On Social Facts

By Margaret Gilbert

Are social groups real in any sense that is independent of the thoughts, actions, and beliefs of the individuals making up the group? Using methods of philosophy to examine such longstanding sociological questions, Margaret Gilbert gives a general characterization of the core phenomena at issue in the domain of human social life. After developing detailed analyses of a number of central everyday concepts of social phenomena--including shared action, a social convention, a group's belief, and a group itself--she proposes that the core social phenomena among human beings are "plural subject" phenomena. In her analyses Gilbert discusses the work of such thinkers as Emile Durkheim, Georg Simmel, Max Weber, and David Lewis. "Gilbert's book aims to ... exhibit some general and structural features of the conceptual scheme in terms of which we think about social groups, collective action, social convention, and shared belief.... [It] offers an important corrective to individualistic thinking in the social sciences...."--Michael Root, Philosophical Review "In this rich and rewarding work, Margaret Gilbert provides a novel and detailed account of our everyday concepts of social collectivity. In so doing she makes a seminal contribution to ... some vexed issues in the philosophy of social science.... [An] intellectually pioneering work."--John D. Greenwood, Social Epistemology

• Language: English
• Publisher: Princeton University Press
• Release date: Jun 30, 2020
• ISBN: 9780691214627

Book preview

On Social Facts

On Social Facts

Margaret Gilbert

PRINCETON UNIVERSITY PRESS

PRINCETON, NEW JERSEY

Published by Princeton University Press,

41 William Street,

Princeton, New Jersey 08540

In the United Kingdom,

Princeton University Press, Oxford

Copyright © 1989 by Margaret Gilbert

Reprinted by arrangement with Routledge,

publishers of the 1989 edition; first Princeton Paperback printing, 1992

All Rights Reserved

Library of Congress Cataloging-in-Publication Data

Gilbert, Margaret.

On social facts / Margaret Gilbert.

p. cm.

Originally published: London; New York, 1989.

ISBN 0-691-07401-1 (cloth)—ISBN 0-691-02080-9 (pbk.)

1. Sociology—Methodology. 2. Social groups. 3. Social action.

4. Collective behavior. I. Title.

[HM24.G479 1992]

301′.072—dc20 91-38727

eISBN: 978-0-691-21462-7

For Miriam Gilbert

CONTENTS

Preface and acknowledgements  ix

I Introduction: everyday concepts and social reality   1

1 Preamble   1

2 The focus and thesis of this book   1

3 Social science and everyday concepts   3

4 Weber on everyday collectivity concepts   6

5 The everyday concept of a collectivity   8

6 Methodology   10

7 The main themes   12

8 Overview of chapters   14

II ‘Social action’ and the subject matter of social science   22

1 Introduction   22

2 Weber’s account of ‘social action’   24

3 The question of collectivities   34

4 Further considerations on Weber’s concept   44

5 Conclusions   55

III Action, meaning, and the social   58

1 Introduction   58

2 The deep level of the discussion   64

3 Winch on rule-following   71

4 Kripke’s Wittgenstein   100

5 The intentionalist programme   128

6 Group languages   132

IV Social groups: a Simmelian view   146

1 Introduction: Simmel’s statement   146

2 On Sharing in an action   154

3 ‘We’   167

4 Social Groups   204

V After Durkheim: concerning collective belief   237

1 Introduction   237

2 Durkheim on social facts: some salient features of the Rules discussion   243

3 Assessing accounts of collective beliefs: some tests   254

4 The simple summative account   257

5 A second summative account: adding common knowledge   260

6 A third summative account: the group as cause   274

7 A nonsummative account of collective belief   288

VI Social convention   315

1 Introduction   315

2 David Lewis on convention   319

3 Critique of Lewis (1): A flawless mechanism?   329

4 Critique of Lewis (2): Lewis’s conditions on convention   339

5 Critique of Lewis (3): Lewis and the ‘ought’ of convention   349

6 Critique of Lewis (4): conventions and collectivities   355

7 Towards an account of social convention   367

8 Social convention   373

VII On social facts   408

1 The structure of everyday collectivity concepts: summary of results   408

2 ‘The actions of participating individual men’   417

3 Concerning ‘individualism’ versus ‘holism’   427

4 A sketch of some further applications   436

5 On social facts   441

6 Afterword   445

Notes   446

Bibliography   496

Index   505

PREFACE AND ACKNOWLEDGEMENTS

I have been thinking about the topic of this book for many years. During this time my views on the character of social phenomena among humans have changed rather radically, becoming more holistic. Around 1982 I came to sense that what I refer to here as the conceptual scheme of singular agency was incapable of properly characterizing the experience of social connection. This initially inarticulate perception has found its expression in the present work, which was completed in 1986.

I have tried to make the book accessible to those previously unfamiliar with the issues and texts I discuss, presenting fairly detailed expositions of the texts in question. Meanwhile, I make no attempt to mention every related work. I have of course attempted to make it clear if anyone has influenced the specifics of what I say here. But I have kept undiscussed references to a minimum.

David Lewis introduced me (and the learned world) to the important notion of what is now generally referred to as ‘common knowledge’. He also provided an elegant account of social convention that demanded attention, and set a new standard of depth, clarity, and precision for such accounts. At one point I favoured an account of social phenomena which took common knowledge phenomena to be paradigmatically social. Later I began to feel that this was not so. Thus I came to appreciate the scepticism about common knowledge that Charles Taylor had been expressing. Though I continue to hold that common knowledge is an intriguing phenomenon of major importance in the context of human interactions, Taylor’s expressions of doubt helped to encourage my sense of the need for more holistic conceptions in the characterization of social phenomena.

My work on the topic of this book has been facilitated over the years by stipendiary research fellowships at St Anne’s College, Oxford University, St Hilda’s College, Oxford University, and the Institute for Advanced Study, Princeton. I have also enjoyed the privileges of a Visiting Fellow in the philosophy department at Princeton University on a number of occasions. I thank these institutions and the faculty members involved for this invaluable support. I should mention in particular Gwynneth Matthews and Gabriele Taylor, Jean Austin, and Morton White. I am also grateful to Albert Hirschmann and to Pat Sher.

