1998, pp. 202-204) reveal some reasons for preferring cooperation to exploitation of one’s partner in an imagined Prisoner’s Dilemma: e.g., a good reputation in one’s group, adopting a group-perspective, moral codes of fairness and equality with respect to group-members as well as strangers, and a weak interest in maximizing one’s gains. 18  According to J.  H. Turner (2002, p.  123) some scholars (Fiske 1991; Cosmides 1989) have argued that there is a hard-wired propensity for humans to assess reciprocity and justice. He thinks that this is credible as our ancestors were under selection pressures to develop mechanisms for bonding. A sense of reciprocity, fairness, and justice would be an effective way to bind people together in cooperative relations. 15

12  Cooperation Rests on Trust—But What Is Trust and How… 

259

In cases of thin mutual respect and thin Genuine trust the motive to see to the other’s right was said to be instrumental, but his equal right is intentionally respected all the same. Here we could speak of a basic quasi-­ moral we-perspective, as the moral motive is lacking, although the moral principle of fairness is practiced. The same holds for cases of team action, lacking agreements, where every member’s contribution is needed for the achievement of a joint goal and the members act for merely instrumental reasons. Still, there is a basic quasi-moral we-perspective in the group, in the sense that all the members recognize, that what they expect from the others, the others have a right to expect from them, for achieving the joint goal. The instrumental motives are here conceptually inherent in the joint action structure in a technical sense, as one cannot intend to act jointly, without intending to perform one’s part, expecting the corresponding action from the others, and granting the others a technical right to expect one’s part-performance. (See R.  Tuomela, and M.  Tuomela 2003 and 2005). The basic social or (quasi-)moral we-perspective is not the same as the “group-perspective” or the we-perspective of a group that involves viewing matters from the point of view of the group and furthering the group’s interest, sometimes even when it does not coincide with the private interests of the members. Still, the “basic (quasi-)moral we-perspective” is a typical feature of any kind of flourishing group life and an inbuilt feature of groups that act as teams, where people stand and fall together achieving the group’s goal. Here, in the context of trust based on a relationship of mutual respect for the other’s right, the focus is on the right of each member, instead of on the group’s interest. The basic moral we-­perspective may well be the view that paves the road to a we-perspective of a group (viz. a group-perspective). It may be, that we first must learn to see the importance of the equal rights of our partners before we are ready to care for the interest of a group and be fit to act in a fair way in a group context, furthering the group’s interest. In a relationship of mutual respect for rights, or in one in which the parties adhere to a basic moral we-­perspective genuine trust can arise and flourish. Moral values like the basic moral we-perspective (entailed in mutual respect for the other’s right) should belong to the mental toolbox of children, to be cultivated by their caretakers. The general recipe for enhancing

260 

M. Tuomela

a basic moral we-perspective and thick genuine trust seems to amount to making groups out of people by encouraging contacts, dependencies, and concrete cooperation. Institutional law enforcement in modern societies has not replaced the need for trust, although it does solve many problems. We still want and need to genuinely trust each other—to find ways to enhance rational Genuine trust and direct cooperation of earlier times.

3 Part II: Cooperation Rests on Trust In a multi-cultural urban milieu, people must be brought together to be able to experience that outsiders may become insiders, when they are given a chance. Rules help people to dare to enter cooperative relationships based on Predictive trust. These relationships may eventually give rise to rational Genuine trust. Rules teach a lesson about equality and fairness and may become internalized with a deep understanding of the values involved. On the other hand, control may lead to cheating, when there is a chance to escape.19 Theses: It is argued that there is a two-way connection between trust and cooperation (R. Tuomela and M. Tuomela 2005). The first thesis is that rational intentional cooperation between “conditional cooperators” requires, at least, Predictive trust. The second thesis claims that successful rational intentional cooperation enhances rational Genuine trust (viz. Social normative trust). The latter thesis can be strengthened in the case of “we-mode” cooperation. Thus, the third thesis is that, given some qualifications, when some persons rationally cooperate in the we-mode, where the members hence are sincerely collectively committed (involving, at least, an instrumental and social obligation towards the others to perform their parts, in a case lacking an agreement), and where they mutually believe that this is the case, their relationship may be characterized as one involving mutual respect for the others’ rights and rational Genuine trust (viz. (subjectively) Rational social normative trust in the sense of (RSNTR)), vis-à-vis their part-performances. Defense:  Cf. (O. O’Neill 2002, pp. 18-19; and H. Nissenbaum 2001).