I thank the following friends and colleagues for reading and commenting on part of the book manuscript or on related material: Polly Allen, Martha Bolton, Myles Burnyeat, Richard Collins, Raymond Geuss, Martin Gilbert, Tim Elder, Gilbert Harman, Rom Harré, Anne Hiskes, Ted Honderich, Anne Jacobson, Leonard Krimerman, Charlotte Katzoff, Saul Kripke, Joel Kupperman, David Lewis, Steven Lukes, the late John Mackie, A.S. McGrade, Judith May, Ruth Millikan, Mark Sainsbury, Thomas Scanlon, Amartya Sen, Jerome Shaffer, Frank Stewart, John Tienson, John Troyer, David Vance, Steven Wagner, Samuel Wheeler. I hope that I will be forgiven for any inadvertent omission. Being read in the course of writing has meant a very great deal to me. In addition, I thank John Albanese, Marie Becker, Richard Collins, Zlatan Damnjanovic, and Shelly Korba, for help of various important kinds.

I dedicate this book to my mother, Miriam Gilbert, who has so often provided a supportive and peaceful setting in which to write. I hope that the book reflects to some degree her clear-sightedness, honesty, and good sense.

Errata and additional bibliographic material may be found on p. 521.

I

INTRODUCTION: EVERYDAY CONCEPTS AND SOCIAL REALITY

1 PREAMBLE

The relationship between human individuals and the social groups to which they belong has long been disputed. Much passion and puzzlement has been associated with this issue.

What precisely is a social group? Does group membership involve a deep transformation of the human individual? If so, what is the nature of the transformation? It is natural to look to the social studies for answers to these questions, and more generally to find out what a social phenomenon is, as opposed to phenomena of other types. But social scientists have disagreed widely over the proper characterization of their subject matter, beginning with the acknowledged founders of sociology, Max Weber and Emile Durkheim. Durkheim sees social groups as dynamic ‘new’ entities, phenomena sui generis that arise when individual humans associate with one another. Weber shuns such description, urging that it is solely the actions of individual humans that constitute human social life.

Such disagreement suggests that the core phenomenon at issue in the human social domain has not yet been isolated. Yet only when this has been done can we properly understand our situation and our tasks as human social animals. In this book I provide a solution to this problem.

2 THE FOCUS AND THESIS OF THIS BOOK

We think of certain phenomena as ‘social’ phenomena. Here are some examples: two people talking on a street corner; a football game; a meeting of the Town Council; the mob storming the Bastille. But what do we mean by ‘social’? What restrictions are there on what we would consider a social phenomenon? How is the everyday, intuitive concept of a social phenomenon to be explicated? This question prompted the investigations in this book. The title is meant to recall Durkheim’s view that the phenomena particularly apt for the label ‘social’ are significantly special. (Durkheim calls these phenomena ‘social facts’.) I agree that our paradigmatic social phenomena have a very special nature. In this book I explain why. My reasons have something in common with Durkheim’s, though I do not claim to be a Durkheimian.

I shall be parsimonious in my initial assumptions about our everyday concept of a social phenomenon. So as not to exclude in advance certain approaches to the issue, I shall not assume at the outset that ‘social’ means something like ‘having to do with a society or smaller social group’. I shall take it for granted, however, that human social groups and their properties are social phenomena par excellence. So a key question can be put in terms which echo Georg Simmel: ‘What makes a collectivity out of a sum of living human beings?’ I construe this as a conceptual question: under what conditions do we count a set of human beings as a collectivity or social group? I maintain my focus on the question of human collectivities until the very end of the book, where the question of nonhuman animals is considered briefly.

In the main body of the book I focus on three everyday concepts relating to group life among human beings: the concept of a collectivity or social group; the concept of the belief of a group; and the concept of a social convention. I also discuss the concept of a group’s language. Investigation of these concepts leads to important results. These results enable us to say something interesting and precise about the concept of a social phenomenon in general. They also bring into focus a consequential range of phenomena.

The main thesis of the book will be that our collectivity concepts incorporate the concept of a plural subject. The nature of plural subject phenomena will be carefully explained. They are so special, and so apt for the label ‘social’ that one can argue it should be reserved for them. My main aim is not to produce an argument about a label, however. Rather, it is to draw attention to the concept of a plural subject and to the precise nature of plural subject phenomena. Thought about our use of the label ‘social’ can lead to such awareness.

There is little question that plural subject phenomena are of the utmost importance in human life. Awareness of these phenomena and clarity about them is essential for the understanding of many issues in philosophy and elsewhere. These issues include large theoretical questions such as the nature of language, the nature of politics and political obligation, and more practical and detailed questions such as the morality of lying and the conquest of loneliness.

I shall not be able to examine all the implications of the existence of plural subjects in this book. Among other things, however, I argue that the results presented here allow us to adjudicate the longstanding controversy between ‘individualism’ and ‘holism’ about social groups. Many great thinkers, both philosophers and social scientists, have spoken on the different sides of this controversy. On the individualist side, along with Max Weber, we find among others John Stuart Mill, who forcefully asserts, in A System of Logic, that people ‘do not form a new kind of substance’ when they come together in society.¹ But in Utilitarianism Mill describes the members of a social group as the ‘members of a body’, sounding closer then to Durkheim and his view of society as a ‘synthesis’ sui generis, and to those who like Plato in the Republic see group members as united (ideally at least) in a single organism.² The fact that Mill can be cited on both sides is significant. Others also fall into this category. I shall suggest that this is because both types of statement contain an important truth. Each of these truths can be clearly and unequivocally stated, as I hope to show.

The results presented here suggest that one risks the impoverishment of social theory if one ignores plural subject concepts. An understanding of them should do more than improve our understanding of human possibilities. If we are blind to our own possibilities, how can we make good decisions about the conduct of our lives? How can we have a reasonable scale of values, if we do not know what is there to be valued? Since much of what follows will be quite abstract and general, it is worth making this point at the beginning. The investigations undertaken are not simply exercises in pure theory or semantics, or matters of mere logic-chopping. Sometimes logic must be chopped, in order that both the way we think, and the way we are, may be seen clearly.

3 SOCIAL SCIENCE AND EVERYDAY CONCEPTS

My primary aim is to make explicit the structure of certain everyday concepts. The articulated concepts appear well suited for the conceptual quiver of social theory.³ In pressing the claims of these concepts on social scientists I run counter to influential ideas expressed by Emile Durkheim and Max Weber. Let me explain.

In his classic work Suicide, discussing how he will define ‘suicide’ for sociological purposes, Durkheim observes that his own technical notion diverges from what is apparently meant by ‘suicide’ in everyday speech. He defines suicide as action undertaken in the knowledge that one’s own death is a highly probable result. Thus a soldier’s maintaining his position in the front line under orders would count as suicide. Meanwhile, the vernacular notion is, rather, the idea of an act in which one aims at one’s own death. Consistently with this, in his methodological treatise, The Rules of Sociological Method, Durkheim strongly attacks the idea of conducting sociological enquiry in terms of vernacular notions, or ‘preconceptions’. After alluding to Descartes’s method of doubt he writes:⁴

the sociologist, either when he decides upon the object of his research or in the course of his investigations, must resolutely deny himself the use of those concepts formed outside science and for needs entirely unscientific. He must free himself from those fallacious notions which hold sway over the mind of the ordinary person, shaking off, once and for all, the yoke of those empirical categories that long habit often makes tyrannical. (p. 73)

One of the things Durkheim says here is quite plausible: our everyday concepts were not formed for scientific purposes. His methodological recommendation in the light of this, however, is unexpectedly strong: the sociologist must not use these concepts. A more plausible conclusion is that social science could use a careful examination of everyday notions. These will in any case tend to inform all our endeavours, as Durkheim suggests. If we find them wanting or insufficient to our needs we can act accordingly. The possibility remains that on close examination they may surprise us with their complexity, beauty, and theoretical utility. The ‘analytic’ branch of philosophy may be expected to be a useful adjunct to social science in the articulation of everyday concepts.⁵

Possibly under the influence of the founders of sociology, social scientists have tended stipulatively to define a set of technical terms, on the one hand, and, on the other, to use vernacular concepts relatively uncritically and without examination. There are often good reasons to do these things. Meanwhile I want to urge the importance of carefully examining the structure of the vernacular concepts relating to the social world.

My putting Weber in the sceptical camp may seem surprising. His attitude to vernacular concepts is in fact complex. One of Weber’s central tenets certainly dictates that sociologists should steep themselves in these concepts. As is well known, Weber believes that sociologists must attempt to grasp the meaning that pieces of human behaviour have for the persons concerned. More will be said about this view later. But it evidently requires that sociologists both grasp and employ the concepts that are in play in people’s minds when they act. To give a simple example, a man moving his arm in a certain way might be doing this with the intention of saluting his country’s flag. In order to describe the man’s intention correctly, it seems that an observer must himself employ the concept of a salute.⁶

If he would allow that in order to describe people’s intentions sociologists must employ those people’s vernacular concepts, Weber would also insist that in describing other aspects of the world sociologists must adopt a critical stance to vernacular concepts. Such concepts should only be used if they conform to certain scientific requirements.

According to this view, the social scientist may be expected sometimes to be in the position I would be in were I to tell you that 5-year-old William is looking for goblins in the garden. I know what goblins are - or are supposed to be - so I know what I am talking about when I say that he is looking for goblins. I know that William believes that there are goblins, small human-like creatures with wings, and whatever. I myself do not think that there are any goblins, and I do not mean to commit myself to the existence of goblins when I say he is looking for goblins. Unlike William, I could not myself seriously set out to look for goblins in the garden. But I can take an interest in how he goes about it, listen with understanding to his progress reports, and report them back to others.

The view that social scientists should not apply vernacular concepts to the world uncritically seems utterly reasonable. It may be your duty as a developmental psychologist to discover the contours of the concept of a goblin, if many children believe in goblins. But it surely cannot be your duty to share their belief. The same goes for all manner of concepts that the social scientist may need to grasp in order to proceed - witchcraft, phlogiston, and so on. One need not believe what others believe, in general, in order to know that they believe it, to investigate why, and so on. Which concepts one applies to the world in propria persona should be a function of the way one takes the world to be.

As described so far, Weber’s position on the role of vernacular concepts in theoretical sociology seems eminently reasonable. He does not deny the theoretical utility of all vernacular concepts, but presumes that such concepts should not be adopted uncritically. There is a more problematic aspect to his views, however.

4 WEBER ON EVERYDAY COLLECTIVITY CONCEPTS

In his classic discussion of the ‘fundamental concepts of sociology’ in Economy and Society, Weber expresses scepticism about one of the very concepts one would expect to demarcate the subject matter of sociology: the everyday concept of a collectivity. His discussion is both complex and compact and it is not entirely obvious what he means to claim. With that caveat, let us look at the details of Weber’s discussion.⁷

On the one hand, Weber accepts that in everyday life people operate with a number of ‘collective concepts’, such as the concept of a ‘state’, a ‘family’, an ‘army corps’, and so on. In so far as people do operate in terms of these concepts, the sociologist ‘cannot afford to ignore’ them. On the other hand, Weber evidently has some trouble accepting that states, families, and so on, exist. This may seem bizarre. How could one deny that families, for instance, exist?

Weber’s reasons are not entirely clear. One thing that troubles him is that people in possession of the vernacular concept of a family, say, allow themselves to speak as if families themselves can act. They also ascribe rights and duties to them. Possibly Weber thinks it obvious that nothing but an individual human being can act, have rights and duties, and so on. He says at one point: ‘. . .for sociological purposes, there is no such thing as a collective personality which acts’. It seems reasonable to assume that we can read ‘for the purposes of scientific description’ in place of or in addition to ‘for sociological purposes’. It is possible, then, to discern in what Weber writes something like the following line of argument. According to the vernacular concept, a collectivity can act. But then there can be no collectivities in the sense in question.