19

12  Cooperation Rests on Trust—But What Is Trust and How… 

261

Here is a summary of our main points for the defense of the theses. First thesis: Each rational party of cooperation must believe that, in order to initiate the cooperative action and to go on cooperating, the other will intentionally do something that furthers the interest of his partner (in general terms: to act with goodwill towards him). However, we have claimed that the mere Goodwill belief of rational Predictive trust is not sufficient, but that at least Predictive trust is needed for rational cooperation (R. Tuomela and M. Tuomela 2005, Section IV.1). The analysis of Predictive trust involves some additional conditions, i.e., the conditions of Genuine trust (RSNTR) (except for that part of the goodwill clause (3) which concerns respect of rights), and the satisfaction of those additional conditions is also needed in the case of rational agents. Also, the conditions of Context (Y) and the first two belief clauses of the trust account must be satisfied for reasons of rationality. The fifth condition must be satisfied for making cooperation possible. If a rational person believes that his potential partner will perform the desired action, but that he lacks an accepting attitude vis-a-vis becoming dependent on his partner (clause (5)), he is not only inconsistent (nonrational), but he is clearly not ready for cooperation. Note that trust does not require action but provides readiness for action. Thus clause (5) is needed for initiating and maintaining cooperation not only on rational but also on conceptual grounds and presumably on psychological grounds. The satisfaction of the fourth condition is not needed for cooperation but lacking a “positive feeling” about the future dependence on another person may well turn potential cooperators to seek other partners. Anyway, it is “rational” to react with positive feelings to becoming dependent when gratification is expected. Second thesis: According to the above result, rational intentional cooperation requires (at least) Predictive trust. Thus, to show that successful intentional cooperation enhances rational Genuine trust (i.e., Social normative trust), we first note that Genuine trust requires a history of “good experiences”—c.f., the parties’ mutual respect of each other’s rights. Then, we may claim that exactly such good experiences are acquired by that successful cooperation based on Predictive trust. The thesis further claims: The latter thesis can be strengthened in the case of “we-mode” cooperation. This we will accept without further ado.

262 

M. Tuomela

Third thesis: Here we-mode cooperation is claimed to be a situation where Genuine trust (involving mutual respect for the other members’ rights) will thrive. The proof goes along the following lines (R. Tuomela and M. Tuomela 2005, Section IV.3). The social commitment involved in collective commitment in we-mode cooperation20 involves that the participants not only expect or are disposed to expect that the others perform their parts, but that they have the right to such expectations and mutually know this (cf. also R.  Tuomela and M.  Tuomela 2003). The social right here can be called “quasi-moral” as it concerns refraining from letting the others down. This phase of the argument thus has established part of the mutual respect assumption needed by Genuine trust (RSNTR). What is yet needed is to show that there is a non-fleeting social relationship of mutual respect. This is normally the case on contingent grounds. According to the account of Genuine trust, the central element is a trustor’s belief of the following kind: “Due to a relationship of mutual respect for the other’s rights, he normatively expects of the others, on social or (quasi-) moral grounds, that they will act with goodwill towards him in those matters where he has a right to expect such action, and he also expects that they will so act, in part due to his right” (cf. clause 3 of Genuine Trust). When the motives become genuinely other-regarding this social normative trust turns into thick Genuine trust. Be it groups of the we-mode or the I-mode kind, a certain amount and kind of experiences of mutual respect for the others’ rights will be subjectively regarded by the participants to be sufficient for them to consider themselves to have a relationship of mutual respect for rights. Given this empirical result, right-involving rational successful I-mode or we-mode cooperation will generate rational Genuine trust among normally rational agents vis-à-vis their justifiably expected contributions. It seems that our joint effort to fit my account of trust into Raimo’s account of cooperation was successful, or so we deemed. There is no foolproof recipe for finding trustworthy participants for cooperative ventures,  In cases of I-mode groups (pro-group I-mode), a promise will do the work of the structural feature of collective commitment in the we-mode group—allowing for expectations of respect of rights. 20

12  Cooperation Rests on Trust—But What Is Trust and How… 

263

nor for creating relationships of mutual respect for rights where Genuine trust can grow and thrive. A good road to Genuine trust is to start by mutual evaluation of features of the person and the situation, and assurance of mutual respect. If this leads to Predictive trust, the parties feel secure to initiate a cooperative action with each other, thus depending on the features of the other and the situation. If all goes well and each party has an experience that he interprets as one in which his rights have been respected and he has respected the other’s rights, he has reason to form the belief that a relationship of mutual respect is being/has been established. That relationship is then the growing ground for Genuine trust. Sociality (viz. to seek cooperation and fellowship) is part of man’s biological make-up and above it has been argued to be the basis for a predisposition to respect others’ rights for instrumental reasons and thus for cooperation based on thin Genuine trust. Sociality may also be argued to be the basis for sympathy, the development of morality, and thus for genuine respect and cooperation based on thick Genuine trust. Cultural co-evolution as an explanatory model for the development of our species may be a relevant frame of reference for the ideas above. The general recipe for enhancing a “basic moral we-perspective” and thick Genuine trust seems to amount to making groups out of people by encouraging contacts, dependencies, and concrete cooperation. Acknowledgements  I thank my late husband, Prof. Raimo Tuomela, for cooperation on our article “Cooperation and Trust.” I also wish to thank Prof. Robert Audi and Prof. Cristiano Castelfranchi for comments on my thesis about trust, as well as Dr. Soc. Sci Pekka Mäkelä, Docent, Dr. Soc. Sci Raul Hakli, and M.A. Kaarlo Miller for useful comments on my work throughout the years. Some of this text is reused with the permission of Sage Publications, Thousand Oaks, Ca, from “Trust (Social)” by Maj Tuomela, in Encyclopedia of Philosophy and the Social Sciences (Ed.) Byron Kaldis, 2013, pp. 1018-1020. A fuller treatment of trust is published in “Rational Social Normative Trust as Rational Genuine Trust”. In L.  Siegal (Ed.), Philosophy and Ethics: New Research. Hauppauge, NY: Nova Science Publishers, Inc., 2006, pp. 1-56, as well as in Trust and Collectives. Research Reports 1/2006, Helsinki: Department of Social Psychology, University of Helsinki.