As I have noted, Weber allows that the sociologist cannot afford to ignore the ‘concepts of collective entities which are found both in common sense and in juristic and other technical forms of thought’. But this is because ‘actors ... in part orient their actions’ in terms of these concepts. In other words, the sociologist will use these concepts in the description of people’s thoughts. Weber also implies that sociologists will undoubtedly find themselves talking about ‘families’, ‘nations’, and so on, because the terminology is at hand, and (I suppose) may seem the natural terminology to use to describe some actual phenomena. However, ‘If purposes of sociological terminology alone were involved’ it would be possible, though ‘extremely pedantic and cumbersome’, to eliminate such terms entirely, and substitute ‘newly coined words’. It seems that we are to understand that in so far as sociologists may say that ‘there are collectivities’ and so on (using terms referring to specific kinds of collectivity), such utterances must be interpreted or handled differently from the vernacular statements of everyday life (and of the law).

Weber writes elsewhere that it is the task of sociology to reduce such concepts as those of a state, an association, and so on, to ‘understandable action, that is, without exception, to the actions of participating individual men.’⁸ What does Weber mean by ‘reduce’ here? Taken out of context his words might be thought to express the view that our common-sense, everyday concepts of ‘collective entities’ can be explicated in a certain way. Given the rest of what he says, the more likely alternative is that the sociologist must replace our common collectivity terms with new ones which are explicable in terms of the actions of individual human beings, the so-called ‘members’ of the ‘collectivity’.

I take it that Weber has one main guiding criterion for an acceptable sociological definition of ‘collectivity’ or of any theoretical term in social science. Reasonably enough, Weber evidently supposes that sociological concepts must be realistic in some sense.⁹ I shall not try to elaborate an appropriate criterion of realism here. Presumably such a criterion will exclude any concept which we know cannot apply to the world, because it is internally inconsistent, for instance.

Weber indicates that he accepts a second criterion, which he seems to see as derivable from a criterion of realism. Witness his remarking that everyday collectivity concepts are to be ‘reduced’ to ‘understandable action’, to ‘the actions of participating individual men’. He makes many similar claims: such as ‘these collectivities must be treated as solely the resultants and modes of organization of the particular acts of individual persons, since these alone can be treated as agents in a course of subjectively understandable action’. We might sum up this criterion as follows: the existence of a human collectivity must be a function of the acts of individual human beings.

In one sense this criterion is unexceptionable. However, it is subtly ambiguous and suggests a criterion against which this book will, in effect, provide an argument. The distinction in question will best be made in the light of ideas I will develop. This issue is discussed fully in the concluding chapter. At this point we may simply note three questions that Weber’s remarks on collectivity concepts in general raise. First, how are our everyday collectivity concepts to be explicated? Second, do these concepts definitely not meet Weber’s criteria for acceptable sociological concepts, as he suggests? Third, are Weber’s criteria reasonable?

This book focuses on the first question. In the light of my answer I shall argue that in so far as the two Weberian criteria noted are reasonable, everyday collectivity concepts do meet them. These concepts are thus freed to be candidates for the role of foundational concepts of social science. I shall argue that from an intuitive point of view this is indeed their proper role. That is, the vernacular concept of a collectivity and its relatives initially locate the concerns of the disciplines most aptly referred to as social sciences. In that sense they will be foundational concepts of social science. This is by no means an unimportant sense. For the concepts which are accepted as foundational in this sense give direction to subsequent enquiry in a given discipline.

5 THE EVERYDAY CONCEPT OF A COLLECTIVITY

It will be useful to preface more detailed analysis with some general remarks about our everyday collectivity concepts.

How do we begin to determine the boundary of our everyday concept of a collectivity? Sociologists and others sometimes make lists of types of social group, of collectivities, or, more grandly, ‘social systems’. For example, Weber lists: a state, a nation, a family, an army corps. Durkheim lists, among the ‘smaller social groups’ which are part of larger societies, religious organizations and literary societies. Some put into their lists some very transient and small-scale affairs, such as a pair of conversationalists on a street corner.¹⁰ I too could make some such list. The conjunction of the above examples, ending with ‘and so on’, will do.

In dealing with such lists the reader is evidently supposed not to balk at an ‘and so on’ at the end. If one is in this position, this suggests that one has grasped a concept or an intuitive principle of some kind, linking all the examples mentioned. What is this intuitive principle? To ask this is another way of asking for an explication of the everyday concept of a collectivity.

Consider the questions: What have typical families in common with schools? What has an army corps in common with two people chatting on a street corner? These questions could be addressed as empirical questions. First we might consider a sample of what we took to be typical families, then a sample of schools, and ask ourselves what in the world they had in common. Evidently, differences might be as salient as commonalities, and it might be quite hard to see where to start such an investigation. But the question about an intuitive principle is of a different order. If it could be given an answer, one would have a special way of answering these other questions: a family and a school have the following in common, in so far as both are collectivities . . . (there would follow an explanation of the nature of collectivities in general).

The question about an intuitive principle is, in effect, a question about what lies behind our listings. On what basis are we willing to put families in, but loth to put in King James the First, or the population consisting of Rudyard Kipling, Julius Caesar, and David Hume? It is clear that most people would not put King James, or the latter trio, in one of these lists. Thus, though there could be disputes about cases, there are also areas of clarity. It should not be said that ‘it’s all a matter of arbitrary choice’.

Let me make a few clear points at once, to fix ideas. It is generally acknowledged, and my example of Rudyard Kipling and the other two randomly chosen people shows, that not just any set of people, in the logician’s sense of set, forms a social group intuitively. Again, it is not enough for certain persons to have some feature or other in common. Thus the population of left-handed people, or the population of women, is not automatically a social group by virtue of having members with a common property. Again, merely inhabiting the same geographical area of the world, perhaps one cut off from others by a surrounding sea or mountain range, is not sufficient to make a population a social group, intuitively. So much, I take it, is common ground. Having said only this much, however, a caveat is in order.