264 

M. Tuomela

References Audi, R. 1993. Dispositional Beliefs and Dispositions to Believe. Noûs 28 (4): 419–434. Axelrod, R. 1984. The Evolution of Cooperation. New York: Basic Books. Aykens, P. 2002. Where’s the Trust? States, Markets and a Social Theory of Currency Crises. paper for panel SA06—Challenging IPE Perceptions of Global Finance and Credit: Disciplinary Narratives—at the 2002 International Studies Association convention in New Orleans, Louisiana, March 24-27, 2002. Baier, A.C. 1995. Moral Predjudices: Essays on Ethics. Cambridge, MA: Harvard University Press. Baker, J. 1987. Trust and Rationality. Pacific Philosophical Quarterly 68: 1–13. Castelfranchi, C., and Falcone, R. 2010. Trust Theory A Socio-Cognitive and Computational Model. Wiley Series in Agent Technology. Chichester, U.K.: John Wiley & Sons Ltd. Dunn, J. 1984. The Concept of ‘Trust’ in the Politics of John Locke. In Philosophy in History, ed. R.  Rorty, J.B.  Schneewind, and Q.  Skinner, 279–301. Cambridge: Cambridge University Press. Fukuyama, F. 1995. Trust—The Social Virtues and the Creation of Prosperity. New York, NY: Free Press. Gambetta, D., ed. 1988. Trust—Making and Breaking Cooperative Relations. New York, NY: Basil Blackwell, Inc. Gintis, H. 2003. Solving the Puzzle of Prosociality. Rationality and Society 15: 155–187. Hardin, R. 2002. Trust and Trustworthiness. New  York, NY: Russell Sage Foundation. Hertzberg, L. (1988). On the Attitude of Trust. Inquiry: An Interdisciplinary Journal of Philosophy 31: 307–322. ———. 1994. The Limits of Experience. Acta Philosophica Fennica (The Philosophical Society of Finland). Jones, K. 1996. Trust as an Affective Attitude. Ethics 107: 4–25. Kollock, P. 1998. Transforming Social Dilemmas. In Modeling Rationality, Morality, and Evolution, ed. P. Danielson, 185–209. New York and Oxford: Oxford University Press. Lagerspetz, O. 1998. Trust: The Tacit Demand. Dordrecht. Netherlands: Kluwer Academic Publishers. Lahno, B. 2002. Der Begriff des Vertrauens. Paderborn: mentis Verlag.

12  Cooperation Rests on Trust—But What Is Trust and How… 

265

Liebrand, W., D.  Messick, and H.  Wilke. 1992. Social Dilemmas. Oxford: Pergamon Press. McGeer, V. 2002. Developing Trust. Philosophical Explorations V: 21–39. Misztal, B. 1996. Trust in Modern Societies. Oxford, U.K. and Cambridge, MA: Blackwell Publishers. Mäkelä, P., and C. Townley, eds. 2013. Trust, Analytic and Applied Perspectives. Amsterdam—New York, NY: VIBS, Rodopi B.V. Nissenbaum, H. 2001. Securing Trust Online: Wisdom or Oxymoron? Boston University Law Review 81: 101–130. O’Neill, O. 2002. A Question of Trust, The BBC Reith Lectures. Cambridge, U.K.: Cambridge University Press. Pettit, P. 1995. The Cunning of Trust. Philosophy and Public Affairs 24: 202–225. Richerson, P., R. Boyd, and J. Henrich. 2003. Cultural Evolution of Human Cooperation. In Genetic and Cultural Evolution of Cooperation, ed. P. Hammerstein, 357–388. Cambridge, MA: The MIT Press (in cooperation with Dahlem University Press). Schmid, H.B. 2017. What Kind of Mode is the We-mode? In Social Ontology and Collective Intentionality: Critical Essays on the Philosophy of Raimo Tuomela with His Responses, ed. G. Preyer and G. Peters, 79–93. Cham, Switzerland: Springer, Studies in the Philosophy of Sociality. Solomon, R.C. 2000. Trusting. In Heidegger, Coping, and Cognitive Science. Essays in honor of Hubert L. Dreyfus, ed. M. Wrathall and J. Malpas, vol. 2, 229–243. Cambridge, MA: MIT Press. de Sousa, R. (1980). Arguments from Nature. Zygon 15: 169–191. ———. (1987). The Rationality of Emotion. Cambridge, Mass: MIT Press. Tuomela, M. 2003. A Collective’s Trust in a Collective’s Action. Protosociolgy. Understanding the Social II: Philosophy of Sociality 18 (19): 87–126. ———. 2006a. Trust and Collectives. Research Reports 1/2006. Helsinki: Department of Social Psychology, University of Helsinki. ———. 2006b. Rational Social Normative Trust as Rational Genuine Trust. In Philosophy and Ethics: New Research, ed. L.  Siegal, 1–56. Hauppauge, NY: Nova Science Publishers, Inc. ———. 2013. Trust (Social). In Encyclopedia of Philosophy and the Social Sciences, ed. B. Kaldis, 1018–1020. Thousand Oaks, CA: Sage Publications. Tuomela, R., and Bonnevier-Tuomela, M. 1992. Social Norms, Tasks, and Roles. Reports from the Department of Philosophy 1, Helsinki: University of Helsinki.