Groups or populations of some of these kinds may be of great interest to demographers and others. For instance, one might be curious to discover whether those with a given feature in common (left-handedness, say) had any other common properties (higher or lower intelligence than right-handed or ambidextrous persons, say). One might be curious to know whether sameness of interests, whatever the particular interests were, tended to make people like one another more, or whatever. So to say that some type of population is not by its nature a social group, intuitively speaking, is in no way to judge the interest of some or all populations of that type from a scientific or other point of view.

A study of some such populations may make good sense within the framework of an interest in the nature and workings of a particular society or type of society. For instance one could usefully study the question of the relative contentment of those at the low end of the economic scale in different types of society. Such a study could be undertaken by someone we are pleased to call a social scientist, without it being the case that the poor in a given society necessarily constitute a collectivity in any intuitive sense. Thus it must be clearly understood at the outset that to argue that populations of certain types are not social groups intuitively implies nothing about when, why, or by whom populations of that type may be studied, nor about their importance in whatever way in the scheme of things. The question here is: when does a population count as a collectivity, and what is especially interesting about this class of populations? I have focused on this question because I think the intuitive concept of a collectivity picks out a phenomenon of the greatest interest.¹¹

6 METHODOLOGY

My approach to the business of conceptual analysis is quite standard. Two points are worth noting at the outset for those unfamiliar with this type of project.

First, I do not simply present my own accounts of the concepts at issue, but will first examine certain plausible alternatives and argue that they are unacceptable. I regard it as an important part of the defence of my own accounts that the inadequacy of these alternatives be clearly seen. In addition, the delineation of alternatives will help to locate the answer I myself propose in what might be called the space of concepts. For every proposed analysis is in effect a definition, the construction of a new concept if not the analysis of an old one. Part of what I want to say about the collectivity concepts that I in effect construct here is: look how different they are from the concepts delimited by other possible analyses.

Second, in assessing a proposed account I will often bring a counterexample against it. This can be a real or an imaginary case in which, say, purported sufficient conditions hold but the concept at issue does not apply. Now, the cases may sometimes seem bizarre or unrealistic, and quite possibly silly, to those unfamiliar with this enterprise and this method of approach. But the main test of the adequacy of a counterexample is not that it describes something which could quite likely happen, but that it describes something logically possible. Mere bizarreness is not a flaw in such examples. They must be taken seriously if our judgments about them contradict a proposed analysis.

We can tell much that we need to know about concepts by telling science fiction tales and such. Some may find it irritating to try to understand anything without constant recourse to the world of flesh, bones, and stones. But nonexperimental, unempirical research can reveal conceptual structures. Once revealed their applicability to the world can be ascertained. Thus the imaginary, the bizarre, and even the silly can help us come to grips with what is real.

There is clearly a question whether, for a given concept, any useful analysis or account in terms of logical entailments or entailers can be given. In advance of work on a given concept, I know of no general argument strong enough to rule out the possibility of such an account a priori.

A note on the terms ‘intuition’ and ‘intuitive’ may be useful. When I speak of our ‘intuitive’ concept of an X, I refer to the concept that is implicit in the judgments we are most immediately inclined to make about what counts as an X and what does not. When I say that something is ‘counterintuitive’ or ‘contrary to our intuitions’, I mean that it does not accord with such judgments. Clearly, ‘intuition’ here is not supposed to be the exercise of some faultless faculty of seeing into the essence of things.

A standard and important route to the analysis or explication of concepts is a consideration of how we use certain key words or phrases, such as ‘social group’, or ‘social convention’. Of course words can be ambiguous. I shall sometimes argue that I have captured a sense of a certain term in English. I may suggest it is the central sense, on the grounds that it is the narrowest and richest sense and other senses can be seen as looser uses. (Consider in this connection the word ‘game’. We may talk about the games people play in their daily interactions, yet at the same time be happy to deny that these are really games.) Or I may speak of a central sense, meaning that I think that there is this sense and it is a distinctive and common one.

I believe that the ultimate interest of what I have to say lies in the articulation of a scheme of vernacular concepts that attaches to important real phenomena. We could possess a given concept without possessing some neat phrase or a single term with which to express it. We could simply tend to agree that certain lists of items looked right; we could have convoluted ways of expressing our views on what the items in question have in common. Meanwhile, the importance of the concepts and the phenomena at issue here would be supported by the fact that they have compact reflections in language. Hence at times I have found it worth arguing that they are indeed so embodied in everyday talk.

7 THE MAIN THEMES

Each of the following chapters represents part of an assault on our everyday concept of a social phenomenon. A major theme running through the whole discussion is: how are we to explicate out everyday collectivity concepts? Until this is decided, we cannot hope to solve the more general problem. Two important issues in this connection are what I shall call singularism and intentionalism. Let me explain roughly and briefly what these are. (Later discussion will help to make things clearer and more precise.)

The issue of singularism concerns a certain restricted conceptual scheme. This may be called the conceptual scheme of singular agency. At its core is the idea of a (human) agent with goals of his own.¹² One acts as a singular agent in so far as one acts in the light of one’s own goals. It may be thought at first blush that the conceptual scheme of singular agency is the only scheme of agency that there can be. That issue will be addressed later. Meanwhile it may be noted that the conceptual scheme of singular agency informs some influential recent attempts to model social processes in terms of the mathematical theory of games. A similar type of modelling is suggested by the scheme of fundamental concepts of sociology that Weber presented in Economy and Society. In the context of an investigation of our vernacular collectivity concepts, singularism is the thesis that these concepts are explicable solely in terms of the conceptual scheme of singular agency. If singularism in relation to our vernacular collectivity concepts is false, then the question of the adequacy of a singularist scheme of technical concepts to describe social processes becomes pressing. The falsity of singularism will be suggested by the results of the next chapter, and later chapters will reinforce this suggestion. My eventual stance, then, will be against singularism. In concluding I shall connect singularism with individualism, arguing against certain precisely specified forms of individualism, as opposed to the contrasting forms of holism.

Intentionalism is the view that according to our everyday collectivity concepts, individual human beings must see themselves in a particular way in order to constitute a collectivity. In other words intentions (broadly construed) are logically prior to collectivities. Intentionalism is surely plausible. Human beings appear to be in an important sense powered by their ideas and views of their situation. Thus many things human may be expected to depend on the existence of certain perceptions and thoughts in people’s minds. Now there are philosophical arguments derived from Wittgenstein to the effect that, as a matter of logic, thoughts or intentions are logically dependent for their existence on the existence of social groups. If the arguments are valid, and intentionalism about everyday collectivity concepts is true, these concepts will be incoherent and there will be nothing to which they apply. I shall argue that intentionalism about everyday collectivity concepts is true, but that this does not mean these concepts are incoherent.

As the falsity of singularism would throw doubt on the adequacy of game theoretic and related approaches to social phenomena, so the truth of intentionalism would suggest the inadequacy of approaches in terms of externally observable structures or systems. Some theorists would propose that the essential mark of a human collectivity is an externally perceivable organization or structure, or systematic relations between the parts, that is, the human members. While there is doubtless something to be said for the view that human collectivities may be viewed externally as systems or structures, I shall argue that according to our intuitive conception they are not systems from an external point of view only. The participants must see themselves as bound together in a highly specific way. This suggests that an explanation for the existence of objective systemic features in collectivities may be found in an internal sense of unity. This would not be surprising, in so far as we accept that human beings are powered by their own perceptions of the situation. There is therefore reason to suppose that the perceptions indicated in our everyday collectivity concepts are the real ‘glue’ of the human social world.

Clearly, it is not enough to argue for the coherence of intentionalism. In order to argue convincingly for an intentionalist account of everyday collectivity concepts, one must attempt to specify precisely what thoughts, conceptions, and perceptions, are the binding glue of collectivities according to that conception. Here my anti-singularist, pro-intentionalist stance finds its positive basis. I argue that people must perceive themselves as members of a plural subject. As will become clear in the course of the book, this thought takes them (and the observer) beyond the conceptual scheme of singular agency.

So much, then, for the main themes. In the next section I sketch the contents of each chapter. The chapters are organized into a pointful sequence and together constitute a lengthy argument. At the same time, each one but the last is to a large extent self-contained. Different readers may wish to concentrate on different chapters. Although each chapter revolves to some degree around the ideas of a particular sociologist or philosopher, this is not primarily a work of exposition or comparison of texts. I have attempted, rather, to come to grips with the issues I am interested in by taking provocative, central statements in the work of others, and discussing and developing them in my own terms.

8 OVERVIEW OF CHAPTERS

Chapter II focuses on Max Weber’s influential technical concept of ‘social action’. Roughly, a person performs a social action when he acts with someone else in mind. For instance, if I tell you something, if I wave goodbye to you, I perform a social action in Weber’s sense. In Economy and Society Weber claims, in effect, that his concept of social action is the fundamental concept of sociology: it is the existence of social actions in his sense that provides social science with a subject matter. Of course there are social actions of many different kinds, and many different types of complex of social action are possible. The sociologist is to develop categories which pick out types or complexes of social action that are deemed to have some special importance. Weber himself articulates a number of such categories. However, the fundamental social fact, if you like, is the fact of social action itself. Weber’s view has been influential. Numerous writers in the social sciences and the philosophy of social science write as if ‘social actions’ in something like Weber’s sense are the central subject matter of social science.

Weber’s placement of his concept of ‘social action’ at the heart of his scheme of sociological concepts is importantly suggestive in two ways. It implies that sociological concepts will appeal to human intentions. It also suggests that the conceptual scheme of singular agency is adequate for sociology.

In Chapter II I consider whether an effective unitary characterization of sciences which could be called social sciences with intuitive aptness can be given in terms of Weber’s concept of social action. I argue that this is not so. I argue that one thing that should trouble us is that Weber’s concept does not enable us to explicate the vernacular concept of a collectivity. I argue for its insufficiency in this respect by considering and rejecting a number of possible simple accounts of a social group using Weber’s notion. Clearly, the relation of Weber’s concept to our collectivity concepts will best be ascertained when a satisfactory analysis of those concepts has been given. Meanwhile, the ground is cleared here to the extent that certain attractively simple analyses in terms of Weber’s concept do not work.

In concluding I argue that it is doubtful whether an action social in Weber’s sense is naturally thought of as a ‘social’ phenomenon in itself. Paradoxically, Weber’s vision of human life gives every indication of leaving out its intuitively social dimension. In my view, it is not Weber’s allusion to intentions that creates the problem, but rather the implied suggestion that sociology stick to the conceptual scheme of singular agency.

As I have indicated, faith in the adequacy of the conceptual scheme of singular agency is still alive in modern social science and is exemplified in attempts to model social processes in terms of the mathematical theory of games. I take a further look at the consequences of this approach when I examine - and reject - the influential game-theoretical account of social conventions proposed by the philosopher David Lewis, in Chapter VI. Chapters IV and V will also consider accounts of social phenomena which embody the limited perspective of singular agency.

There are traces in Weber’s writing of the view of social phenomena that I shall put forward here. Why did this view not surface more? I am inclined to conjecture a disinclination to acknowledge certain facts, a disinclination which is expressed in Weber’s animadversions against the existence of collectivities which act. There is doubtless a sense of ‘action of a collective’ such that Weber’s negative existential claims in this regard are warranted. However, it will be the main thrust of this book that there is an important and theoretically respectable sense in which collectivities can act, and, indeed, think, have attitudes, and hold to principles of their own. I am in little doubt that Weber was aware of that sense at some level and that it informed his own daily life. So it would not be surprising if this awareness manifested itself at times in his writing. I say this not because I know anything significant about Weber’s daily life, but because I think we all know this at some level. Normal human beings regularly, if not always and inevitably, see themselves as members of a collective agent or, more generally, of a collectivity with specific psychological properties. Moreover, they constantly act in terms of this perception. This book as a whole will be an explanation of why I say this. In the concluding chapter, I explain how the structure of our intuitive collectivity concepts makes it understandable that someone might wish to deny the existence of collective psychological properties.