266 

M. Tuomela

Tuomela, R., and M.  Bonnevier- Tuomela. 1998. Norms and Agreement. In Sources of Law and Legislation, ed. E.  Attwool and P.  Comanducci, vol. 3, 87–93. Stuttgart: Franz Steiner Verlag. Tuomela, R., and M. Tuomela. 2003. Acting as a Group Member and Collective Commitment. Protosociology 18: 7–65. ———. 2005. Cooperation and Trust in Group Contexts. Mind and Society 1 (4): 49–84. Wolfe, A. 1989. Whose Keepers? Social Science and Moral Obligation. Berkeley: University of California Press.

Index1

A

Acting together, 64, 96, 221, 222, 232, 234n8, 238 Action theory, 17 Agency transformation(s), 23, 25, 212 Artificial Intelligence (AI), 32 B

Bacharach, Michael, 22–25, 212 Bratman, Michael, 42, 43, 46, 48, 97, 164n9, 180, 222, 229, 230, 232 C

Collective acceptance, 63, 63n5, 65, 66, 80, 83, 105–115, 115n3,

117–126, 164n11, 169, 202, 203, 214, 233, 234n8, 235–237 Collective action(s), 42n3, 44n5, 50n7, 56–58, 73, 76, 77, 80, 88, 95, 117, 118, 152–163, 160n7, 161n8, 164n11, 166, 168, 209, 213 Collective activity(ies), 56, 57, 87 Collective agency, 86, 153n2, 160n7 Collective agent(s), 63, 66, 67, 74, 152 Collective attitudes, 61, 67, 72, 107, 126 Collective belief, 53–55 Collective commitment, 63–65, 107, 111, 119, 122, 153, 166, 169–171, 170n14, 173, 174, 232, 234, 236, 262, 262n20

 Note: Page numbers followed by ‘n’ refer to notes.

1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Garcia-Godinez, R. Mellin (eds.), Tuomela on Sociality, Philosophers in Depth, https://doi.org/10.1007/978-3-031-22626-7

267

268 Index

Collective intentional action(s), 160, 168 Collective intentionality, 63, 81–83, 91–94, 109, 110, 153, 156, 160, 160n7, 162, 163, 166–174, 177, 178, 180–185, 187–189, 198, 208, 210 Collective intentional state of acceptance, 122 Collective intention(s), 59–66, 69, 95, 123, 125, 171 Collectively accept(ed), 83, 116, 233, 234 Collectively committed, 112, 132n2, 145n20, 260 Collective reasoning, 60–67, 73–77 Collective responsibility, 26–31 Collective subject(s), 183–186, 191 Collectivism, 14, 60 Common knowledge, 42, 43, 45, 48, 55, 62, 75, 230 Conditional intention, 164n10, 171 Conditional we-intention(s), 18, 160–167, 165n12 Constitutive rule(s), 129–137, 139–148, 153, 156, 158, 161, 162, 164–165 Cooperation, 55, 77, 169, 173, 174, 222, 256–258, 258n17, 260, 261, 263 Customs, 80, 83, 87 D

Decision-making, 12, 22, 23, 26, 30, 95, 97, 100, 172, 247

E

Emergent coordination, 94–99 F

Fiction, 181 Fictionalist, 182, 198 Fictitious, 91, 92, 122n9, 142n16, 179, 181 Formal roles, 58 G

Game theory, 22, 24, 43, 44, 44n5 Generalised conditional we-intention(s), 164, 171 General trust, 254–255 Genuine trust, 244–253, 249n6, 250n9, 251n11, 255, 257–263 Gilbert, Margaret, 42, 43, 46, 48, 65, 183, 228n6, 233, 234n8, 236, 239 Group, 172n16 act, 80 action, 22, 27–29, 32, 42, 52, 53 activity, 62 agency, 28, 63, 91 agent(s), 25, 26, 32, 142n16, 153n3, 160n7, 161n8, 182n2, 190, 198, 234, 238, 251n11 attitudes, 25, 26, 30, 80, 122n9 belief, 39, 41–47 believe, 44 ethos, 80, 203, 205 intention(s), 25, 41 level responsibility, 30

 Index 

morality, 201, 206, 213 preference(s), 22–25, 31 reasoning, 172 reason(s), 63, 100, 171n15, 205, 232, 234–236 responsibility, 27–29, 31 Group-based commitment, 165n12 Group-based reason(s), 26, 237 Group-normativity, 203 Group-social normativity, 221, 222, 231, 233, 237, 238 H

Habits, 80–82, 87, 93, 94, 98–99, 244, 255

269

Intentional action(s), 30, 32, 84–86, 95, 98, 153n2, 168, 249n6 Intentional attitudes, 142n16 Interactive knowledge, 44–45, 44n5, 48–56, 50n6, 50n7, 80 J

Joint action(s), 19, 20, 39, 41, 59–62, 64, 65, 67, 68, 70, 71, 74–77, 96, 178, 209, 259 Joint intentional action, 161, 164–168, 221, 235 L

I

I-mode, 14, 21, 24, 32, 60, 61, 63, 64, 66–68, 87, 95, 110, 168, 174, 196, 197, 222–232, 232n7, 235, 238, 262 I-mode action, 14 I-mode we-intention, 168 Individualism, 31, 60, 63, 67–77, 94 Individualism—collectivism debate, 30 Institutional practice(s), 163, 166 Institutions, 37, 63n5, 84, 92, 97, 101, 105–107, 109–113, 111n1, 117, 118, 119n6, 120, 120n8, 124–126, 129–131, 132n2, 133, 135, 136, 138, 138n9, 139, 140n11, 141–145, 148, 163, 164n11, 165, 173, 177, 179, 198, 202, 213, 214

Level problem, 39, 48–53, 55 Lewis, David, 42, 43, 49 Ludwig, Kirk, 152–168, 152n1, 153n3, 155n4, 157n5, 160n7, 164n10, 164n11, 165n12, 167n13, 170–173, 174n18, 180, 208, 209n3 M

Meshed cognition, 98–99 Moral normativity, 65, 202 Moral norms, 201, 206–208, 211, 213, 214, 225, 226, 226n5 Moral responsibility, 28–31 Mutual belief(s), 19, 20, 40, 41, 43, 45, 46, 48–49, 55–57, 202, 209, 223–225, 229, 233, 235, 237

270 Index N

Nash equilibria, 24 Nonintentional action(s), 85, 86, 110 P

Participatory intention(s), 19, 20, 170, 171 Pattern-governed behavior(s), 83–87, 89–91, 93–95, 99, 100 Pettit, Philip, 29 Practical reasoning, 19, 24, 25, 170, 247, 256n14 Predictive trust, 244–247, 249n6, 250–255, 250n9, 260, 261, 263 Preference transformation(s), 18, 23–25 Prisoner’s dilemma, 18, 23, 213, 258n17 Pro-group I-mode, 22, 23, 61–64, 262n20 Proxy agency, 151–160, 155n4, 162, 163, 168, 172–174 Proxy assertion, 154 Pure I-mode, 22 R