Recall that intentionalism concerning our collectivity concepts is the view that an explication of these concepts will appeal to human intentions. In Chapter III I defend the coherence of intentionalism in general against some well-known arguments derived from Wittgenstein’s work. Many whose primary interest has been the philosophy of social science have first encountered Wittgenstein’s thought through the interpretation in Peter Winch’s monograph The Idea of a Social Science and its Relation to Philosophy. Winch claims that Wittgenstein has shown it is logically impossible for a being outside society to have intentions. That is, our concept of a being with intentions involves our concept of a society or social group. This would clearly be a problem for intentionalism. Winch’s understanding of Wittgenstein has points of contact with that of many others. As I shall explain, however, the argumentation he presents is inconclusive. At a crucial point he assumes rather than argues for the unavailability of an acceptable ‘nonsocial’ model of language and, ultimately, of intention.

Not only is this assumption crucial to Winch’s argument for the society-dependence of both language and thought. It is also deeply disturbing. For, as I argue, our everyday notions of language and thought imply that an acceptable nonsocial model is indeed available. If Winch’s assumption is true, there can be no such thing as language or thought as intuitively conceived. Though Winch does not see the import of his assumption this way, it does seem to have this powerful consequence. In other words, if Winch is right about this assumption, then language and thought do not have a social nature; they are simply impossible: they cannot exist.

Is there a philosophically acceptable nonsocial model of language and thought, then, or is there not? This was a main focus of Wittgenstein’s concerns, according to the interpretation proposed in Saul Kripke’s Wittgenstein on Rules and Private Language. Kripke derives from Wittgenstein powerful arguments against a nonsocial or ‘private’ model. I argue that in spite of the power of these arguments one type of private model has not been shown to be incoherent and can be argued, then, to be philosophically acceptable.

Kripke suggests that rather than concerning himself with the formulation of necessary and sufficient conditions for thought and meaning, or with the analysis of the concepts of thought and meaning, Wittgenstein prefers to drop such an enterprise and consider what ‘language game’ we play with words such as ‘thought’ and ‘meaning’. I argue that were we to accept this approach, it would be possible to accept a form of intentionalism. My conclusion in relation to intentionalism, then, is that it can be allowed to stand, as far as the arguments in Winch and Kripke go. Thus the ground has been cleared for an intentionalist approach to the logical construction of collectivities.

Though I do not accept a society-dependence thesis of thought and language in general I believe that a full understanding of the nature of human linguistic phenomena demands special consideration of what it is for a group, as opposed to an individual, to have a language. Indeed, awareness that a radical distinction can be made between a group’s language and that of an individual may help to defuse a fairly prevalent sense that language as such is society dependent, and a possibly consequent judgment on the nature of thought. I therefore conclude Chapter III with a brief account of the concept of what I call a group language. This is a subtle and important collectivity concept. To my knowledge, it has not been explicitly noticed, let alone examined with any care, in previous literature on Wittgenstein or on language in general. I shall explain how Wittgenstein’s remarks on criteria, which have puzzled many commentators, are strikingly appropriate to an account of a group language, as opposed to the language of an individual. (I do not claim that Wittgenstein was prepared to make the distinction I am stressing here.) A fuller understanding of group languages requires an understanding of the concept of a social group, the topic of Chapter IV.

Chapter IV is perhaps the core of the book. I develop a new account of our concept of a human collectivity, inspired by a remark of the sociologist Georg Simmel’s in ‘How is society possible?’. Simmel says that humans must see themselves as unified in some way in order that they constitute a collectivity. This is clearly an intentionalist view. It is distinguished by the precise content of the thought supposed to be crucial. The account I develop from this beginning has points of contact with other authors also, in particular with Rousseau.

I argue that our concept of a collectivity is the concept of a plural subject of action, belief, attitude, or other such attribute. Such subjects exist when people do things together, for instance. A very simple example of people doing things together is that of two people going for a walk together. (I am happy to be committed to the existence of small, ephemeral collectivities. I believe that these are the bedrock of human social life.) I argue that in order to do things together people must view themselves in a certain special way.

I discuss in some detail what it is for people to do things together, or to share in an action. After considering various possible conditions on sharing in an action I conclude that it is necessary that the participants express to each other willingness to be part of a plural subject of a certain goal, for instance the goal that A walks in the woods and B walks in the woods and that they do so in one another’s company. Their being part of a plural subject of this goal is, importantly, not equivalent to each one’s accepting the goal as his own.

There are various ways of attempting to say what it is to express willingness to be the plural subject of a goal but none of these appears to involve what would normally be thought of as an analysis. For instance, one might say that each participant must manifest willingness jointly to accept responsibility for achieving the goal in question. But such willingness is in its turn best explicable as willingness to constitute a plural subject of the goal.

What I have said so far, then, does not give us a definition of what it is to be the plural subject of something. Rather, it gives us a logically necessary condition for the existence of such a subject. The statement of this condition itself uses the notion of a plural subject. It is a notion the members of a plural subject must have in order to become members of a plural subject. For they must experience and manifest willingness to be members of such a subject.

To illuminate the plural subject notion further we must turn to a description of a kind used by Rousseau. One is willing to be the member of a plural subject if one is willing, at least in relation to certain conditions, to put one’s own will into a ‘pool of wills’ dedicated, as one, to a single goal (or whatever it is that the pool is dedicated to). It is logically sufficient for the existence of such a pool that it be ‘common knowledge’ among the people in question that, roughly, everyone has expressed his personal willingness that his own will be part of it. (Common knowledge is a technical term that will be defined later.)