Rational genuine trust, 250 Rational Predictive trust, 244 Rational social normative trust in the sense of (RSNTR), 260 Regulative rules, 130–134, 133n3, 136, 138–142, 142n16, 144–147, 144n19

Role-consciousness, 192, 193 Role-mode, 177–180, 183, 189–198 Role(s), 20, 46, 47, 54, 55, 72, 73, 111, 118, 124, 159, 164–167, 165n12, 170–174, 177–179, 181, 184, 189–198, 213 bearer(s), 190, 191, 193 consciousness, 192 holder(s), 153, 162, 165n12, 166–168, 171–173 intentionality, 178, 179, 190 S

Schmitt, Michael, 60 Scientific realism, 13 Searle, Earlier, 138n8 Searle, John R., 19, 20, 63n5, 69, 113, 130–139, 132n2, 133n3, 135n4, 139n10, 141n14, 142, 145–147, 158, 162, 180, 193, 209n3, 214 Sellars, Wilfrid, 13, 84, 86, 89, 121, 202, 207, 208, 208n2, 210–212, 214–216, 215n5 Shared agency, 101 Shared beliefs, 178, 183–185, 223 Shared cooperative activity, 42 Shared intention(s), 96, 97, 221, 222, 229, 230, 232 Shared reason(s), 80, 168 Shared responsibility, 173 Shared we-attitude(s), 14, 85, 87, 89, 93 Social action(s), 17, 29, 62, 85, 87, 88, 93, 112, 117

 Index 

Social activity, 79 Social group(s), 20, 29, 62, 66, 71, 74, 161, 163, 205, 206, 211, 213, 233 Social normative (genuine) trust, 255 Social normative trust, 245–251, 248n3, 260–262 Social normativity, 201, 202, 239 Social norm(s), 20, 71, 107, 112, 117, 130, 144, 145, 205, 206, 208, 216, 219–239, 221n2, 226n5, 228n6, 234n8, 244, 250, 250n10 Social order, 173, 220, 221, 228 Social practice(s), 79–81, 85–95, 99, 101, 107, 110–112, 111–112n1, 125, 126, 130, 131, 133, 137, 138, 138n9, 142–145, 214, 223, 228 Social reason(s), 85, 88, 93, 117 Social roles, 20 Social structures, 109 Social trust, 244, 251 Status role(s), 155–159, 157n5, 161, 162, 166, 173 Sugden, Robert, 22, 25 T

Team reasoning, 12, 22, 23, 25, 26, 31, 212 Trust, 114, 156, 244–247, 249, 249n6, 249n8, 250n9, 251, 251n11, 252, 254–257, 255n13, 256n14, 258n15, 259–262 Trust context, 247, 248n3

271

U

Unconditional intention(s), 164n10, 171 Unconditional we-intention, 18 W

We-attitudes, 63, 67, 72, 80, 85–94, 91n2, 96, 99–101, 112, 115n3 We-commitment, 152, 163–166 We-intentionality, 174, 180, 184, 211 We-intention(s), 17–21, 25, 61, 64, 89, 153, 158, 161, 163, 167, 168, 172, 202, 207–211, 213, 214, 216, 235, 238 We-mode, 14, 21–23, 25, 28–30, 39, 60, 62–67, 80, 87, 100, 107, 113, 115n3, 125, 126, 166, 168–172, 174, 174n18, 178, 180n1, 183, 189–198, 204, 208–212, 209n3, 214–216, 232–238, 232n7, 260, 262 action, 14, 29, 30 attitudes, 63, 107 cooperation, 21, 22, 262 groups, 21, 30, 31, 201–206, 215, 262n20 preferences, 22, 32 reasoning, 23, 95 reason(s), 65, 66 social group(s), 202 thinking, 25, 63, 80 we-attitudes, 173 We-mode(s)/I-mode, 196, 209n3 We-perspective, 14, 187, 209n3, 214, 216, 258–260 We-reasoning, 12, 21–26, 65, 73, 172n16