I argue that the English pronoun ‘we’ has a central sense in which it refers to a plural subject. Its referents are not only the plural subjects of goals. ‘We’ is also properly used at least of those who are ready to share in certain actions when the time comes, or jointly ready. Joint readiness involves a plural subject. This makes it more plausible to define a social group in terms of ‘we’. Intuitively speaking, not all social groups involve the current pursuit of a goal or end, in spite of the impression given by some writers.

I conjecture that plural subjecthood is the crucial constituent of a social group. In brief, social groups are plural subjects. In the following two chapters plural subject phenomena other than joint readiness for action or actual joint action are discussed in detail. Further consequences and aspects of plural subjecthood are discussed in the concluding chapter.

In the remainder of Chapter IV I address various aspects of my account of collectivities. These include the processes of group formation, the relation of group formation to language, the crucial minute phenomenon of mutual recognition, the possibility of large groups and inactive groups, the occasionally striking phenomenology of plural subjecthood. I reply to a set of possible objections, and note the relation of my account to the remark of Simmel’s which provoked it.

Chapter V addresses the topic of a group’s beliefs. I do not assume at the outset that the account of a social group proposed in the previous chapter is correct. This enables the discussion here to serve as a test of the idea that social groups are plural subjects. I start with the fact that we often say things of the form ‘Group G believes that p’. Under what conditions are such locutions most appropriate intuitively? Which actual phenomena most clearly warrant this kind of description?

It is widely assumed that any account of collective belief must be ‘summative’. As I use the term, a summative account of collective belief holds it to be a necessary condition of a group’s believing that p that all or most members of the group believe that p. A weaker requirement on an account is that it merely be ‘correlativist’. A correlativist account holds it to be a necessary condition of a group’s believing that p that at least one member of the group believe that p. As far as they have been characterized here, correlativism and summativism are consistent with a singularist view of collectivity concepts. In an exegetical passage of some length I argue that Durkheim, in The Rules of Sociological Method, held an account that was not even correlativist, let alone summative. I shall myself argue for a nonsummative, noncorrelativist stance.

I give summative theory a good run for its money. I start with an account that many would accept, what I call the simple summative view. This holds that in order for a group to believe that such-and-such it is both necessary and sufficient that most members of the group believe that such-and-such. I consider two other summative accounts, each one more complex than the preceding account, whose problems it attempts to avoid. One account brings in a common knowledge condition, and one is a ‘causal’ account. (‘Causal’ is in quotation marks since the causal nature of the account could be disputed.) Before describing and assessing these accounts I formulate some intuitive tests by which the aptness of the label ‘belief of a group’ can be assessed.

Summativism does not do well on the intuitive tests. A detailed example of a situation in which a group may be held to have formed a belief is then presented (the case of the poetry discussion group). Given the details of this case I argue that correlativism is false as a general theory of the conditions under which a group believes something. A fortiori, so is summativism.

I argue that the poetry group example and other data considered support a radically different type of account of collective belief: roughly, a group belief is a jointly accepted view. That is, it is a view that each of a set of persons has shown willingness to accept jointly with the others. A jointly accepted view is not to be confused with a view that all or most of the people concerned personally accept. One may accept a view jointly with certain others but not believe it personally.

That a certain view is jointly accepted by certain persons is an extremely important type of fact. It can influence what people themselves think, and in any case it constrains their behaviour in important ways. I conjecture that the phenomenon I characterize under the label of ‘jointly accepted view’ is both widespread and influential. It would be too bad if misinterpretation of everyday language, or anything else, resulted in obliviousness to this phenomenon.

As it turns out, the account of a group’s belief arrived at here supports the account of a social group proposed in the previous chapter. That which jointly accepts a view is a plural subject. Chapter VI proceeds in much the same way as Chapter V. It consists of a relatively independent and self-contained investigation of the everyday concept of a social convention. I first present a detailed critique of David Lewis’s influential game-theoretic account of social convention. (There is a brief introduction to game theory for readers who are unfamiliar with it.) This is an account of a social convention as, in effect, a ‘certain kind of sequence’ of ‘actual and possible social actions’ of certain specific types. (The quoted phrases are from Weber.) I then consider an alternative which resembles a number of the available accounts of ‘social norms’ or ‘social rules’ in making no appeal to the concept of a plural subject. (In H. L. A. Hart’s influential account in The Concept of Law, for instance, there is no explicit appeal to plural subjects. What is only implicit is hard to gauge.) Both accounts are argued to be inadequate as accounts of the intuitive concept of a social convention.

Having arrived on the basis of discussion of these accounts at a precisely articulated set of criteria for evaluating an account of social convention, I propose a new account which satisfies the criteria better than the others do. According to my account a social convention is a principle of action jointly accepted by the members of some population, a principle of the ‘flat’ form. Intuitively, a population which jointly accepts such a principle thereby constitutes a collectivity. This is an intuitively appealing result from the point of view of social convention, and at the same time helps to confirm the plural subject view of social groups. For by virtue of joint acceptance of a principle members of a population constitute a plural subject. In the course of this chapter the connections of social conventions to a number of other phenomena are investigated. In particular, I distinguish between social conventions and what I call ‘linguistic conventions’, and between a group’s morality, its conventions, its agreements, and its laws.

In the concluding chapter I summarize the positive results of the previous chapters and then turn to some applications of these results. I discuss three main issues. The first is Weber’s proposed constraint on admissible collectivity concepts in sociology, to the effect that these are to be analysable in terms of the ‘actions of participating individual men’. The second is the general issue of individualism versus holism about social phenomena. The third is the question with which this book began: the nature of social phenomena in general as intuitively conceived. The thesis that we have and live in terms of the conceptual scheme of plural subjects has implications for all these issues, and more. I explain by brief reference to examples why I take our possession of this scheme of concepts to be relevant to political philosophy, and to psychology.

Should the reader wish to concentrate on my own positive accounts of the various collectivity concepts considered, this material is largely to be found in the following parts of the book: