Social Functions in Philosophy: Metaphysical, Normative, and Methodological Perspectives [1 ed.] 1138351326, 9781138351325

Citation preview

Social Functions in Philosophy

Social functions and functional explanations play a prominent role not only in our everyday reasoning but also in classical as well as contemporary social theory and empirical social research. This volume explores metaphysical, normative, and methodological perspectives on social functions and functional explanations in the social sciences. It aims to push the philosophical debate on social functions forward along new investigative lines by including up-to-date discussions of the metaphysics of social functions, questions concerning the nature of functional explanations within the social domain, and various applications of functionalist theorizing. As such, This is one of the first collections to exclusively address a variety of philosophical questions concerning the nature and relevance of social functions. Rebekka Hufendiek is an Assistant Professor at the University of Basel. Her research interests lie in philosophy of mind and philosophy of science, particularly in empirical and ideological dimensions of research on cognitive and behavioral features. Her book Embodied Emotions: A Naturalist Approach to a Normative Phenomenon (Routledge 2015) provides a noncognitivist theory of emotions. Daniel James is a Postdoc at Heinrich Heine Universität Düsseldorf. His historical research concerns the intersection of Hegel’s metaphysics with his political philosophy. With a view to contemporary debates in social philosophy, he is interested in the concept of social power and its fruitfulness for social-scientific inquiry, as well as in social dispositions and their connection to social-structural explanation. Raphael van Riel holds a position as a Senior Lecturer at the University of Duisburg-Essen, where he directs a research group that focuses on theories of explanation. In his book The Concept of Reduction (2014), he offers a novel explication of reduction claims in the philosophy of mind and the philosophy of science.

Routledge Studies in Contemporary Philosophy







For more information about this series, please visit: https://www. routledge.com / Routledge-Studies-in- Contemporary-Philosophy/ book-series/SE0720

Social Functions in Philosophy Metaphysical, Normative, and Methodological Perspectives Edited by Rebekka Hufendiek, Daniel James, and Raphael van Riel

First published 2020 by Routledge 52 Vanderbilt Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2020 Taylor & Francis The right of Rebekka Hufendiek, Daniel James, and Raphael van Riel to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data A catalog record for this title has been requested ISBN: 978-1-138-35132-5 (hbk) ISBN: 978-0-429-43539-3 (ebk) Typeset in Sabon by codeMantra

Contents

Preface

vii



1 The Empirical Issues in Functional Explanations in the Social Sciences

18

H A RO L D K I N C A I D

2 Do Organizations Adapt?

28

DA N I E L L I T T L E

3 Social Dysfunctions

45

H E I N E R KO C H

4 In Search of the Missing Mechanism. Functional Explanation in Social Science

70

R A P H A E L VA N R I E L

5 From Natural Hierarchy Signals to Social NormEnforcers. What Good Are Functional Explanations of Shame and Pride?

93

R EBEK K A H U FEN DIEK

6 What Explains Social Role Normativity?

122

CH A R LOT T E W I T T

7 The Social Function of Morality A N DR EAS MÜ LLER

135

vi Contents 8 The Function of Gender as a Historical Kind

159

M A R I M I K KO L A

9 Function Without Intention? A Practice-Theoretical Solution to Challenges of the Social Domain

183

A MR EI BA HR





Notes on Contributors Name Index Subject Index

247 249 253

Preface

This book is the result of seven years of ongoing philosophical exchange between philosophers interested in such diverse areas as social philosophy, philosophy of social sciences, philosophy of biology, philosophy of mind, metaphysics, philosophy of language, ethics, feminist philosophy, and Marxism. The book was motivated by the hunch that social functions play an important role in many social theories but are strangely undertheorized at the same time. Furthermore, existing reasoning about social functions can be found in such different places as functionalism in the social sciences, analytic Marxism, and debates on explanations in the philosophy of science. Our idea, therefore, has been to bring together philosophers with expertise in the relevant subfields and organize an ongoing discussion on the normative, metaphysical, and methodological questions that can be asked about social functions. The German Research Foundation generously funded the research network “Social Functions. Normative, ontological and explanatory aspects of functions in social philosophy” between 2015 and 2019. During this period, members of the network – Amrei Bahr, Miguel Hoeltje, Rebekka Hufendiek, Daniel James, Heiner Koch, Mari Mikkola, Andreas Müller, Raphael van Riel, and Markus Wild – were able to meet on a regular basis. Moreover, members of the network organized a lecture series and three international conferences on the topic. Many of the invited speakers, friends, and colleagues provided feedback and inspiration over the years. We would like to thank all those who contributed in one way or another to the project and this volume, in particular: Johannes Achatz, Brian Epstein, Sally Haslanger, Frank Hindriks, Wybo Houkes, Ludger Jansen, Daniel Little, David Livingstone Smith, Beth Preston, Peter Railton, Hans-Bernard Schmid, and Tommie Shelby for the talks they delivered during workshops and conferences organized by the network; Kai Ploemacher and Christoph Thies for their proofchecking and formatting under stressful conditions (needless to say: all remaining errors are the sole responsibility of the editors); and in particular: Miguel Hoeltje, Mari Mikkola, and Markus Wild for intense discussions and feedback at every step of the project. Berlin/Bern 12.2019

Taylor & Francis Taylor & Francis Group

http://taylorandfrancis.com

Introduction Rebekka Hufendiek, Daniel James, and Raphael van Riel

Why Social Functions Matter When we talk about an object’s function, we usually have in mind a particular type of effect of an event the object is involved in, which is beneficial in some way, either for the object itself or for something else. Social functions are effects that are beneficial at a societal level. Social functions thus appear to differ from paradigmatic functions of human artifacts, or functions of human organs, whose effects are located at the individual level. Eating, watching television, or doing sports seem to serve some individual needs. But they may also have some social function  – like maintaining society as a whole or contributing to the well-being of the public. Typically, however, when we consider social functions, we consider not only functions whose effects concern the societal level, but, rather, functions whose bearers are social objects, or at least involve some social dimension. Classifying social objects in terms of their functions plays a crucial role for our conceptualization of social reality. It is common wisdom that the function of schools and universities is to educate, that the function of money is to serve as a means of exchange, and that the function of newspapers is to distribute information. Prima facie, it seems that many ordinary conceptions of social reality involve function ascriptions. At least sometimes, we classify things in terms of the social functions they have or are believed to have (think of the various institutions and procedures in a democracy which are involved in an electoral system; or the mechanisms of state intervention in market economies). An investigation into the nature of social functions may thus inform us about some aspects of our ordinary conception of social reality, but also, to the extent that this conception gets it right, about social reality itself. An important idea in the history of the social sciences, and also in philosophy, has been that some alleged social functions need to be uncovered. It need not be an a priori matter that the function of schools is to educate, but it is not some deeply hidden fact about social reality either. Whereas some alleged functions of institutions seem to be due to explicit intentions of those who “designed” these institutions, some

2

Rebekka Hufendiek et al.

authors have suggested that many social functions are due to processes beyond intention or design and are, as a consequence, more difficult to identify. So, identifying functions or offering well-grounded functional explanations may inform us about the structure of social reality, but doing so may be a difficult task. One crucial question here concerns the concept of a benefit. In many cases, we may be in a position to spell out the concept of a beneficial effect in terms of an effect which meets success conditions determined by some particular intention or goal. But sometimes, beneficial effects are independent of explicit intentions or goals. Social traits were considered to meet universal human needs (Malinowski 1944); they were considered to be in the (objective) interest of some groups in a given society and, via satisfying these interests, were thought to contribute to their own continued existence; and they were taken to contribute to the existence of a social whole (Radcliffe-Brown 1952). All these accounts, however, raise further philosophical questions and problems. The concept of a benefit is key to understanding another reason for why social functions have been at the center of attention in the social sciences for quite some time, a reason already hinted at above: uncovering social functions may give an insight into dependencies among social processes or institutions and, thus, provide us with explanations. At a very abstract level, we can put this point as follows: as we have just seen, beneficial effects may be construed as effects which contribute to the continued existence of some social trait, of some social object, or satisfy presumed universal human needs (which, if necessary for the continued existence of the human species, may hence contribute to the continued existence of society as well). Uncovering such dependencies will provide us with an explanation of why some social trait continues to exist; it may shed light on the stability of a system, and it may, potentially, provide us with some general understanding of the organization of society. Finally, in addition to purely descriptive purposes, we may resort to social functions when we justify social facts or subject institutions to moral criticism. Political exchange about the adequacy of an institution will typically involve reflection on the expected functional effects of the institution. At the same time, appeal to social functions may become a forceful element of social critique, namely, when we can show that an institution does not live up to its own standards, i.e. when institutions fail to fulfill the function which would otherwise justify their existence. It is easy to find examples where it may not be obvious whether an institution really serves its official purpose (if any), rather than some other, quite different and perhaps not so pleasant purpose; consider, for instance, particular behavioral patterns of intellectuals (Bourdieu 1979) or social inequalities (Tilly 1998). It is because of these aspects that we may care about social functions. We conceive of social reality in terms of social functions; yet uncovering

Introduction  3 functions or offering solid functional explanations is a difficult affair. The normativity involved in the concept of a beneficial effect raises ontological questions. At the same time we often justify the existence of institutions or patterns in terms of their (alleged) functions, and appeal to functions may support various forms of critique. Further questions arise concerning the history of function talk in the social sciences and its relation to biological functions. We now turn to a quick look at the history of social functionalism.

History – A Short Overview Historically, the notion of a social function traces back to the emergence of sociology and anthropology as autonomous areas of inquiry in the 19th century (although it has its precursors in the theories of 18th-century Enlightenment philosophers such as Millar and Ferguson in Scotland and Condorcet and Turgot in France, see Harris 1968). This period is also considered the heyday of social functionalism, which, in the narrow, widely accepted sense of the term, encompasses a large family of macrosociological theories arising in the mid-19th century broadly unified by the aim to account for various parts of a social system in terms of their function for the maintenance or stability of this system as a whole (for an overview of the varieties of social functionalism, see Turner and Maryanski 1979). In a broad sense, social functionalism encompasses any theory that aims to explain the nature, existence, persistence, or resilience of some entity in terms of its social function, so as to include not only theories historically associated with this term, but also many sociological and philosophical theories that do not operate under this name. Several threads of the history of social functionalism in this broad sense are tightly intertwined with the analogy between biological and social phenomena that underpins the very idea of functional explanation in social science. Broadly, we can distinguish two versions of this analogy, depending on which account of functional explanation in biology serves as its source. The first version is the analogy between social entities and living beings, both maintaining themselves through the maintenance of their parts. The second version is the analogy between social function bearers and biological traits, both resulting from adaptation to a given environment. The former version underlies an organizational view, and the latter a selectionist view of social functions and functional explanation in the social domain. Although social functionalism is a fairly recent theoretical development, the organismic version of the biological analogy underlying early social functionalism has a far longer historical pedigree and is, indeed, just about as old as the concept of an organism itself, tracing back to ancient and, later, early modern theories of the “body politic” (for an overview of the history of this analogy, see MacLay 1990). However, not

4

Rebekka Hufendiek et al.

until the early 19th century was this analogy elaborated into a full-blown model of society intended to support a particular kind of explanation. This endeavor has resulted in a view of social functions which assumes that societies are – analogous with living beings – self-maintaining systems and accounts for social functions in terms of an objective teleology. We can find an early version of this view in G. W. F. Hegel’s Philosophy of Right. Analyzing Hegel’s account as a predecessor of what is nowadays called the organizational view is the main subject of Daniel James’s contribution to this volume. Although Hegel rarely figures among the protagonists in the dominant narrative about early social functionalism, his theory arguably influenced the social theories of several among those authors who undoubtedly are. These theories include Karl Marx’s conception of capital as an “organic unity”, in particular as presented in the Grundrisse (see Marx 1858 – cf. Meaney 2002). Somewhat more controversially, they may also include Émile Durkheim’s conception of society as a functionally integrated system, and Talcott Parsons’s view that shared internalized social norms explain the stability of society (see Durkheim 1902; Parsons 1951; cf. Knapp 1982, 1985, 1986). However, despite its – more or less hidden – influence on his early functionalist successors, many of them took issue with the objective teleology implied in the version of the organizational view developed by Hegel. Thus, we can identify one crucial thread extending from Hegel with the effort to exorcize the apparent teleological implications of the organizational view (see Turner 1986, 21–22) – an effort that still finds its echo in contemporary debates over functional explanations in the social sciences. Among the early social functionalists making use of the organismic analogy, most ink has arguably been spilled on the one who prominently coined the term “social organism,” that is, Herbert Spencer (see Spencer 1893). Interestingly, Spencer is also widely regarded as one of the founding figures of social evolutionism, which had its heyday in the latter half of the 19th century and faded after the turn of the century.1 It is, however, by no means clear whether any of the early theories of social evolution also amounted to an etiological theory of social functions in the sense that has been the focus of contemporary discussions (see “Functions in Philosophy of Science and Philosophy of Biology” section). For, although many of these theories, at least implicitly, relied on some notion of adaptation they often tied to social functions, and, in some cases, also identified causal mechanisms that facilitated social evolution, it is doubtful whether they understood this notion in the sense of a social selection mechanism (Giddens 1981). Indeed, as some commenters have – often critically – remarked, most, if not all of these theories were arguably committed to a developmentalist (as opposed to a selectionist) understanding of social evolution, meaning that they took social evolution to be a process in which some disposition of the society in question manifests itself. This is true even of those early social

Introduction  5 functionalists who, such as Marx, explicitly invoke Darwin (Sanderson 2016). For this reason, it would be anachronistic and, indeed, misleading to confound talk of social evolution among the early social functionalists with a specific understanding thereof modeled after Darwinian evolution. 2 The same is arguably true even for later social functionalists such as Parsons (see Gould 1977; van Parijs 1981). Despite this overall commitment, some of the classic social evolutionists already identified specific causal mechanisms that facilitated evolutionary social change and, thus, anticipated the contemporary selectionist view.3 However, this view of social functions would not be elaborated before the revival of social evolutionism in anthropology against the background of the so-called “modern synthesis” of Darwinian natural selection and Mendelian genetics in the 1930s and 1940s. Social anthropologists, like Radcliffe-Brown and Malinowski, did not try to uncover selectionist processes; instead, they engaged in what is today often interpreted as functional analysis (see “Functions in Philosophy of Science and Philosophy of Biology” section). The goal was to analyze the functional organization of a social system. Individual cultural practices, institutions, or traits were considered to contribute to the stability of a social system, without being recognized by those involved in the system and thus bypassing explicit goals or intentions of individual agents. A now-famous example is that of the rain dance (also employed by Robert K. Merton). People gathering to perform this dance were hoping for rain. Gatherings of this sort, however, were assumed to contribute to social stability by reinforcing group identity – that was considered their function, independently of whether gatherers held a corresponding belief or intention. As Raphael van Riel points out in his contribution to this volume, Radcliffe-Brown took the hypothesis that social wholes are functionally organized systems to be a viable hypothesis. The hope was to identify laws of causal organization that are general in the sense that we encounter them across all societies. Functionalism, in this sense, was thus closely tied to a conception of the social sciences oriented towards the virtues of the natural sciences, and functional analysis was supposed to uncover the laws of social organization. In a slightly different spirit, Malinowski (1944) developed a view according to which social institutions serve universal human needs, such as hygiene. Radcliffe-Brown and Malinowski were both opposed to historical approaches to culture and society; the actual development of social institutions did not play a role in their work. Accordingly, functional analysis did not serve the purpose of explaining the etiology of social institutions. As van Riel argues, Radcliffe-Brown thereby manages to offer social functional explanations without committing to a selection mechanism of the sort criticized by Elster (see “Functions in Philosophy of Science and Philosophy of Biology” section).

6

Rebekka Hufendiek et al.

More recently, theories of social evolution have become prominent again. In his contribution to this volume, Matthieu Queloz highlights an underappreciated tradition of “pragmatic genealogy” that traces back to authors who describe fictional states of nature ranging from David Hume and Thomas Hobbes to Miranda Fricker and Philip Pettit. Such pragmatic genealogies end up ascribing social functions to particular concepts or practices which one does not necessarily expect to have a social function at all. On Queloz’s interpretation, these genealogical explanations are dynamic models that primarily reveal and ground ascriptions of functionality, but which can then be used in more or less ambitious ways to explain the resilience or even the persistence of concepts or practices. Idealizing a bit, we thus arrive at the following picture: social theory which invokes notions of social functions has been associated with (i) selectionist, (ii) developmentalist, (iii) organizational, and (iv) analytical approaches, the former, (i) and (ii), taking a historical perspective, whereas the latter, (iii) and (iv), do not. Often, social functions have been regarded as a means to assimilate the social sciences to the natural sciences. Once the view that we can theorize societies in analogy to organisms became suspect, more local versions of social functionalism have been developed. These developments raise the question of what social functions are, and what theoretical or explanatory role reference to social functions may play. Issues concerning the explanatory and theoretical role of functions, in general, have mainly been discussed in connection with biological functions. The next section will thus broaden the perspective to provide some theoretical background on philosophical theories of functions in general, many of which have been inspired by the philosophy of biology.

Functions in Philosophy of Science and Philosophy of Biology It is widely agreed that in some cases, functionality arises from intentions or intentional actions of individuals. And some functions in the social realm may arise straightforwardly from intentions associated with the function bearer. Dan Little, in his contribution to this volume, focuses on how interacting though different intentions of individuals determine the ability of organizations to adapt to novel situations. The guiding idea is that multifaceted individual intentions give rise to various mechanisms, which may determine the potential for organizational change. Functionality which is independent of intentions or goals has received particular attention by philosophers, in particular in the philosophy of the life sciences. Biologists and physiologists ascribe to the heart not only the causal property but also the function of pumping blood. They use terms such as “neurotransmitter” or “receptor” to describe entities

Introduction  7 not in terms of size, shape, and motion, but of their functional role in the behavior of a system (Craver 2013). Psychologists and anthropologists often provide functional explanations of cognitive and behavioral human features. For instance, it has been suggested that punishing behavior has the function to maintain and stabilize cooperation (Bowles and Gintis 2011), or that cultural learning has the function to quickly adapt to information-poor environments (Henrich and Henrich 2007). Accordingly, both general philosophy of science and philosophy of biology have a long history of inquiring into the nature and the value of such function ascriptions and functional explanations. Two different views of functional explanation have been particularly prominent in recent decades: the etiological theory and functional analysis (or causalrole theory). The etiological theory holds that the function of an entity is what explains why it is there or why it has the structure that it has (Wright 1973). Most etiological theorists tell a story about the causal origin of the function bearer and, thus, commit themselves to a particular kind of feedback relation: X is part of a system S and does Y. By benefiting system S, Y leads to the reoccurrence of X. Accordingly, to have a function, an item must instantiate a type whose tokens are (or once were) disposed to have certain effects. A functional item must have the right kind of history, in which the effects of the items have led to the reproduction of the type they are tokens of (Millikan 1984, 1989; Neander 1991a, 1991b, 2017). The core idea is that the function of an item is to do what it was selected for. In functional explanations of this kind, phylogenetic natural selection is the standard mechanism assumed to underlie such selection processes (however, other kinds of selection like cultural selection might explain selected functions as well). Etiological explanations of social items are highly prominent in evolutionary psychology and evolutionary anthropology but have often been criticized for their conjectural nature. Functional analysis is the view that takes functions to be causal roles performed by components of complex systems. Functional analysis demands an explanatory strategy that has to define a system within which the functional item is embedded (Cummins 1975). Drawing on Carl Hempel, Robert Cummins argues that functions just are the causal roles that a part has in relation to the system as a whole. Cummins’s ideas are currently undergoing a revival in the debate about mechanist explanations (Craver 2007; Bechtel 2008; for a reductivist interpretation, see van Riel 2014). According to the new mechanists, we can ascribe functions to parts or components of systems that realize certain activities, where the relevant components can be described as mechanisms. Therefore, to ascribe a function is to describe why the mechanism does what it does. If we make use of functional descriptions, we have to consider the nested architecture of biological or social systems made of

8  Rebekka Hufendiek et al. mechanisms organized in higher-level mechanisms. Functional analysis often assumes that the ontological status of a function can ultimately be unpacked in causal terms, such that whether a particular causal role for a system as a whole is relevant will inevitably depend on our perspective. It can be more or less adequate with regard to the underlying causal structure, but a relevant description of the mechanisms within a system and their functions ultimately relies on our decisions as to what matters in the respective system (Craver 2013). Such mechanistic explanations, including the analysis of functions as causal-role functions, have been widely accepted in current philosophy of science but hardly discussed with regard to the social sciences. Philosophers of science have discussed the difference between analytic and etiological accounts as a difference in explanatory aim, as well as a difference in metaphysical and normative assumptions. Those who highlight the difference in explanatory aim also highlight that etiological theories and functional analysis are not mutually exclusive but rather asking slightly different questions (Godfrey-Smith 1993; Griffiths 1993; Huneman 2013). Both the etiological and the causal-role theory hold that functions play an important role in explanations, but they differ in how they describe that role. Examples of questions paradigmatic etiological accounts aim to answer are: “Why do we have a heart?” or “Why do we have universities?” The etiological answer to such questions is a causalhistorical reconstruction, which accounts for the existence of these things in terms of the real or intended benefit. By contrast, functional analysis aims to answer how-questions like “how does the heart work?” or “how does the university deliver education?” where the answer is a description of what the heart or the university do within the relevant system (that distinction is widely accepted and goes back to Ernst Mayr’s distinction between questions raised by physiologists and questions raised by evolutionary biologists, see Mayr 1961). The etiological account assumes that the function of X being Y explains the presence of X, whereas the causal-role theorist assumes that the function of X being Y explains or contributes to an explanation of the general proper activity of a system which includes X (McLaughlin 2000; Huneman 2013). With regard to ontological assumptions, the main difference between the etiological account and functional analysis is that the etiological account defends a realist concept of function according to which certain kinds have functions as a result of a history of selection, in which earlier tokens of the kind served a purpose for the (biological or social) system in question. A function, according to this account, is an objective property that is acquired through a process of selection. As a (highly disputed) consequence, an item that accidentally happens to serve a function does not have a function in the relevant ontological sense. This consequence has inspired harsh criticisms of functional explanations in the social sciences: Elster has argued that since we typically lack any idea

Introduction  9 of the selectionist mechanism which is required for a functional explanation in the social domain, most attempts to functionally explain social facts are doomed to fail.4 This issue directly leads us to the normative dimension of functional explanations. This dimension is closely tied to the metaphysical one since the fault line running through the metaphysical dispute seems to follow the question of norm and value. To characterize something as having a function is to view it as a means to an end. But in what sense does such a characterization presuppose a valuation of the relevant end? Is such a value intrinsic or relative to a particular perspective (McLaughlin 2000, 4)? The standard position functional analysts adopt commits them to a mechanistic worldview, suggesting that “the causal structure of the world is disenchanted and purposeless,” and that “functional descriptions … presuppose a vantage point on the causal structure of the world, a stance which intentional creatures take when they single out certain preferred behaviors as worthy of explanation” (Craver 2013, 134). Etiologists object that such a perspectivalist account cannot adequately capture the function-dysfunction distinction. This shortcoming creates practical problems, since … physiologists are interested in explaining pathological processes (such as the growth of tumors), as well as normal ones. A component can have a Cummins function [a causal role function in Cummins’ sense of the term] because of its causal contribution to an organism’s malignancy. Thus, even if we accept that Cummins’ account might allow token malfunction, it will, in some explanatory contexts, turn the function-dysfunction distinction on its head. (Neander 2017, 54) The concept of a mal- or dysfunction has played a crucial role in the social sciences as well. For instance, Durkheim considered some states of a society in terms of anomie, and Merton argued that mal- or dysfunctional effects play a core role in the study of society (see Heiner Koch’s contribution to this volume, which highlights the role of dysfunctions in accounting for social change). Although the etiological and the causal-role theory continue to dominate the debate over functions and functional explanation, alternative accounts have emerged in the last few years. One such account worth mentioning because of its historical pedigree and its more recent currency is the organizational theory of functions. Defenders of this account propose that we can account for both the teleological and normative dimension of functions in terms of the notion of a closed and differentiated self-maintaining organization (Mossio, Saborido and Moreno 2009, 815). Organizational differentiation is a process within which a system generates and maintains distinct structures that each contribute in a

10  Rebekka Hufendiek et al. distinct way to the self-maintenance of the system of which they each are a part. Organizational closure is the mutual dependence among distinct structures that constrain and thereby result: structures that constrain and thereby maintain each maintain each other’s activity. Jointly, they amount to self-determination. Based on these two notions, the organizational account identifies functions with contributions to the selfmaintenance of a system which are due to organizational differentiation and closure. Through organizational closure, a self-determining system contributes to the maintenance of the conditions of its own existence. Mossio and Bich (2017) identify the conditions of the system’s existence of a circular causal regime maintained by the activity of its components with the goal of its organization. In this sense, (biological) organization is supposed to be teleological. In his contribution to this volume, Harold Kincaid offers a simple characterization of functional explanations that does not commit to teleology or a normatively loaded concept of beneficial effects. Instead, Kincaid discusses the conditions under which the social sciences can offer adequate empirical evidence for functional explanations. He argues that it is very common in the social sciences to identify some positive aspect of a social practice and to infer from that that the practice exists in order to provide that aspect. Such inferences, however, are unsubstantiated unless we can find evidence that the practice in question persists because of its positive consequences. Such evidence is often not offered. This is a main reason for the bad reputation of functional explanations, where in principle it is possible to present functional explanations as solid causal explanations. Rebekka Hufendiek, in her contribution to this volume, discusses functional explanations of pride and shame to demonstrate that in the case of functional explanations of cognitive and behavioral traits sufficient evidence for functional explanations (understood as causal explanations) can hardly ever be offered. As a consequence, she suggests to understand functional explanations as eclectic explanations that entail elements of rational and interpretive modeling.

Normativity As we stated in the beginning, when we talk about an object’s function, we usually mean a type of effect of an item, which is beneficial in some way, either for the item itself or for something else. It seems that when we ascribe a function to an entity this sets a standard according to which we can judge the entity in question. A knife has the function to cut, so we can judge whether a knife is good or not by asking whether it cuts well or not. The same is true for social functions. If the function of traditional newspapers is to spread information, we can judge them on the question of whether they fulfill this function appropriately.

Introduction  11 Two questions arise from that: what kind of normativity is this? And: what is its source? A weak answer to the first question is that functional descriptions ground instrumental normativity: when we assume that the function of a knife is to cut, while maintaining that it is not the function of a fork to cut, it seems rational to use a knife instead of a fork if we want to cut something. We may even say that we should use a knife when we want to cut something. A stronger claim would be that cutting is, in some sense, an essential part of what a knife is, and that sets a standard to tell good and bad knives apart. This distinction matters when applied to the social domain: if newspapers are spreading false information, we might want to say that they fail to fulfill their function of disseminating true or valuable information. If distributing information is an instrumental function, what follows from that is that in case that true or valuable information is what we want, we should turn to other sources. If spreading true or valuable information is constitutive of being a newspaper, that sets a normative standard we can draw on to distinguish good from bad newspapers and criticize media politics, without invoking independent moral or other evaluative premises. Most accounts that offer a stronger understanding of the kind of normativity that can be grounded in function descriptions do so by making strong ontological commitments in answering the second question. We have seen in the previous section that claims about normativity being grounded in functions are closely related to metaphysical claims about the objectivity of functional properties. The Neo-Aristotelian tradition draws on a particularly strong conception of how to describe the relation between functions and norms, suggesting that the function of something is part of its nature and that having a function sets a standard that can ground noninstrumental norms. Several Neo-Aristoteleans have defended that we can even refer to functions to ground normative judgments on what is morally good and bad to do (Nussbaum 1992; Foot 2001; Thomson 2008). Alasdair MacIntyre argues that a social practice is a socially established cooperative human activity through which goods internal to this form of activity are realized (MacIntyre 1981, 187). Social practices have an internal normativity that follows from their nature, their history, and their level of expertise. This sets a standard according to which we can speak about good or bad political practices, excellent or lousy guitar playing, or a wonderful teacher in a noninstrumental sense. Social practices do not need to be measured with regard to the question of whether they serve a particular purpose of some agent or whether they are in everybody’s best interest. In being functional they set their own purposes and standards. In a similar vein, Charlotte Witt (2011, this volume) argues that social roles demarcate ways of being a person (being a mother, being a professor, etc.) that normatively structure our agency. Following Aristotle, Witt assumes that, first, a social role is constituted by a function and,

12  Rebekka Hufendiek et al. second, that by being functional social norms are irreducibly normative. Such Aristotelean accounts stand in contrast not only to ascriptivist accounts but also to naturalist accounts that offer etiological explanations of social functions. As we have already seen in the previous section, etiologist explanations explicitly aim for an objective understanding of functional properties that can ground claims on what is functional, malfunctional, or dysfunctional and thereby set a normative standard from which actual facts can deviate. Different meanings of malfunction and dysfunction and the relation between ontological commitments and explanatory fruitfulness are discussed in Heiner Koch’s contribution to this volume. As Koch points out, the most interesting feature of social dysfunctions is their tendency to promote social change. The field in which the normative standard set by etiological explanations is tested most radically for its vindicative or debunking powers is the vibrant literature about the evolution of morality. As Andreas Müller shows in his contribution to this volume, etiological explanations of morality that take morality to be a biological adaptation and assume that the truth of moral judgments can be reduced to the function morality was selected for, hardly seem convincing. Müller discusses the more promising vindicative accounts developed by Philip Kitcher and Philip Pettit asking whether they are more successful in meeting the naturalist challenge and account for true moral judgments in a natural world. Both Kitcher and Pettit assume that morality has the social function of solving certain problems and both take this function to play a central role in the vindication of morality. Some authors concerned with social functions adopt a mechanistic world view, claiming that functions do not exist objectively and independent of human ascriptions. Instead, functions in the social domain originate in collective ascriptions. A paradigmatic case is that of money. We collectively ascribe the function of being a means of exchange to pieces of printed paper within a particular currency union, thereby imposing a normative status (Searle 1995). A status function is a function its bearer can only perform in virtue of the fact that the community in which the function is performed assigns a particular status to the object, person, or entity. By assigning status functions we create institutional facts (concerning money or presidents), determined by constitutive rules, as opposed to regulative rules, and these institutional facts have deontic powers (see Searle 2010). The interesting normative dimension here is that the ascriptive account spells out the rights, authorizations, obligations, permissions, duties, etc. that belong to the social domain and explains them as the result of the ascription and acceptance of status functions. On this view, all genuinely political power is a matter of status functions, and because status functions require collective recognition “all genuine political power comes from bottom up” (Searle 2010, 165).

Introduction  13

Outlook: Some Implications for Social Metaphysics In the emerging fields of social ontology and social metaphysics, social functions have, so far, received little explicit attention compared to other social phenomena such as social institutions, practices, structures or social kinds. However, some philosophers have explicitly invoked social functions to account for the nature, the unity, the persistence, and the existence of social phenomena. As such, social functions are, again increasingly, invoked to do important explanatory work in social metaphysics. In this last section, we will provide an overview of some such uses which appear to be most important for the contributions assembled in this volume. We can find a particularly promising area of application for the notion of a social function in the metaphysics of social kinds. Outside of debates in social metaphysics, some theorists of kinds in both the biological and social domain have already suggested that the members of at least some putative kinds are grouped together primarily because they share the same causal history, rather than the same causal profile (see Dupré 2004; Bird 2008; Garson 2011; Khalidi 2013). Whereas the latter groupings are similar because of laws of nature, the former are so because of a historical copying process. Echoing Millikan’s distinction between “eternal kinds” on the one hand, and “historical kinds” on the other, Khalidi has coined the term “etiological kinds” for them (see cf. Millikan 1984, 1999, 2000, 2005, 307–308; Khalidi 2013). Most philosophers take biological kinds to be paradigms of etiological kinds. However, Khalidi explicitly notes that “some social kinds are etiological” (Khalidi 2013, 130, footnote 4), as well. Likewise, although Millikan has tailored her account to biological kinds, she explicitly extends it to include various artifactual and social kinds (see Millikan 1999, 56–57; cf. Elder 2004). Thus, she has already laid the groundwork for a theory of social kinds in etiological terms, a theory that carries over many of the metaphysical implications of her etiological theory pertaining to both functions and their bearers to the social domain. A prominent recent elaboration of such a theory within the field of social metaphysics is Bach’s account of gender as a “natural kind with a historical essence” (see Bach 2012). Explicitly drawing on Millikan’s as well as other recent accounts of kinds tailored to biological species, he claims that gender kinds are nonarbitrary by virtue of their historical essence, which depends on a distinctive social copying mechanism, namely gender socialization, through which gendered individuals are reproduced (cf. Griffiths 1999; Millikan 1999). Thus, the property in virtue of which an individual is a member of the kind men or women is participation in a lineage: an individual must be a reproduction of ancestral men or women, that is, they must have undergone the ontogenetic processes through which a gender system replicates gender roles. The background assumption is that the dual-gender system had a stabilizing

14  Rebekka Hufendiek et al. effect and that gender roles were culturally selected for this effect (Bach 2012). Because of this, men and women are, on Bach’s account – by their very nature – historical and, therefore, social-functional kinds. Bach’s account is motivated by explanatory a well as political considerations. For one thing, he takes this conception of gender kinds as natural kinds with historical – and functional – essences to provide a better foundation for our inductive practices; for another, he takes it to provide a better foundation for feminist politics. In her contribution to this volume, Mari Mikkola objects to Bach’s account on both explanatory and political grounds. As we have seen already, social functions raise particular ontological questions, where they can’t be sufficiently explained by reference to intentions or intentional actions alone. In her contribution to this volume, Amrei Bahr discusses cases of artifact functions that seem hard to explain by reference to individual intentions. She suggests to bring social practices into the picture to solve the problem. Sally Haslanger, for example, has pointed out that “[s]ocial practices are not fully intentional, as such, but neither are they mere regularities in behavior such as blinking or squinting in bright light” (Haslanger 2018, 237). Bahr suggests that practices turn initially unintended effects into social functions, whereas these functions can then serve to individuate the respective social practices. These recent examples from the emerging fields of social ontology and social metaphysics suggest that there is yet a lot of ground to cover in exploring the explanatory power of social functions in accounting for the nature of social reality, as well as on other, prima facie nonsocial phenomena. With the contributions assembled in this volume, we hope to not only further bolster the case for social functions in philosophy but also shed light on the metaphysical, normative, and explanatory dimensions of social functions that bestow such power on this notion.

Acknowledgment We would like to express our gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

Notes 3 For an overview of the causes facilitating social evolution identified by some of the classic evolutionists, and Spencer, Morgan, and Tylor, in particular, see Sanderson (2016, 21–27). 4 See van Riel’s contribution to this volume; see also Pettit (1996) for an alternative reconstruction of functional explanation in the social domain,

Introduction  15 Kincaid (2007) for the claim that in some cases social sciences can detect selectionist mechanisms through causal modeling, and Little in this volume for a reconstruction of organizations in terms of micro-processes, some of which favor functionalist interpretations while doing justice to the requirement that function ascriptions require particular mechanisms.

References Bach, Theodore. 2012. “Gender Is a Natural Kind with a Historical Essence.” Ethics 122, no. 2: 231–272. Bechtel, William. 2008. Mental Mechanisms: Philosophical Perspectives on Cognitive Neuroscience. New York; London: Routledge. Bird, Alexander. 2008. “Causal Exclusion and Evolved Emergent Properties.” In Revitalizing Causality: Realism about Causality in Philosophy and Social Science, edited by Ruth Groff, 163–178. London; New York: Routledge. Bock, Kenneth. 1980. Human Nature and History: A Response to Sociobiology. New York: Columbia University Press. Bourdieu, Pierre. 1979. La distinction. Critique sociale du jugement. Paris: Les Editions de Minuit. Bowles, Samuel, and Herbert Gintis. 2011. A Cooperative Species: Human Reciprocity and Its Evolution. Princeton, NJ: Princeton University Press. Craver, Carl. 2007. Explaining the Brain. Oxford: Oxford University Press. ———. 2013. “Functions and Mechanisms: A Perspectivalist View.” In Functions: Selection and Mechanisms, edited by Philippe Huneman, 133–158. New York: Springer. Cummins, Robert. 1975. “Functional Analysis.” Journal of Philosophy 72, no. 20: 741–764. Dupré, John. 2004. “Human Kinds and Biological Kinds: Some Similarities and Differences.” Philosophy of Science 71, no. 5: 892–900. Durkheim, Emile. (1902) 1986. De la division du travail social: Etude sur l’organisation des societes superieures. Paris: PUF. Elder, Crawford. 2004. “Artifacts and Other Copied Kinds.” In Real Natures and Familiar Objects, 131–162. Cambridge, MA: MIT Press. Foot, Philippa. 2001. Natural Goodness. Oxford: Oxford University Press. Garson, Justin. 2011. “Selected Effects and Causal Role Functions in the Brain: The Case for an Etiological Approach to Neuroscience.” Biology and Philosophy 26, no. 4: 547–565. Giddens, Anthony. 1981. A Contemporary Critique of Historical Materialism. Berkeley; Los Angeles: University of California Press. Godfrey-Smith, Peter. 1993. “Functions: Consensus without Unity.” Pacific Philosophical Quarterly 74, no. 3: 196–208. Gould, Stephen. 1977. Ever Since Darwin. New York; London: Norton. Griffiths, Paul. 1993. “Functional Analysis and Proper Functions.” British Journal for the Philosophy of Science 44, no. 3: 409–422. ———. 1999. “Squaring the Circle: Natural Kinds with Historical Essences.” In Species: New Interdisciplinary Essays, edited by Robert A. Wilson, 209–228. Cambridge, MA: MIT Press. Harris, Marvin. 1968. The Rise of Anthropological Theory. New York: Crowell.

16  Rebekka Hufendiek et al. Haslanger, Sally. 2018. “What Is a Social Practice?” Royal Institute of Philosophy Supplement 82: 231–247. Henrich, Natalie, and Joseph Patrick Henrich. 2007. Why Humans Cooperate: A Cultural and Evolutionary Explanation. Evolution and Cognition. Oxford; New York: Oxford University Press. Huneman, Philippe, ed. 2013. Functions: Selection and Mechanisms. New York: Springer. Khalidi, Muhammad Ali. 2013. Natural Categories and Human Kinds: Classification in the Natural and Social Sciences. Cambridge: Cambridge University Press. Kincaid, Harold. 2007. “Functional Explanation and Evolutionary Social Science.” In Philosophy of Anthropology and Sociology, 213–247. Amsterdam: Elsevier. Knapp, Peter. 1982. “The Question of Hegelian Influence upon Durkheim’s Sociology.” In The Sociological Galaxy, edited by C. Babbitt, 78–89. Philadelphia, PA: Sociological Society. ———. 1985. “The Question of Hegelian Influence upon Durkheim’s Sociology.” Sociological Inquiry 55, no. 1: 1–15. ———. 1986. “Hegel’s Universal in Marx, Durkheim and Weber: The Role of Hegelian Ideas in the Origin of Sociology.” Sociological Forum 1, no. 4: 586–609. Lopreato, Joseph. 1990. “From Social Evolutionism to Biocultural Evolutionism.” Sociological Forum 5, no. 2: 187–212. MacIntyre, Alasdair. 1981. After Virtue: A Study in Moral Theory. Notre Dame, IN: University of Notre Dame Press. MacLay, George R. 1990. The Social Organism: A Short History of the Idea that a Human Society may be regarded as a Gigantic Living Creature. Great Barrington, MA: North River Press. Malinowski, Bronislaw. 1944. A Scientific Theory of Culture and Other E ssays. Chapel Hill: The University of North Carolina Press. Marx, Karl. (1858/1903) 1993. Grundrisse: Foundations of the Critique of Political Economy, translated by Martin Nicolaus. London: Penguin Classics. Mayr, Ernst. 1961. “Cause and Effect in Biology.” Science 134, no. 3489: 1501–1506. McLaughlin, Peter. 2000. What Functions Explain: Functional Explanation and Self-Reproducing Systems. Cambridge, MA: Cambridge University Press. Meaney, Mark E. 2002. Capital as Organic Unity: The Role of Hegel’s Science of Logic in Marx’s Grundrisse. Dodrecht, Boston, London: Kluver Academic Publishers. Millikan, Ruth. 1984. Language, Thought and Other Biological Categories. New Foundations for Realism, vol. 14. Cambridge, MA: MIT Press. ———. 1989. “In Defense of Proper Functions.” Philosophy of Science 56, no. 2 (June): 288–302. ———. 1999. “Historical Kinds and the ‘Special Sciences.’” Philosophical Studies 95, no. 1–2: 45–65. ———. 2000. On Clear and Confused Ideas: An Essay About Substance Concepts. Cambridge, MA: Cambridge University Press. ———. 2005. Language: A Biological Model. Oxford: Clarendon Press.

Introduction  17 Mossio, M., C. Saborido, and A. Moreno. 2009. “An Organizational Account of Biological Functions.” The British Journal for the Philosophy of Science 60, no. 4: 813–841. Mossio, Matteo, and Leonardo Bich. 2017. “What Makes Biological Organisation Teleological?” Synthese 194, no. 4: 1089–1114. Neander, Karen. 1991a. “The Teleological Notion of ‘Function.’” Australasian Journal of Philosophy 69, no. 4: 454–468. ———. 1991b. “Functions as Selected Effects: The Conceptual Analyst’s Defense.” Philosophy of Science 58, no. 2: 168–184. ———. 2017. “Functional Analysis and the Species Design.” Synthese 194, no. 4: 1147–1168. Nussbaum, Martha. 1992. “Human Functioning and Social Justice: In Defense of Aristotelian Essentialism.” Political Theory 20, no. 2: 202–246. van Parijs, Philippe. 1981. Evolutionary Explanation in the Social Sciences: An Emerging Paradigm. Totowa, NJ: Rowman and Littlefield. Parsons, Talcott. 1951. The Social System. London: Routledge and Kegan Paul. Pettit, Philip. 1996. “Functional Explanation and Virtual Selection.” The British Journal for the Philosophy of Science 47, no. 2: 291–302. Radcliffe-Brown, A. R. 1952. Structure and Function in Primitive Society. Glencoe, IL: The Free Press. Sanderson, Stephen K. 2016. Evolutionism and Its Critics: Deconstructing and Reconstructing an Evolutionary Interpretation of Human Society. New York: Routledge. Searle, John. 1995. The Construction of Social Reality. New York: Free Press. ———. 2010. Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press. Spencer, Herbert. 1893. The Principles of Sociology, vol. 1. London; Edinburgh: Williams and Norgate. Thomson, Judith. 2008. Normativity. Chicago, IL: Open Court. Tilly, Charles. 1998. Durable Inequality. Berkeley; Los Angeles; London: University of California Press. Turner, Stephen. 1986. The Search for a Methodology of Social Science: Durkheim, Weber, and the Nineteenth Century Problem of Cause, Probability, and Action. London, New York: Springer. Turner, Jonathan, and Alexandra Maryanski. 1979. Functionalism. San Francisco: Bejamin-Cummings Publishing Company. van Riel, Raphael. 2014. The Concept of Reduction. Heidelberg: Springer. Witt, Charlotte. 2011. The Metaphysics of Gender. Studies in Feminist Philosophy. Oxford; New York: Oxford University Press. Wright, Larry. 1973. “Functions.” Philosophical Review 82, no. 2: 139–168.

1

The Empirical Issues in Functional Explanations in the Social Sciences Harold Kincaid

Introduction Since at least the 19th century, social scientists have asserted that various social phenomena are explained by their functions and the purposes they serve. Marx, for example, thought the state exists in order to promote the interests of the ruling class. Durkheim claimed that the division of labor in society exists in order to promote social solidarity. Appeals to functional explanations in the social sciences continue unabated up to the present across the social science, as I will document below. Of course, functional explanations in the social sciences (and to some extent in the biological sciences) have been quite puzzling to commentators. These explanations look like attributions of needs and intentionality where there is no real agent who has those needs or intentions. Functional explanations seem like “free-floating teleology,” meaning a teleology that cannot be tied down to real-world causal processes. Attempts to give functional explanations an empirical grounding by appeal to Darwinian mechanisms seem to ignore the vast differences between evolutionary accounts in biology and the way the social world works. In what follows I sketch how functional explanation in the social sciences can be explicated, why some standard objections are wrong, what might count as evidence for functional explanations, and some open and interesting questions about them. The material below builds on previous work (Kincaid 1990, 1996, 2002, 2006a, 2006b) which I think still largely stands as a possible route to understanding functional explanations. Details of most of what I say can be found there.

What Is the Project? What are we trying to do when we analyze functions or functional explanations? I fear much of the debate has been unproductive because it presupposed an ill-founded and an inherently unresolvable question. The question that has generally been pursued concerns what functional explanations are, where the fundamental task is to provide a conceptual analysis. It is assumed that until we have the necessary and sufficient

Empirical Issue in Functional Explanations  19 conditions for “x is the function of y” or “a functional explanation of y is x,” functions and functional explanation, whether in the social sciences or elsewhere, remain suspect. In my view, this approach is fundamentally misguided. The reasons are well known, though still often ignored. Naturalism, paradigmatically found in Quine, doubts that such conceptual analysis can get very far. The deep reason is that there is no a priori, unrevisable knowledge provided by philosophical reflection on meaning and concepts. A weaker version of these doubts would claim that scientific practice is clarified by philosophical analysis only if it is strongly constrained by scientific practice, leaving open the prospect of a priori knowledge by conceptual analysis in other areas. The grounds for these naturalist claims are at least three: (a) the holism of belief makes theoretical claims ultimately subject to empirical evidence; (b) our ordinary concepts are often ambiguous, vague, and not uniform across speakers; and, moreover, (c) it is not clear what we learn about the science from analyzing those concepts. Analyzing scientific concepts – as opposed to everyday concepts – shows more promise. Yet, it is still a dubious goal to try to predict scientists’ linguistic behavior. What scientists say is influenced by all sorts of factors, can be ambiguous and vague, and may not be representative of what scientists actually do. Of course, some attempt to explicate – which is not the same as define – scientific concepts, if tied tightly to scientific practice, certainly has value. This latter project, embodied in an attempt to understand the multiple facets of actual social science research, is what I want to do here and has been my goal in previous work on functional explanation. Thus, my project is to look at functional explanation in the social sciences and to identify claims associated with those explanations that can be given a clear exposition and then to assess their possible empirical applications. In the section below, I identify some causal claims that can be fairly well outlined. I deny that these claims are the only or necessary ways to think about functions or functional explanations in the social science. I do claim, however, that these claims can be clear enough to state and assess, that there are no conceptual reasons to think they must be wrong, and that they fit some standard social science practice.

Clarifying Functional Explanations and Their Evidence Let me start off with a range of typical functional explanations across the social sciences, from past to present: • •

Marx (2008) claimed that the function of the state is to promote the interests of the ruling class. Durkheim (1933, 1961, 1965) claimed that the division of labor serves to promote social solidarity.

20  Harold Kincaid • • • • • • • • •

•

Parsons (1951) claimed that patterns of interaction exist because they promote societal needs. Radcliffe-Brown (1948) emphasized “the sum total of all social relationships of all individuals at a given time” (1948, 55) that exist to serve societal needs. Blau and Duncan (1967) claimed that the occupational structure exists to promote social efficiency. Rappaport (1984) claimed that pig slaughter among the Maring exists because it promotes ecological balance. Hannan and Freeman (1989) claim that organizational strategies exist in order to take advantage of the relevant environments. Tilly (1998) claimed that categorical inequality exists for the purpose of solving organizational problems. Briton and Nee (2002) argue that norms exist to capture the gains from cooperation. Aoki (2001) similarly argues that property rights and community norms serve to regulate external diseconomies. Contemporary historians frequently employ functional explanation: O’Brian (2014) says that the 18th-century British mercantile state existed to promote efficiency, while Ogilvie (2019) denies that medieval guilds existed to promote efficiency as is often claimed, but instead asserts that they existed to protect the interests of certain classes. Contemporary defenders of cultural evolution claim that culture and other social practices change and persist because they promote fitness, either biological, social, or both (Boyd and Richerson 2005).

Even this long list is just a fraction of those that could be cited. There is not much doubt that explaining by functions has and continues to be a mainstay of the social sciences. There are at least two distinct notions of functions or functional explanation invoked in these explanations. The first notion describes the roles social practices play, usually as part of some larger social system. The second notion asserts that social practices exist because of their effects, usually because of their beneficial effects. These are two different claims, with the claim that some social practice “exists in order to” most strongly raising the worries about free-floating teleology. Both types of functional explanation are found in the samples quoted above. There are numerous ambiguities in the research cited above. The most basic unclarity is probably blurring the two different types of explanation just noted – social scientists often go seamlessly from identifying effects of social practices as “functions” and then on to the claim that they exist because of these effects. Other unclarities concern the exact nature of the benefits produced by a practice, who or what receives the benefits, and what would count as evidence for the latter.

Empirical Issue in Functional Explanations  21 The idea that social phenomena can be explained by social roles has not attracted as much critical attention as have functional explanations of the teleological sort, perhaps because social roles are less controversial and because the basic idea is fairly simple. In describing the organization of a society, social scientists find it helpful to show the characteristic behavior – the causes and effects – of specific practices which they standardly called roles. Initially, these categories were taken from commonsense sociology, as it were. Family roles are a paradigm example, and anthropologists showed, going beyond (Western) common sense, that there is a diversity of such roles across societies and cultures. “Characteristic behavior” is generally fleshed out in practice by describing the social expectations associated with the role and the causal influences on roles and of roles on other social aspects. Appeal to functional roles in the social sciences often adds to the role claims that there is a set of interrelated roles and that society is organized, or more strongly, made stable by that set of roles. Sometimes even more is claimed – social roles meet “social needs.” The idea of roles as typical causes and effects of some entity or practice and that roles constitute an interlocked system is not unique to the social sciences. It is advocated in philosophy of mind and some parts of cognitive science as an account of mental functioning (Cummins 1975). Such explanations have their criticisms, but they are largely not criticisms of the very idea of a function in this sense. Something having a characteristic set of properties and causal relations seems uncontroversial itself, but then debate does begin when the idea is tied down with specifics and made a general theory of society, for example, where it is an empirical issue how such explanations work and how far they go. The second sense of functional explanation goes beyond describing causal roles of social practices to assert that practices exist because of or in order to bring about their effects. This claim is the one that most seems to threaten free-floating teleology. Nonetheless, as I have argued elsewhere (Kincaid 1990, 1996, 2002, 2006a), these explanations by effects can be translated into ordinary causal relations, albeit relations of a specific type. One such translation involves at least two causal claims: 1 F causes B. 2 F persists because it causes B. The first claim is an ordinary causal claim of the sort social scientists make all the time. We may wonder how well social research supports causal claims like this, but as it stands, (1) is no different than other causal claims in the social sciences, and (2) does much more work – it translates the “in order to” in functional explanations. The idea is that F not only has a function or standard causal effect, but in some sense, it

22  Harold Kincaid exists because it does so. Again, this second component is also a causal notion. It says that when F causes B at one time, then that fact explains F’s existence at another time. In simple statistical terms, (2) can relate two binary variables: “F caused or did not cause B at t1” with another binary variable “F exists or does not exist at t2.” This suggests, as at least an initial piece of evidence, using panel data for a logit estimation of E = a + xC + e where E = exists a t2, C = F had causal effect at t1, a is a constant, and x is a regression coefficient measuring (tentatively and crudely) the effect of C on E. We could stop with these two conditions and would have found a way to make some sense out of teleological sounding claims. So, for example, if we believe that norms exist in order to promote cooperation, we need evidence that norms causally contribute to cooperation and that their existence at time t + 1 is caused by their contribution at t1. Thus, I claim we can turn functional explanations into nonmysterious causal relations by the translation or model represented by (1) and (2). I do not believe that this analysis is the only or “right” way to understand commonsensical or informal functional language. However, efforts like the above to turn vague concepts and relations into explicitly testable and tested causal claims seem to me important for the social sciences to advance. I think this search for clarity is an important legacy from positivism even if their axiomatic and operationalist ambitions were misplaced. In previous work, I have suggested a third element in functional explanations. A little thought shows that (1) and (2) by themselves are also compatible with explanations that do not seem to have any commonsense ‘in order to’ reading or teleological implications. Two physical bodies in a stable mutual interaction will fit (1) and (2). The orbit of Saturn is caused in part by Jupiter’s orbit, and Jupiter’s orbit persists into the future in part because it does so. Mutual interaction or causation may be one way to spell out functional explanations in terms of roles (as does Faia (1986) explicitly), but these accounts are not specific enough to fully get at the teleological element. The problem with mutual or reciprocal causation is that there is no reason to pick either factor out as existing “in order to.” Appeals to social practices existing in order to bring about their consequences do not imply in addition that their consequences exist because of their function. For example, if norms exist to promote cooperation, that does not entail that cooperation exists in order to promote norms. Functional explanations thus seem to have a kind of asymmetry. When F persists because it causes B, but we have no evidence to think that B persists because it causes F, then we have the clearest case of a functional explanation in sense (2).

Empirical Issue in Functional Explanations  23 The account given so far has only described abstract causal relations. It has said nothing in itself about what sort of entities Fs are nor what kind of things Bs are (see van Riel’s contribution to this volume). Furthermore, my account has made no appeal to selectionist processes or mechanisms – (1) and (2) could hold in a universe with no other entities than Fs and Bs and thus no differential survival. This generality helps defuse a number of confusions and objections in response to functional explanation in the social sciences, as we will see below. As causal claims, functional explanations of the sorts I have described above are supported by whatever kind of evidence social sciences provide for causal claims. Full-fledged statistical causal models, comparative case analysis, process tracing, ethnographic field work, and so on are potentially relevant. Then, the interesting question is to what extent we can have functional explanations that are well supported by such evidence in social science research. Thus, blanket judgments about how well functional explanations are confirmed are probably misguided. Instead, detailed case studies are needed. That is beyond my purview here, though I have made efforts in that direction elsewhere (Kincaid 1996).

Objections Diffused, Open Questions, and Complications In this section, I use the account above to answer some standard objections to functional explanations, to further clarify the abstract schema above, and to note some open questions about functional explanation in the social sciences. First, we need to note that nothing said so far determines how widely functional explanations apply in the social sciences. There may be many or few social phenomena where compelling evidence for (1) and (2) can be established. It is also worth noting that functional explanation as defended here does not entail or require an evolutionary story about societies. Evolution on its own is just change in state over time. There is plenty of grand theorizing about societal evolution, stages, etc. in the social sciences, but functional explanations do not presuppose them. Nor does functional explanation need to be a full account of society, nor is it necessarily committed to conservative claims that all major institutions exist for some beneficial reason, a view sometimes advocated by functionalism in the 1950s and 1960s. Another point about the scope of functional explanation concerns causal complexity in the social sciences. Because (1) and (2) are causal claims, they are compatible with a more complex picture than what I have presented. The two causal claims could be embedded in more complex causal structures where the effects in both cases, for example, also had other nonfunctional causes. This is a common situation and not unique to functional explanation, but it does raise the bar for evidence

24  Harold Kincaid and explanation. For example, when beneficial effects are cited that might have causes in addition to or instead of the alleged functional causes, then these complexities need to be addressed. A related causal complexity surfaces in what are now standard cultural evolutionary models (e.g. Boyd and Richerson 1985). These models are inspired by population genetics. They treat society as a population of individuals and then describe and explain changes in the cultural characteristics of individuals. These characteristics are usually beliefs or other states of individuals. Thus, culture is treated as the set of characteristics in the head of individuals, as it were. These models ignore important causal complexity and reflect an individualistic bias. There are serious, albeit abstract, reasons from philosophy of mind and cognitive science to think that “what’s in the head” has determinate states only as part of a surrounding social environment. In general, accounts of individual behavior in the social sciences often work by assuming an institutional background. For example, game theory explains institutions as equilibria in games. Those explanations describe preferences, strategies of individuals, individual endowments, and so on. Yet those explanations take as given those preferences, strategies, initial endowments, payoffs, amounts of interaction, etc. – in short, they presuppose as unexplained the rules of the game, institutions, social structure (e.g. who owns what), and so on. This means that even if these functional explanations succeed, they leave much unexplained. Moreover, there is no reason that functional explanations have to be entirely about individuals. As we saw above, the general requirements for functional explanation are silent on what they apply to. There is no a priori reason we cannot explain the traits of social entities, e.g. of organizations and corporations, as existing in order to cause social effects such as survival or copying by other social entities. Social group selection is one such mechanism. Unfortunately, much of the cultural evolution literature is ambiguous on these questions, sometimes talking about differential survival of individual traits because of group membership and at other times about differential survival of groups themselves because of group traits. These two processes are different and lead to different functional accounts. Even current philosophy of biology and evolutionary theory sometimes are unclear about which process is involved. Another complication with functional explanations concerns mechanisms. I think this is a complication that is often exaggerated. The general literature on mechanisms is quite ambiguous in my view. I have discussed these issues elsewhere (Kincaid 1996, 1997, 2012), but here is a synopsis. Saying that we need mechanisms is ambiguous on multiple fronts. We might want mechanisms to help confirm or want them to provide better explanations – these are different things. We might want mechanisms in the sense of the underlying detail of causal factors – how they are realized in lower-level compositional structure – that explains

Empirical Issue in Functional Explanations  25 aggregate causal properties or we might want to know intervening variables. These, too, are different things. I might want mechanisms because I want a deeper explanation even if I already know a causal claim is empirically well confirmed. We had good evidence for genes long before we knew their molecular structure – before we knew their constitutive mechanisms as it is sometimes called in the literature. However, asking how sequences of DNA produce phenotypic traits, most directly proteins, is asking for the intervening process between genes and other things. Providing such mechanisms is describing not the constituents of genes but how genes fit into a pathway to organismal traits. Whether we need or gain from having mechanisms along any of these parameters depends on the context, e.g. what else we know, what we are trying to do, and so on. Causal claims are confirmed and taken as explanatory all the time without a full account of intervening variables or of underlying details – think of baseballs breaking windows. Nothing about functional accounts is inherently different, for functional explanations are just a certain sort of causal explanation. Of course, adding more detail increases epistemic confidence and explanatory depth. Yet, we do not have to have everything to have something. To assess our epistemic confidence, for example, we need to assess how much we know about the causal relations involved, what the causal relations logically assume about underlying and/or intervening mechanisms, and what we know about them. Depending on the context, we may need to know much or little about mechanisms. This point about mechanisms is one about principle. That principle can be right while specific functional explanations can be quite suspect because they are incomplete and in that sense lack mechanisms. It is very common in the social sciences to identify some positive aspect of a social practice and to infer from that that the practice exists in order to provide that aspect. However, such inferences are unsubstantiated unless we can find evidence for (2) – evidence that the practice persists because of its positive consequences. Frequently, the evidence needed to establish the crucial step (2) is never given. Much of the bad reputation of functional explanations comes from this failing.

Open Issues and Conclusion There is a great deal in the above that needs further elaboration and investigation. Game theory is a case in point. Game theory explanations claim that social phenomena exist because the behaviors involved are the best responses to each other. So, strategies exist in a population because they are optimal – in order to get the best individual outcome given what everybody else does. These accounts support functional explanations, but how well confirmed they are and how much they explain in the social sciences still needs further investigation in my view.

26  Harold Kincaid Similarly, claims that game theory can unify the social sciences are another research question (Gintis 2009). It is hard to see how they provide more than a common formalism, though that is certainly one way that the natural sciences are sometimes thought to achieve unification. There are general philosophy of science questions involved here, and it would be interesting and useful to have a look at them in this context. Another topic needing attention concerns the level at which we give functional explanations. Functional explanations are as often used to talk about the properties of social entities – groups, practices, etc. – as they are about individual behavior. Do traits at the collective or macro level used in functional explanations require some commitment to group goals, purposes, etc.? Can those uses be defended? I think they can in some social science explanations (see Kincaid 2019), but that is controversial and entirely undiscussed, as far as I know, in the case of functional explanations. Likewise, almost entirely unexplored are explanations invoking multiple levels. However, hopefully the considerations above show two things. First, it seems clear that functional explanations are a key part of the social sciences. Second, there are ways to make functional explanations nonmysterious and amenable to empirical investigation.

Acknowledgments Thanks to Rebekka Hufendiek and Raphael van Riel for helpful comments on an earlier draft.

References Aoki, Masahiko. 2001. Toward a Comparative Institutional Analysis. Cambridge, MA: MIT Press. Blau, Peter, and Duncan, Otis. 1967. The American Occupational Structure. New York: The Free Press. Boyd, Robert, and Richerson, Peter. 2005. The Origin and Evolution of Cultures. Oxford: Oxford University Press. ———. 1985. Culture and the Evolutionary Process. Chicago: University of Chicago Press. Brinton, Mary, and Nee, Victor, (eds.). 2002. The New Institutionalism in Sociology. Palo Alto, CA: Stanford University Press. Cummins, Robert. 1975. “Functional Analysis.” The Journal of Philosophy 72, no. 20: 741–765. Durkheim, Eemile. 1933. The Division of Labor in Society. New York: Macmillan. ———. 1961. The Elementary Forms of Religious Life. New York: Collier. ———. 1965. Rules of the Sociological Method. New York: Free Press. Faia, Michael. 1986. Dynamic Functionalism: Strategy and Tactics. Cambridge: Cambridge University Press.

Empirical Issue in Functional Explanations  27 Gintis, Herbert, 2009. The Bounds of Reason: Game Theory and the Unification of the Behavioral Sciences. Princeton, NJ: Princeton University Press. Hannan, Michael T., and Freeman, John. 1989. Organizational Ecology. Cambridge, MA; London: Harvard University Press. Kincaid, Harold. 1990. “Assessing Functional Explanation in the Social Sciences.” PSA 1990, vol. 1: 341–354. ———. 1995. “Optimality Arguments and the Theory of the Firm.” In The Reliability of Economic Models: Essays in the Epistemology of Economics, edited by Daniel Little, 211–236. Boston, MA: Kluwer. ———. 1996. Philosophical Foundations of the Social Sciences: Analyzing Controversies in Social Research. Cambridge: Cambridge University Press. ———. 1997. Individualism and the Unity of Science: Essays on Reduction, Explanation, and the Special Sciences. Lanham, MD: Rowman and Littlefield. ———. 2002. “Functionalist Successes and Excesses in the Social Sciences.” Analyse und Kritik 24, no. 1: 60–71. ———. 2006a. “Functional Explanation and Evolutionary Social Science.” In Handbook for the Philosophy of Science: Philosophy of Anthropology and Sociology, edited by Risjord and Turner, 213–249. Amsterdam: Elsevier. ———. 2006b. “Evolutionary Social Science Beyond Culture.” Behavioral and Brain Sciences 29, no. 4: 356. ———. 2012. “Mechanisms, Causal Modeling and the Limitations of Multiple Regression.” The Oxford Handbook of the Philosophy of Social Science, edited by H. Kincaid, 46–65. New York: Oxford University Press. ———. 2019. “Beyond Causalism and Acausalism.” In Explanation in Action Theory and Historiography, edited by Gunnar Schumann, 179–194. New York: Routledge. Marx, Karl and Engels, Friedrich. 2008. The Communist Manifesto. London: Penguin. O’Brian, P. 2014. “The Formation of States and Transitions to Modern Economies.” In The Cambridge History of Capitalism, Vol. 1, The Rise of Capitalism, edited by L. Neal and J. Williamson, 357–403. Cambridge: Cambridge University Press. Ogilvie, Sheilagh. 2019. The European Guilds: An Economic Analysis. Princeton, NJ: Princeton University Press. Parsons, Talcott. 1951. The Social System. New York: The Free Press. Radcliffe-Brown, Alfred R. 1948. A Natural Science of Society. Glencoe, IL: The Free Press. Rappaport, Roy. 1984. Pigs for the Ancestors. New Haven, CT: Yale University Press. Tilly, Charles, 1998. Durable Inequality. Berkeley: University of California Press.

2

Do Organizations Adapt?1 Daniel Little

The Setting: Social Functionalism Social functionalism maintains that social features – institutions, values, laws – have the characteristics they possess because they serve important needs of the social order. Or to put the point the other way around, a social order comes to have features that are functionally well adapted to serving the needs of society (or powerful groups within society). Social anthropologists Radcliffe-Brown, Evans-Pritchard, and Malinowski were early advocates of such a view, and Marx makes such a statement when he asserts that the superstructure of capitalist society exists to stabilize the property relations of the economic structure. Robert Merton offered a much more qualified and careful theory of social functions as a limited and empirically testable notion. In Social Theory and Social Structure (1963), he finds that earlier versions of sociological functionalism were deeply flawed. Here is a fairly clear statement of the logic of functional analysis as practiced by social anthropologists like Evans-Pritchard and Malinowski, according to Merton: Substantially, these postulates hold first, that standardized social activities or cultural items are functional for the entire social or cultural system; second, that all such social and cultural items fulfill sociological functions; and third, that these items are consequently indispensable. Although these three articles of faith are ordinarily seen only in one another’s company, they had best be examined separately, since each gives rise to its own distinctive difficulties. (Merton 1963, 79) Merton offers careful critique of all three postulates. He disputes that we have a basis for postulating the full integration of societies; he doubts that there is evidence to support the idea that all persistent social characteristics are “functional”; and he notes that there are often alternative arrangements that would work as well to bring about the assumed benefits, so the selected item is not “indispensable” (90).

Do Organizations Adapt?  29 Merton’s own theory of social functionalism draws an important distinction between manifest and latent functions – between the directly observable beneficial consequences of a given social arrangement and the indirect and unobservable consequences that the arrangement may have. Further, he believes that there is good sociological reason to believe that latent functions exist. Philosophers of social science and biology have expressed deep skepticism about sociological functionalism. Particularly cogent were the compelling arguments offered by Jon Elster (1983) in his discussions of Gerald Cohen’s Karl Marx’s Theory of History (1978). Fundamentally the critique of sociological functionalism depends on discrediting an analogy between biology and sociology. The language of functional adaptation of organs to system finds its origin in biology; the heart serves to sustain the organism by pumping blood. But, as Elster and others point out, there is a crucial disanalogy between biology and sociology. This is the fact of natural selection and evolution in the biological case, which has no close analogue in the social case. We have a very good understanding of the mechanism that leads the physiology of a species to be well adapted to serving the reproductive and survival needs of the organism; this is the mechanism of natural selection. But social entities are not subject to analogous selection pressures. So the functionalist view of society is suspect because it depends on a fundamental mystery: how can the effects of a social feature be thought to explain its emergence and structure? How can the fact that a particular system of norms about marriage has the effect of creating a docile workforce explain the existence of that set of norms in society? Doesn’t this get the temporal order of causation backwards? It is not impossible to provide answers to this question, but we cannot take it as a dogmatic assumption that society will unavoidably be organized in such a way that its needs are satisfied by well-tuned institutions. There is one answer to this question that makes perfect sense: the mechanism of purposive design. There is no mystery in the fact that a rudder exists to steer the ship; it was designed by the ship builder to do exactly that. And perhaps it is no less straightforward to assert that the audit function in a corporation exists to prevent dishonest behavior within the organization. Intelligent managers, foreseeing the temptation to fiddle accounts, designed a system for detecting, deterring, and punishing fiddling. What is more difficult to interpret is a claim about the social function of a feature where the postulated consequences of the workings of the features are unknown to participants (latent, in Merton’s terms) and where the feature took shape as the result of the independent actions of numerous actors over time. The first point makes intelligent design difficult to carry out, and the second makes it implausible to assume the existence of the intelligent designer at all.

30  Daniel Little However, it is not impossible to imagine social mechanisms that might bring about a functional adaptation of a social arrangement to the needs of the system. Here is a mechanism that complements the intelligent design mechanism. Call it the mechanism of “slow accumulation of selfinterested adjustments.” Suppose that a social arrangement like a set of norms regulating a common kind of human situation is present in a population at a given time. And suppose that the details of the arrangement are open to modification over time by actions of various actors in the situation. At any given time there will be groups and individuals whose interests are affected by the arrangement. When modifications are possible, groups can seek to implement the modifications of the arrangement that favor their interests or beliefs. And if there are groups within society that wield disproportionate power and influence, we can expect that the modifications that are selected will more often than not be ones favored by this group. This mechanism implies that the social arrangements will be “tuned” over time to work in such a way as to serve the interests of the influential group. This mechanism is different from the intelligent design mechanism because it assumes only local modifications over time – not a unitary optimal planning and design process. It looks more like an evolutionary process – myopic but directional. But the mechanism of innovation and selection is different from the biological case. Innovations survive because they are preferred by influential stakeholders. And the mediumand long-term result is a drift in the direction of fit between the working characteristics of the social arrangement in question and the needs and interests of the influential groups. This gives us a different and more sociologically credible way of defending the idea that important social arrangements will be “functionally adapted” to satisfying the needs of the powerful groups in society – not because those groups directly designed the arrangements, but because they had influence on the process of refinement and adjustment that led to the present configuration.

The Problem: Adaptive Organizations? Let us now consider a specific kind of social entity, a formal organization through which various social activities are managed and directed. Corporations, police departments, labor unions, community activist clubs, governmental safety agencies, and universities provide numerous familiar examples of organizations. Organizations operate in ways designed by founders and adjusted by current actors to bring about certain kinds of effects, they depend upon the coordinated efforts of numerous individuals, and they exist in environments that both affect and react to their ongoing activity and performance. Moreover, organizations are to some extent plastic: the systems, rules, and practices that make them

Do Organizations Adapt?  31 up can change over time. Sometimes these changes happen as the result of deliberate design choices by individuals inside or outside the organization; for example, a manager may alter the rules through which decisions are made about hiring new staff in order to improve the quality of work. And sometimes they happen through gradual adjustments over time of which no one is specifically aware. These facts about organizations imply that the mechanism of slow accumulation of self-interested adjustments is likely to occur in the processes of change in organizations over time. The question arises, then, whether organizations evolve towards higher functioning based on signals from the environments in which they live; or on the contrary, whether organizational change is stochastic, without a gradient of change towards more effective functioning. Do changes within an organization add up over time to improved functioning? Are formal organizations such as business firms, labor unions, or nonprofit community organizations subject to identifiable social forces that lead them to improve their functioning over time? Is there any reason to expect that organizations adapt or evolve into higher levels of performance? In order to expect adaptation and functional improvement we need to be able to identify feedback mechanisms that lead from features of performance to alterations in the workings of the organization that will improve performance. Are there such mechanisms in the case of formal organizations? What kinds of social mechanisms exist that might bring about such an outcome?2 We can abstractly characterize an organization in terms of four features: mission, arrangements, performance, and change. An organization possesses a charter of activity – the outcomes it is designed to bring about (“mission”). The internal workings of organizations (“arrangements”) determine the way in which the organization operates and therefore its ability to bring about intended goals. Often the achievement of goals can be assessed by participants and outsiders (“performance”). The internal workings of organizations change over time, often leading to effects on the effectiveness with which they bring about their purposes (“change”). It is clear that intentions and purposes come into organizations from the start, at various levels. Organizations are intended by a variety of actors to bring about certain goals, and participants exercise intentionality as they carry out their activities. And, as we will see below, the arrangements of an organization are subject to the intentional activities of actors outside the organization as well. The question here is thus a clear one: is there any reason to expect that the arrangements of organizations will tend to change in a direction that leads to improved performance relative to its mission? Are there forces, mechanisms, or contextual factors that lead to a tendency for organizations to improve their performance over time through adjustment of internal arrangements and activities? Is organizational change directional,

32  Daniel Little or is it a more or less random process, leading to worsened performance about as often as it does to improved performance? There are important analogies between organizations and technical artifacts. One reason that we are comfortable using functionalist language in referring to artifacts like tools, aircraft, or skyscrapers is that they are human-made and designed with commonly understood purposes, and human beings are often able to observe their performance and improve their efficiency and efficacy. Innovations are introduced into materials, designs, and technical practices that improve the functioning of artifacts. Experts – engineers, architects, chemists, physicists  – use their knowledge and skill to improve the design and functioning of the artifact. Users themselves make innovations, and they imitate the innovations of others. Further, the process of review, evaluation, and redesign is generally a carefully managed one in which the efforts of multiple experts are marshalled in pursuit of a single goal. Consequently, there is good reason to expect that the properties of an artifact will be adjusted over time so as to achieve the best feasible performance relative to the desired uses of the artifact in the current environment of human needs and physical materials. Jon Elster (1983) offers a very thorough treatment of the social processes through which technical change occurs in artifacts and practices. This process is the result of temporally extended and socially distributed intentionality. Organizations have some important features in common with artifacts. They are often the result of deliberate design by founders with careful attention to the overall process and performance of the organization, and they are also often subject to ongoing review to evaluate how the components are working relative to the goals of the organization or its stakeholders. Founders, executives, and managers are intelligent manipulators of organizations who have the ability to introduce performance-enhancing innovations. Therefore, we might expect that organizations will show a tendency to improve their functioning over time, similar to technical artifacts and practices. There are several complications in this story, however, that make the analogy with artifacts misleading. First, the various stakeholders and actors within an organization often have conflicting goals which lead them to favor different kinds of innovations. Second, there is a high degree of decentralization of authority in many organizations, with the result that it is difficult or impossible for the executive or the governing board to command efforts towards a particular kind of improvement of function. Third, organizations are subject to capture and mission drift, where the founders’ purposes are put aside and new influencers focus on a new set of goals. Features that were “functional” in the former statement of mission are now subordinate or dysfunctional in the new statement of mission. And fourth, there is a wide range of viability for organizations, which means that they do not need to be carefully tuned

Do Organizations Adapt?  33 to the environment to survive, so viability is rarely the decisive factor. So if there is a mechanism leading to organizational evolution and adaptation, it must lie elsewhere, internal to the workings of organizations and the forms of agency that exist there. Therefore, we need to look more closely at how organizations work and the processes that lead to organizational change.

Organizations and Change Discussion of the functionality and performance of an organization unavoidably requires that we address the question, “with respect to what set of goals and purposes?”. Organizations are loosely paired with more or less explicit missions, the goals the organization is expected by various influential actors to accomplish. The goal, purpose, or mission of the organization is stated in simple terms, and the activities, rules of operation, and staff of the organization are coordinated so as to achieve the mission effectively. We can further differentiate the idea of mission by distinguishing among the conceptions of the purpose and goals of the organization held by (a) the founders, (b) external stakeholders (investors, regulators, legislators, the press), (c) internal stakeholders (executives, managers, workers), and (d) the public. Each of these configurations of individuals and groups have powers and opportunities enabling them to influence the behavior and structure of the organization, depending on the degree to which they are satisfied with the level of attainment of mission it provides according to their definition. It would be reasonable to expect that there is a degree of overlap among these different conceptions of mission – presumably all groups agree on the most general description of the purpose of the organization. However, it is apparent that there may be significant differences across the different groups. This makes it clear that there is no single overriding answer to the question, what is the mission of the organization? It also means that a feature may be functional with respect to one group’s conception of mission, and dysfunctional with respect to another group’s conception. Now consider the arrangements that constitute the workings of the organization. An organization is differentiated in its internal activities and the behaviors of the individuals who make it up (Crozier and Friedberg 1980; Lawrence and Suddaby 2006). Abstractly, an organization is an ensemble of specific elements: actors and motives, rules of action for the actors, authority relations among the actors, activities carried out by individuals and groups, and inputs and outputs. An organization is different from a crowd or a street protest. An organization consists of rules of procedure, departments, supervisors, executives, agents, resources, and information stores. Organizations are hierarchical systems. Organizations have modes of functioning; they have processes for making decisions, moving materials, storing and retrieving information, putting staff to

34  Daniel Little work, supervising effort and malfeasance, and so forth. And these functionings are the result of the internal configuration of the organization; so the organization can be modified by changing the internal configuration. There are formal decision-making processes, lines of authority, rules and procedures, and systems for oversight and supervision. The arrangements of the organization serve to break down the activities of the organization into a number of sub-tasks and functions. They provide for decision-making, planning, logistics, marketing, information-gathering and storage, and oversight of the activities of various groups and individuals within the organization. (Charles Perrow (2014) provides an overview of the development of organizational sociology since World War II.) The normal workings of an organization are the result of the arrangements of the organization, the states of agency of the incumbents, and the forms of oversight and feedback provided by external stakeholders. If the arrangements of the organization are reasonably well designed to bring about the goals of the organization, and if the incumbents are provided incentives and motivations to carry out their roles as specified by the arrangements, then the organization is likely to function reasonably well relative to its mission and sub-goals. If either assumption fails, then a range of dysfunctions are likely to occur over time, and dysfunctions are likely to bring some response from stakeholders. An electric power company that consistently fails to restore power to consumers after storm-induced outages is likely to come under strong pressure to change its arrangements, from the public, from the media, and from government regulators. There are identifiable feedback mechanisms that can be expected to arise in response to organizational dysfunction. (See Little (2015) for an effort to understand organizational change within universities from the perspective taken here.) The workings of an organization affect human interests and concerns. Organizations do things that we care about. Sometimes we may focus on the public interest served by a well-functioning organization; sometimes it is the interests of the “clients” of the organization that we are thinking about; sometimes it is the other persons or groups who are affected, negatively or positively, by the activities of the organization; and sometimes it is internal stakeholders whose interests are particularly salient. And, finally, organizations are mostly influenced and directed by intelligent actors who are interested in maintaining or modifying the configuration and functioning of the organization in light of their perceptions of how well or badly things are going, from their individual points of view. Intelligent refinement of processes at a range of levels within an organization is something we can expect to occur, when incumbents have an internal or external incentive to improve performance. So a specific answer to the question posed above of the adaptability of organizations appears to be this: there is a positive feedback mechanism created from good and bad performance to maintenance or refinement

Do Organizations Adapt?  35 of organizational arrangements, flowing from constituents to managers. However, only a subset of these effects play into feedback loops that induce improvement of functioning. The arrangements of the organization have most often been designed and implemented by actors at various levels who have a conception of how the arrangement will work and how it fits into the overall performance of the organization in fulfillment of components of its mission. Implementation and design take place over an extended time, so gradual incremental changes are as important as moments of grand architectural design. This means that there is a degree of rational planning that goes into the creation, maintenance, and modification of the arrangements of the organization. However, there is also a degree of conflict and discord across a typical organization about the modes of functioning of the organization. Various actors have a range of private interests in addition to their organizationally defined roles. And there may be disagreements among actors about how best to accomplish various tasks. So it is important to avoid the conception of a unified rational will that exists within the organization; rather, there are a variety of actors in a variety of positions of power and decision-making authority, leading to organizational actions and decisions that may seem inconsistent. It is also an important fact that organizations have informal processes that are sometimes at odds with the formal processes described here. Individual actors within an organization not only have rule-defined roles, but they also have opportunities for action that are contrary to the rules but difficult to observe or discipline (Crozier and Friedberg 1980; Perrow 1999; Mueller 2015). Individuals sometimes disregard directives, misappropriate resources, or pursue their own conceptions of how the organization should behave. Organizations are frequently “loosely coupled”, with the result that individuals are able to behave at some variance from what the authority hierarchy and the rules of the organization prescribe (Weick 1976; Orton and Weick 1990; Beekun and Glick 2001). Another key fact is that organizations operate within a broader environment of external actors (Aldrich and Pfeffer 1976). These include consumers, stockholders, investors, regulators, media commentators, and numerous other individuals and organizations who are affected by the functioning of the organization in question. The performance of the organization leads various external actors to take actions affecting the executives of the organization in various ways, giving them incentives to consider changes.

Intentions and Purposes A recurrent theme here is that organizations are importantly driven by intentions and purposes of a variety of actors. In particular, organizations are subject to many different kinds of intentions, at many levels of

36  Daniel Little actors and actions. There is the public purpose of an organization – the commonly recognized set of goals the organization is believed to exist in order to further. (For example, “maintain the safety of food and drugs,” “generate profits for the shareholders,” “maintain law and order subject to constitutional protections of citizens,” and “educate the young.”) Often an organization was explicitly created in order to further specific goals (by a legislature or elected executive, by a group of investors, by a group of social activists). We refer to this as the mission and purpose of the organization. But it is also necessary to consider the private purposes that influential participants inside and around an organization have. Executives may have private financial interests that are independent from the public purpose of the organization (e.g., preparing for a lucrative exit following acquisition by another company). Staff of an organization may have private interests in maintaining a comfortable work environment that are likewise independent from the public purpose (e.g., resisting departmental reorganization aimed at establishing a more efficient workflow, resisting investigations by Internal Affairs into complaints of police misconduct (Skogan 2008)). And outsiders may have private interests in preventing effective performance of the organization’s activities (e.g., owners of chemical or nuclear plants who have an interest in minimizing regulatory inspection visits to their facilities, see Perrow (2011)). The actual configuration of the organization and its decisions and actions are influenced by these private interests by various powerful actors. Robert Bates draws attention to these kinds of conflicting incentives within organizations in his treatment of economic institutions in African development (Bates 1981). In their account of systems of government regulation in advanced democracies, Baldwin et al. (2012) draw a similar distinction between the “public interest” thought to be served by a system of regulation and the private interests, both within and outside of government, that often drive the design and enforcement of a system of regulation. These dysfunctions derive from the perennial possibility of conflicts of interest and purpose between executives and employees (for example, between the central tax authorities and the regional tax inspectors; Klitgaard 1988). These conflicts are described as “principal-agent problems” (Stiglitz 1989; Miller 2005). A chemical company executive directs that an inspection of the XYZ chemical plant will occur in the month of September, lower-level officials tasked to implement this directive “slowwalk” the implementation, and inspectors accept valuable favors from the local plant officials in exchange for a cursory look at maintenance records. The result is that the inspection occurs in the month of January, and it fails to identify a chronic pattern of inadequate maintenance. This troubling scenario illustrates a principal-agent problem at multiple levels. Anyone who has worked in an organization knows that there are multiple strategies available to executives to minimize the discrepancy between directive and performance. Spot inspections by independent

Do Organizations Adapt?  37 auditors can reveal faulty performance, frequent communication and visits by executives to the field can help, careful personnel processes aimed at recruiting employees of high reliability helps, and so forth. But the point remains that there is predictably a degree of slippage between higher-level decision-making and on-the-ground implementation of the organization’s directives. Organizational theorists refer to this characteristic of organizations as “loose coupling” (Weick 1976; Orton and Weick 1990; Beekun and Glick 2001). So a nuanced understanding of the role of intentions and interests within organizations tends to undermine the expectation of straightforward functionality within the arrangements of an institution or organization: organizations are subject to multiple and often conflicting purposes, both public and private, both internal and external. The topic of conflict of interest and commitment is an ineliminable feature of organizational behavior (Rodwin 2011; Abend 2014). These conflicting purposes and goals are pursued by actors who have an ability to influence the activities and arrangements of the organization. We certainly cannot assume as an axiom that an organization is well designed to carry out its publicly stated mission because the internal arrangements of the organization are themselves subject to conflicting purposes, intentions, and interests of actors whose interests may be at odds with the public goals of the organization. This account of the internal workings of typical organizations emphasizes the fluidity and multiple agency that is embodied in virtually all organizations, from the Defense Department to the local religious institution. Neil Fligstein and Doug McAdam provide a theoretical exposition of these ideas in their theory of strategic action fields in A Theory of Fields (Fligstein and McAdam 2012). This theory presents an innovative way of thinking about the composition of organizations and institutions, and it helps to capture some of the institutional realities mentioned above. The basic idea is that the fundamental structure of social life is “agents behaving strategically within a field of resources and other agents.” On this approach, organizations and institutions are ensembles of agents-in-fields, at a range of levels. According to this theory, organizations are configured around incumbents who are assigned roles and powers that give them both an interest and an ability to maintain the workings of the organization. So stability is not a primitive quality of an organization; instead, organizational persistence over time is a consequence of the specific interlocking assignments of interests and powers within the various networks of agents that make up the organization. Stability is a dynamic feature of the organization, reproduced by the actions of incumbents. And change in the organization occurs when there is significant alteration in those interests and powers. On this approach, then, stability of an organization is a consequence of the configuration of a given system of strategic fields,

38  Daniel Little rather than an axiomatic property of the organization. Both stability and change need to be explained by recourse to the agents and coalitions that are in play. Fligstein and McAdam provide a powerful theory in terms of which to conceptualize organizations, and it is a theory that appears to have a great deal of empirical reality when one considers a range of case studies of organizational dysfunction, from the failure of United States antisubmarine warfare in World War II (Cohen and Gooch 1990) to the failure of regulation of the nuclear industry in the late 20th century (Perrow 2011), to the highest-level conflicts between and within the departments of State and Defense during the Vietnam War (Halberstam 1972).

How Organizations Change What, then, are the primary mechanisms of change within a social organization? There are many, but almost all derive from the deliberate interventions of influential actors who are subject to specific incentives and interests. The most common answer to the question of how organizations improve over time is the factor of “organizational learning”. The primary mechanism of organizational learning is quasi-intentional. We may postulate that intelligent actors control various aspects of the arrangements of an organization, these actors have a set of interests that drive their behavior, and actors fine-tune the arrangements of the organization so as to serve their interests. This is a process we can describe as quasi-intentional to convey that the organization itself has no intentionality, but its behavior and arrangements are under the control of a loosely connected set of actors who are individually intentional and purposive. Organizations are the joint product of the actions and plans of multiple intelligent, goal-directed actors, and it is reasonable to expect that intelligent actors with positions of influence within an organization will see opportunities for improved functioning, and they will undertake the steps of organizational reform needed in order to capture those opportunities. In this case, we need to provide a theory of the mechanisms that might connect the mission of the organization to the interests, desires, and goals of the actors who are in a position to affect its activities through some kind of identifiable feedback process. An important set of systemic pressures for change in an organization derives from the facts of observability and accountability of the performance of the organization. An organization has stakeholders, and its operations have effects on a broader population of actors who have various means of protest and criticism at their disposal. When the stakeholders observe poor or unsatisfactory performance, they demand change. Financial performance is the most visible driver of change for business organizations; if a company is consistently losing money, its investors will demand change or will sell their shares. Bad publicity, poor financial

Do Organizations Adapt?  39 performance, a rash of harmful accidents, or public exposure of internal improprieties can lead to an effective demand for change which motivates leaders to seek out changes. Public embarrassment is another important driver of change. When it was revealed in 2015 that Red Cross International (RCI) had apparently squandered hundreds of millions of dollars in disaster assistance to Haiti, the organization was compelled to change and to document the changes it had implemented (Elliott and Sullivan 2015; American Red Cross 2017). This case invokes both embarrassment and revenues; if RCI loses its reputation as a trustworthy agent in the world of disaster relief, its primary source of revenue (gifts) will disappear. Internal planning and intelligence gathering are other drivers of change. Organizations devote resources and time to attempting to predict the environmental, technological, and business changes that are on the horizon, and they make efforts to adjust their plans in such a way as to take advantage of those prospective changes. (Willy Shih provides fascinating detail for the reasons that Kodak was unable to reorient its business rapidly enough from film to digital photography; see Shih (2016).) Finally, there are sources of change that can only be described as stochastic. Managers and directors change, and they make small changes in the processes they administer. New habits emerge which over time begin to replace the formal processes of the office. Staff misunderstand or misremember the processes of the office. And the new arrangements are now passed on and become the new normal. So we can summarize the main drivers of organizational change in these terms. (i) The business or political environment in which the organization conducts its activities may change, leading to changes in demand for its services, or changes in the cost of the resources it needs to carry out its mission. (ii) Technological change may lead to large changes in the products the public wants from the organization, or the ways in which those products are produced, or the ways in which the organization needs to communicate with its stakeholders. Technological change may be predictable (the digital revolution in photography was certainly foreseeable in 1990), or it may fall from the sky. Either way, many organizations are faced with an existential necessity to adapt or fail. (iii) Many organizations perform their activities within a sphere in which state regulators have substantial influence on the nature of their activities. Changes in vehicle mileage and emissions affect the auto industry, changes in financial regulation affect the banking industry; changes in environmental protection affect the energy industry, and changes in labor law create a necessity for change in labor unions. These are some structural forces driving organizational change. Demand for change is voiced by different configurations of actors. Stakeholders can demand improved performance in a variety of ways, from shareholder resolutions to public protests. The public can voice its anger

40  Daniel Little and dissatisfaction with an organization through opinion polling, boycotts, and shifts of demand and support. Bad publicity can put pressure on an organization to improve its performance and its public image. Each of these forms of stakeholder voice has implications for the incentives and interests affecting decision-makers within the organization. Executives can lose their jobs (or their bonuses), funding can dry up for a nonprofit organization with a bad reputation, leading to reduction or extinction, and public outrage about an organization’s failures of performance can lead to legal and regulatory inquiries. So these factors are often effective inducements to senior executives within organizations to introduce changes leading to improved performance. The problem-solving that the organization needs to do in order to address these problems of performance is common in virtually every organization. Human resource specialists are called upon to suggest remedies for the various failures identified; executive coaches and consultants may be hired to bring external expertise to bear on the problems identified in organizational structure and culture. An organization is a large, complex system involving interlocking forms of behavior and motivation by actors within the organization. Correcting the functioning of the organization is akin to the problem of identifying and correcting a systems problem in a complex design of a technology system. Change may come in the form of small-scale, local adjustments in many parts of the organization; it may come in the form of a major effort to restructure the operation in a systematic and centralized way; or it may even take the form of a change of the mission and scope of the organization. Organizational history is full of examples of all these kinds of organizational change. Here is the key question, however: to what extent do these various forces for change and correction constitute feedback mechanisms that lead to a direction of change towards improved functioning in a system including organizational arrangements, environment, and stakeholders? The workings and outputs of the organization are the dynamic results of the set of interlocking practices and behaviors described above. The aggregate performance results of the organization – quality of output, profits, safety, reputation, satisfaction – are the outcomes that ensue from the normal workings of these arrangements in a given environment. The performance results influence the interests and behavior of multiple groups of actors – the executive, the governing board, internal actors, the public, the regulators, and other stakeholders. This description thus captures a feedback process from arrangements to performance and output, back to the actors who have interests in either maintaining or changing the arrangements of the organization. Poor profitability performance will influence stakeholders and governing board to apply pressure on the executive for change. Poor safety performance will influence regulators and the public to demand change. As change is formally adopted, the

Do Organizations Adapt?  41 governing board and the executive modify the arrangements of the organization (by implementing new rules, roles, and practices into the definition of the organization). Workers are influenced by the actual practices and behaviors (working conditions), and also have the constrained ability to modify these practices and behaviors. This representation of an organization allows us to identify the dynamics of change, in that influential actors react to outcomes and provide incentives to internal actors to modify the workings of the organization. We can also see that not all these forces for change are “functional” from the social point of view. The pressure for enhancing profits may lead to changes that give less attention and resources to safety and may create incentives for concealing defects and bad business practices. The private incentives executives have for security, income, and wealth may lead to changes leading to short-term profitability but long-term stagnation. The incentives workers have for comfortable working conditions may lead to changes that interfere with safety, quality, and profitability.

Assessment This analysis provides an account of the causes of organizational change and an answer to the question of organizational adaptation. Organizations are indeed subject to feedback processes that are likely to lead to change in the arrangements that embody them, and some of those processes are likely to lead to “improved performance.” The reason for this expectation is that the feedback processes noted here lead to new or reinforced incentives for actors who are in a position to modify the arrangements of the organization so as to respond to the negative feedback information. Finding solutions to problems of cost, quality, consumer appeal, safety, or environmental sustainability that lead to negative feedback signals is of course difficult, and not all intelligent efforts at organizational reform are successful. But this process of feedback and adjustment is one that is plausibly as likely to succeed as the processes of technical adaptation that Elster describes for technology. If the executives and other actors were solely motivated by the publicly stated mission of their organization, and if they had full control of the arrangements and functioning of the organization, then this would have been a satisfying conclusion. It would sound quite a bit like Mandeville: there is a happy coincidence between the public’s interest and the private interests of the leaders and executives of organizations, and this coincidence can be expected to lead to improving performance over time. However, this optimism reflects one of the most important errors of functionalist sociology: the assumption that social institutions and practices work for the advantage of society as a whole. Rather, it was emphasized above that organizations are subject to a range of interests by powerful actors that affect their performance, both

42  Daniel Little internally and externally. So the dynamics leading to favorable change are not the last word. The internal dynamics of an organization as an actually existing human system are more conflictual than a simple model of market-driven efficiency suggests. Recall the theory of the organization sketched by Fligstein and McAdam (2012), as a set of agents and alliances with often conflicting interests. The conclusion to be drawn from this fact points in the opposite direction from functional adaptation. We have identified important currents of organizational change that correspond to the “blocking” power that powerful insiders and outsiders have to make reform of ineffective institutions and organizations enormously difficult. These examples suggest that the public’s interests in safe, high-quality consumer products, stable banking institutions, and a well-regulated nuclear industry may well be defeated by the private interests of coalitions of self-interested actors who have the ability to block reforms or to create innovations that favor their private interests at the expense of the public good. This conclusion is supported by Charles Perrow’s careful review of the regulatory systems the United States has created to oversee risky industries (Perrow 2011). Jack Knight reaches a similar conclusion: Institutions [organizations] are not created to constrain groups or societies in an effort to avoid suboptimal outcomes but, rather, are the by-product of substantive conflicts over the distributions inherent in social outcomes. According to this conception, the main goal of those who develop institutional rules is to gain strategic advantage vis-à-vis other actors, and therefore, the substantive content of those rules should generally reflect distributional concerns. (Knight 1992, 40) Two things emerge from this discussion. First, we have identified clear mechanisms that lead executives and other actors within an organization to have incentives to improve the performance of the organization, noting the difficulty and uncertainty attaching to solving problems of organizational performance in a complex area of activity. But we have also noted the existence of conflicting interests among the actors who control and influence organizational change, and the fact that these interests do not necessarily align with the public interest or with the achievement of the official mission of the organization. So here we have the basis for a partial “yes” to the question posed above: there is an identifiable chain of causal mechanisms that is likely to lead to improving performance in organizations over time. At the same time we have identified systemic reasons for expecting the development of “dysfunctional” innovations within organizations when the interests of internal and external actors prefer them. It is therefore impossible to give an unambiguous answer to the question of the adaptation of

Do Organizations Adapt?  43 organizations. Each case of organizational history is sensitive to the particular incentives, beliefs, powers, and opportunities available to the various actors and stakeholders, and the net result of these feedback processes is indeterminate. Organizational change is not purely stochastic, but its direction is unpredictable nonetheless.

Notes

References Abend, Gabriel. 2014. The Moral Background: An Inquiry into the History of Business Ethics, Princeton Studies in Cultural Sociology. Princeton, NJ: Princeton University Press. Aldrich, Howard E., and Jeffrey Pfeffer. 1976. “Environments of Organizations.” Annual Reviews 2: 79–105. American Red Cross. 2017. “American Red Cross Responds to Latest ProPublica and NPR Coverage.” Accessed May 29, 2017. www.redcross.org/news/press-release/ American-Red-Cross-Responds-to-Recent-ProPublica-Report-on-Haiti. Baldwin, Robert, Martin Cave, and Martin Lodge. 2012. Understanding Regulation: Theory, Strategy, and Practice. 2nd ed. Oxford; New York: Oxford University Press. Bates, Robert H. 1981. Markets and States in Tropical Africa: The Political Basis of Agricultural Policies, California Series on Social Choice and Political Economy. Berkeley: University of California Press. Beekun, Rafik I., and William H. Glick. 2001. “Organization Structure from a Loose Coupling Perspective: A Multidimensional Approach.” Decision Sciences 32 (2): 227–250. Cohen, G. A. 1978. Karl Marx’s Theory of History: A Defence. Princeton, NJ: Princeton University Press. Cohen, Eliot, and John Gooch. 1990. Military Misfortunes: The Anatomy of Failure in War. New York: Vintage. Crozier, Michel, and Erhard Friedberg. 1980. Actors and Systems: The Politics of Collective Action. Chicago, IL: University of Chicago Press. Elliott, Justin, and Laura Sullivan. 2017. “How the Red Cross Raised Half a Billion Dollars for Haiti and Built Six Homes.” ProPublica 2015. Accessed May 29, 2017. www.propublica.org/article/how-the-red-crossraised-half-a-billion-dollars-for-haiti-and-built-6-homes. Elster, Jon. 1983. Explaining Technical Change: A Case Study in the Philosophy of Science. Studies in Rationality and Social Change. Cambridge; New York; Oslo: Cambridge University Press; Universitetsforlaget.

44  Daniel Little Fligstein, Neil, and Doug McAdam. 2012. A Theory of Fields. New York: Oxford University Press. Halberstam, David. 1972. The Best and the Brightest. New York: Random House. Klitgaard, Robert E. 1988. Controlling Corruption. Berkeley: University of California Press. Knight, Jack. 1992. Institutions and Social Conflict. The Political Economy of Institutions and Decisions. Cambridge; New York: Cambridge University Press. Lammers, Cornelis J. 1978. “The Comparative Sociology of Organizations.” Annual Review of Sociology 4: 485–510. Lawrence, Thomas B., and Roy Suddaby. 2006. “Institutions and Institutional Work.” In Sage Handbook of Organizational Studies, edited by Stewart R. Clegg, Cynthia Hardy, Tom Lawrence and Walter R. Nord, 215–254. London; Thousand Oaks, CA: Sage Publications. Little, Daniel. 2015. “Guiding and Modelling Quality Improvement in Higher Education Institutions.” Quality in Higher Education 21 (3): 312–327. Merton, Robert K. 1963. Social Theory and Social Structure. New York: Free Press. Miller, Gary J. 2005. “The Political Evolution of Principal-Agent Models.” Annual Review of Political Science 8: 203–225. Mueller, Julia. 2015. “Formal and Informal Practices of Knowledge Sharing between Project Teams and Enacted Cultural Characteristics.” Project Management Journal 46 (1): 53–68. Orton, J. Douglas, and Karl E. Weick. 1990. “Loosely Coupled Systems: A Reconceptualization.” Academy of Management Review 15 (2): 203–223. Perrow, Charles. 1999. Normal Accidents: Living with High-Risk Technologies: With a New Afterword and a Postscript on the Y2K Problem. Princeton, NJ: Princeton University Press. ———. 2011. The Next Catastrophe: Reducing Our Vulnerabilities to Natural, Industrial, and Terrorist Disasters. Princeton, NJ: Princeton University Press. ———. 2014. Complex Organizations: A Critical Essay. 3rd ed. Brattleboro, VT: Echo Point Books and Media. Rodwin, Marc. 2011. Conflicts of Interest and the Future of Medicine: The United States, France, and Japan. Oxford; New York: Oxford University Press. Shih, Willy. C. 2016. The Real Lessons from Kodak’s Decline. Cambridge, MA: MIT Sloan Management Review. Skogan, Wesley G. 2008. “Why Reforms Fail.” Policing & Society 18 (1): 23–34. Stiglitz, Joseph. 1989. “Principal and Agent.” In The New Palgrave: A Dictionary of Economics, edited by John Eatwell and Murray Milgate, 966–971. London: Macmillan. Weick, Karl E. 1976. “Educational Organizations as Loosely Coupled Systems.” Administrative Science Quarterly 21 (1): 1–19.

3

Social Dysfunctions Heiner Koch

The conservative tendency of functionalism lies in its functionally explaining institutions as sustaining (existing) society. There is no conservatism when institutions, and society itself, are explained as serving a development of power which prevails against forms of society resisting it. (Cohen 1978, 285) To the extent that functional analysis focuses wholly on functional consequences, it leans toward an ultraconservative ideology; to the extent that it focuses wholly on dysfunctional consequences, it leans toward an ultra-radical utopia. (Merton 1968, 114)

Introduction In the most general sense of the term, dysfunctions are bad effects. Depending on what you want to achieve by talking about dysfunctions, much stronger claims must be made. In this chapter, I am mainly interested in the role social dysfunctions play for emancipatory transformations of society. But because dysfunctions can, in general, be a relevant factor for the explanation of social change and the destabilization of social entities or even societies as a whole, I want to provide a basic framework for investigating social dysfunctions. I will do so mainly by addressing conceptual problems rather than detailed explanatory models that involve dysfunctions. Thus, I do not want to emphatically argue for a specific account of dysfunctions. Instead, I want to investigate the benefits and costs of adopting one concept of dysfunction compared to another. My investigation will, therefore, be hypothetical in character: if you adopt this concept of dysfunction, rather than another, then you will be committed to these (normative, ontological, and other) claims and you will be able to explain these phenomena but not others. A broad understanding of dysfunctions as just any bad effect is easy to get (all you need is a normative assumption about what “bad” is), but it is of little explanatory value since bad things do not need to have much in common.

46  Heiner Koch A more demanding understanding of dysfunctions, on the other hand, which presupposes some notion of social evolution or self-reproduction could be of great explanatory value, but the costs are so high that we might not always be able to afford them. It might, for example, turn out that social evolution is something that just does not occur because there is nothing like competition-based differential reproduction in the social world. I summarize the different possible concepts of dysfunction in Table 3.1 at the end of part 2.3. There, I mainly distinguish them based on their explanatory context, the normative claims they involve, and the way bad effects that are not supposed to be covered by the dysfunction concept are excluded. Because functions are often assumed to have some positive effect on entities they are part of, they tend to be used to explain the persistence or stability of the function bearer. If we are interested in social change, it might be helpful to have a closer look at negative effects that destabilize social entities. In early theories of society, these negative destabilizing effects have sometimes been called “anomie” (e.g., Durkheim) or “dysfunction” (e.g., Merton). What motivated my interest in dysfunctions was a functionalist interpretation of Marx in analytical Marxism (especially Cohen 1978). Although I have never been convinced by a strong version of historical materialism – that is, the claim that there is some necessary direction in historical development and that certain forms of social change are inevitable – I was always compelled by the idea of stable tendencies that promote social change towards a more emancipated society. E.O. Wright’s (2010) real utopias project can be seen as a general framework for this kind of ideas. The main question is: what path can we choose today that will help us to achieve a better society tomorrow? Analyzing dysfunctions can be a helpful tool in this quest. Sure, dysfunctions can have catastrophic effects on society and hinder emancipation instead of promoting it. But there can also be dysfunctions that destabilize societies that are based on domination, exploitation, oppression, and violence and at the same time also promote social change towards a more emancipated society. Some of the examples that Marx provides might be better suited if we do not reconstruct them as functional explanations but as explanations that appeal to dysfunctions – even if not in that terminology. I do not think that his arguments are compelling for current configurations of society, but they are interesting examples of the role dysfunctions play in explanations of emancipatory social transformations. Consider the following passage from the Manifesto of the Communist Party of Marx and Engles (1969, 119): The essential conditions for the existence and for the sway of the bourgeois class is the formation and augmentation of capital; the condition for capital is wage-labor. Wage-labor rests exclusively on competition between the labourers. The advance of industry, whose

Social Dysfunctions  47 involuntary promoter is the bourgeoisie, replaces the isolation of the labourers, due to competition, by the revolutionary combination, due to association. The development of Modern Industry, therefore, cuts from under its feet the very foundation on which the bourgeoisie produces and appropriates products. What the bourgeoisie therefore produces, above all, are its own grave-diggers. Its fall and the victory of the proletariat are equally inevitable. The mechanism of social change rests on the dysfunctionality of worker association. Technological development (the advance of industry) unwittingly leads to the association of workers. Worker association is (or maybe was) functional for the organization of labor and capitalistic production. But, at the same time, worker association is dysfunctional for capitalistic power structures because it organizes the revolutionary working class. In this context, the functional explanation concerns the rise and stability of a dysfunction that explains a (future) change or a tendency to change relations of production. I suggest that we should understand dysfunctions in analogy to functions. For reasons of consistency, some of the commitments we make concerning functions will also carry over to dysfunctions. If we, for instance, assume a specific form of normativity or social systems in defining functions, we should assume the same in the case of dysfunctions. Unfortunately “malfunction,” “dysfunction,” “disfunction,” and “maladaptation” are often used interchangeably in the literature. In my understanding, these words do not necessarily mean the same as I will argue later on in more detail. The Greek prefix “dys-” can mean bad, ill; hard, difficult; abnormal, imperfect. It thus allows for at least three different interpretations of “dysfunction.” First, it can mean that there is a deviance from normality (in this case from normal functioning). How to define “abnormal” social entities, however, is unclear (I will say more about this below). Second, it can mean that something ceases to fulfill its function. Institutions like the university could, for example, fail to educate students due to being underfunded. It is the third meaning I am mainly interested in: “dysfunctional” in the sense of “bad effect.” A dysfunctional police force1 in this third sense would be, for example, a threat to people instead of protecting them. One crucial distinction between the first two cases and the last case is that the latter case does not necessarily requires that the dysfunctional entity has a function or was, at some time, functional. Climate change or an economic crisis can be dysfunctional (for society or economy) even if they never had a function. I think it is acceptable to include all three cases under the umbrella of “dysfunction.”2 Nonetheless, I want to restrict the meaning of “dysfunction” in this paper to the third case because it is more specific, otherwise not covered in its specificity, and also theoretically fruitful for explanations of social change.

48  Heiner Koch Early social theorists frequently discussed dysfunctions. Their interest might have stemmed from the predominant representation of societies as organisms (Maasen, Mendelsohn, and Weingart 1995). In current social science, on the other hand, dysfunctions are rarely explicitly discussed. Therefore, I will have a closer look on dysfunctions in early social theory to provide an overview of different concepts of dysfunctions, their explanatory value, and the theoretical problems to which they give rise. Based on the overview in “Early Social Theory” section, I will turn to specific conceptual issues in “Dysfunctions” section. Two main aspects are relevant in defining dysfunction. First, the sources of normativity have to be made explicit (“Normativity” section). Second, because not everything that has a negative effect is a dysfunction, unwanted cases have to be excluded. I discuss different ways to exclude unwanted effects in “Exclusion of Unwanted Cases” section. In “Possible Definitions of Dysfunctions in Relation to Functions” section, I will then compare the different more or less demanding concepts of dysfunction based on their purpose and criteria of exclusion. In “Dysfunctions in Explanations and Prospective Research” section, I will conclude my discussion with some remarks concerning the relation between explanations and dysfunctions and sketch some ideas for prospective research.

Early Social Theory Explicit reference to dysfunctions can be traced back to the very beginning of social sciences when social theorists often conceptualized societies in analogy to organisms and mostly viewed evolution as the development of the social organism (Sanderson 1990). Several similar terms are commonly used interchangeably. Merton (1938), for example, uses the terms ‘malfunctional,’ ‘anomie,’ ‘malintegration,’ ‘maladjust,’ and never ‘dysfunction,’ whereas, in his later work (1968), he mainly uses the term ‘dysfunction’ and never ‘malfunction.’ Radcliffe-Brown (1935) objects to Durkheim’s use of the term ‘anomie’ and prefers ‘dysnomia’ and ‘disfunction.’ I will mainly focus on the work of three theorists: on the one hand, on Radcliffe-Brown and Merton as proponents of structural functionalism (functional roles) because they are the first theorists who make explicit use of the term ‘dysfunction;’ on the other, on Marx because his work contains interesting examples which we can reconstruct as involving etiological functions even if he does not refer explicitly to functions or dysfunction. But first, I want to retrace the concept of dysfunction to Durkheim’s theory of anomie to better understand the motives behind early discussions of dysfunctions. Durkheim (1893) used the concept of anomie to describe pathological states of society. He adopted this concept from Guyau, who reintroduced the greek concept into the philosophical debate (Orru 1983). For Durkheim, anomie is not just norm- or lawlessness. It is a mismatch between individual norms and actions, on the one hand, and the norms,

Social Dysfunctions  49 organization, or needs of society, on the other. We could describe this mismatch as dysfunctional. Interestingly, Guyau (1884) promoted the anarchistic thought that anomie is the future form of an individualistic morality in an industrialized society, whereas Durkheim saw anomie as the negation of morality.3 Guyau believed that education promotes anomie, while Durkheim considered education its solution (Orru 1983). Merton (1938, 1968) follows Durkheim in analyzing the consequences of an anomic society’s failure to provide the means for its members to achieve goals that are also set by the norms of society. Disintegration and deviance are possible outcomes. Contrary to Durkheim, Merton is explicitly morally neutral towards anomie:4 It would be a shortsighted view and a concealed ethical judgment, moreover, to assume that even the deviant behaviour which is dysfunctional to the current values of the group is also ethically deficient … the concept of social dysfunction is not a latter-day terminological substitute for ‘immorality’ or ‘unethical practice.’ A particular pattern of behaviour which departs from the dominant norms of the group may be dysfunctional in lessening the stability of the group or in reducing its prospect of achieving the goals it values. But, judged by one or another set of ethical standards, it may be the norms of the group which are at fault, not the innovator who rejects them. (1968, 236) Radcliffe-Brown (1935) also takes up ideas from Durkheim but is very critical about his concept of anomie. Since dysfunctions are nowadays mainly discussed in philosophical concepts of disease, 5 it is interesting to see that Radcliffe-Brown does not believe that the concept of disease can be transferred to the social domain. In relation to organic structures we can find strictly objective criteria by which to distinguish disease from health, pathological from normal, for disease is that which either threatens the organism with death (the dissolution of its structure) or interferes with the activities which are characteristic of the organic type. Societies do not die in the same sense that animals die and therefore we cannot define dysnomia as that which leads, if unchecked, to the death of a society. Further a society differs from an organism in that it can change its structural type or can be absorbed as an integral part of a larger society. Therefore, we cannot define dysnomia as a disturbance of the usual activities of a social type – as Durkheim tried to do. (Radcliffe-Brown 1935, 398) If we look at social dysfunctions from an organismic point of view (society as some form of social organism in analogy to living beings), we quickly get into normative trouble. Social dysfunctions then naturally

50  Heiner Koch seem to be social pathologies, social diseases that have to be eradicated to save the social organisms6 – ideas that Merton and Radcliff-Brown explicitly reject. In general, it is unclear how to construct reference classes in social science to determine statistical deviance from normal or usual activities. In biology, reference classes can be defined by picking out members of the same natural kind (species) with relevant similar properties (such as age) which have an impact on the activities or effects in question (Boorse 1977). Societies as a whole are so different from each other that it might be far-fetched to subsume them under one (social) kind which we can use as reference class. While Radcliffe-Brown might be right in assuming that this will not be possible for social types (the social type “society” might not be a social kind), it might nonetheless be possible for structural types.7 Structural types are types of social entities such as institutions that are organized in the same way. We can sometimes compare economic units such as factories and then regard statistical deviance as dysnomia. Nonetheless, societies are not likely to “die” if some kind of dysnomia occurs but will instead change their structure or organization. Subsystems of society such as groups, organizations, institutions, or corporations might still cease to exist due to dysnomia or dysfunctions and therefore be, at least in principle, subject to some kind of selection process. In the context of social functioning and social dysfunctions, RadcliffeBrown (1935, 397) prefers to appeal to structural unity, which he describes as “… a condition in which all parts of the social system work together with a sufficient degree of harmony or internal consistency, i.e., without producing persistent conflicts which can neither be resolved nor regulated.” Opposition, as he adds in Footnote 3, is not dysfunctional but an essential feature of societies that might produce conflict but in a regulated manner. Radcliffe-Brown also develops a basic form of dysfunctional explanation: … while an organism that is attacked by a virulent disease will react thereto, and, if its reaction fails, will die, a society that is thrown into a condition of functional disunity or inconsistency (for this we now provisionally identify with dysnomia) will not die … but will continue to struggle toward some sort of eunomia, some kind of social health, and may, in the course of this, change its structural type. (1935, 398) Merton argues quite similarly: … when the net balance of the aggregate of consequences of an existing social structure is clearly dysfunctional, there develops a strong and insistent pressure for change. It is possible, though this

Social Dysfunctions  51 remains to be established, that beyond a given point, this pressure will inevitably result in more or less predetermined directions of social change. (1968, 113) However, dysfunctions will not inevitably lead to social change. Organisms do not die from most diseases; they can even suffer from chronic illness. And if societies rarely die, it is plausible to assume that societies can be in a persisting state of “illness,” functional disunity, or dysfunctionality. To make the claim plausible that dysfunctions lead to social change, change of type, or a specific predetermined change, we would have to find the mechanism that explains the occurrence of change in more detail and supports a causal description of what is happening. Without a mechanism, we would have to at least find a lawlike statistical connection between (specific) dysfunctions and a (specific) social change. These would be easier to get if particular dysfunctions (for social needs) would lead to the “death” or demise of society or a subsystem of society. Then, that social system would either die or adapt. In doing functional analysis, Radcliffe-Brown was not so much interested in explaining social change but, instead, how social elements work together and constitute a functional unity: There is not, and cannot be, any conflict between the functional hypothesis and the view that any culture, any social system, is the end-result of a unique series of historical accidents … One “explanation” of the race-horse is to be found in its history – how it came to be just what it is and where it is. Another and entirely independent “explanation” is to show how the horse is a special exemplification of physiological laws. Similarly one “explanation” of a social system will be its history … Another “explanation” of the same system is obtained by showing (as the functionalist attempts to do) that it is a special exemplification of laws of social physiology or social functioning. (1935, 400–401) Things are different if we try to theorize history in analogy to natural evolution. In Karl Marx’s Theory of History: A Defence, Cohen (1978) tried to reconstruct historical materialism in terms of functional explanation. But how to understand functions in Marx’s work is disputed. For instance, interpreters have advanced interpretations in the spirit of functional analysis (e.g., Iorio 2003) and evolutionary theory (e.g., Cohen 1978). In the introduction, I already gave an example which, in my opinion, involves a very specific argument based on the dysfunctional effects of workers’ association. We can find another less specific version

52  Heiner Koch of dysfunctional explanation without reference to a particular mechanism in the preface to Marx’s A Contribution to the Critique of Political Economy from 1859: At a certain stage of their development, the material forces of production in society come in conflict with the existing relations of production, or – what is but a legal expression for the same thing – with the property relations within which they have been at work before. From forms of development of the forces of production these relations turn into their fetters. Then comes the period of social revolution. With the change of the economic foundation the entire immense superstructure is more or less rapidly transformed. (1904, 12) The exact nature of the causal process that leads to social revolution due to the conflict between productive forces and production is unclear and leaves room for interpretation. We can distinguish at least three main lines of analysis: rational adaptation, competition, and dysfunction. The first two analyses are commonly spelt out in functional terms without explicit reference to dysfunctions. But there is no clear line between these interpretations, for we can also understand rational adaptation and competition in terms of dysfunctions. Furthermore, rational adaptation and competition are not exclusive explanantia and can be combined. Cohen’s (1978, 150–160) main argument is that it is irrational to stick to suboptimal economic practices and, for this reason, people change them (intentional primacy). This kind of explanation has some similarities to what we find in Radcliffe-Brown’s work. If some elements of society are not in harmony, society will adapt. This process seems to involve – in Radcliffe-Brown as well as in Cohen – some kind of intentional adaptation. Nolan (2006), Wray (2002), and Carling (1993) criticize this line of argumentation. They think that Cohen (1978) also allows for another version of functional explanation, in which selection processes in competitive situations play a decisive role (competitive primacy). This view amounts to an evolutionary interpretation of Marx, which ascribes a mechanism similar to Darwinian natural selection to Marx’s theory, even if different selection processes are assumed. But it is unclear what the level of selection should be. In Marx and Engels, we can find examples that would support nations (which we could regard as states or societies) to be the level of selection: as they put it, “[t]he cheap prices of commodities are the heavy artillery with which it batters down all Chinese walls … It compels all nations, on pain of extinction, to adopt the bourgeois mode of production …” (Manifesto of the Communist Party 1969, 112). But examples of sub-society levels (economic units like

Social Dysfunctions  53 enterprises) can also be found. Competition between national economies works very differently than competition in biological ecosystems. And, to again cite Radcliffe-Brown: societies rarely die. Intentional primacy and competitive primacy are compatible with the assumption that the relations of production at a given time are perfectly functional but that there are other relations of production that would have a higher functionality (even if the scale is unclear). Due to higher functionality, they are intentionally chosen or supersede the old ones by competition. But “fetters” can also be suboptimal, nonfunctional, or dysfunctional relations of production. Suboptimal relations could turn out to be functional in one respect but at the same time dysfunctional in another if they, for example, cause an economic crisis (e.g., overproduction). In my view, the following example taken from the Communist Manifesto suggests that dysfunctional elements are, at the very least, involved: The productive forces at the disposal of society no longer tend to further the development of the conditions of bourgeois property; on the contrary, they have become too powerful for these conditions, by which they are fettered, and so soon as they overcome these fetters, they bring disorder into the whole of bourgeois society, endanger the existence of bourgeois property. The conditions of bourgeois society are too narrow to comprise the wealth created by them. And how does the bourgeoisie get over these crises? On the one hand by enforced destruction of a mass of productive forces; on the other, by the conquest of new markets, and by the more thorough exploitation of the old ones. That is to say, by paving the way for more extensive and more destructive crises, and by diminishing the means whereby crises are prevented. (Marx and Engels, Manifesto of the Communist Party, 114) It makes a difference for explanations of social change if we see fetters as being functional and replaced by more functional relations of production or if we see fetters as dysfunctional, causing problems that have to be dealt with whether there are or could be better relations of production or not. This understanding of explanations in terms of dysfunctions seems to be close to Merton’s and Radcliffe-Brown’s dysfunctional explanation in the context of functional analysis. But it can also be combined with evolutionary explanation models if we see fetters as formerly functional adaptations that became dysfunctional maladaptations due to a changing environment (development of technology and the productive forces). This view is, again, similar to Cohen’s account (1978, 232). Social change in historical materialism typically has some form of directedness. This feature also distinguishes the historical materialist understanding of social change from evolution in nature since nature

54  Heiner Koch knows no direction. Carling (1993) argues that historical materialism would be better off with combinations of theories that are general and undirected or directed but specific to historical cases. Merton (1968, 113) thought that it is at least possible that dysfunctions lead to a predetermined direction of social change, but he was also unsure whether dysfunctions even lead to a social change that reduces the number of dysfunctional elements (129). So far, we have seen that, in early social theory, dysfunctions can play different roles. They can be more descriptive (regarding the consistency of social processes), or they can explain social change (in historical materialism even a directed social change). Depending on what dysfunctions are supposed to do, we theorize them in slightly different ways. In the next section, I will go into more detail of the involved dysfunction concepts and also include current debates and positions.

Dysfunctions The nature of dysfunctions is unclear in the history of social theory. Dysfunctions are rarely used as a concept in contemporary social science. In selectionist accounts of social evolution, dysfunctions do not play a significant role because they are not selected and therefore do not persist. In functional analysis, dysfunctions are not really supposed to happen because the systems do not need them for stability and will consequently not organize in a dysfunctional way – however, from a theoretical point of view they are nonetheless perfectly possible.8 Luhmann, who developed the most advanced theory of social systems, frequently refers to functions but (almost) never to dysfunctions or malfunctions. One area of inquiry in which dysfunctions are still extensively discussed is the life sciences, which attempt to define or understand diseases. Since the analogy between organisms and social systems or societies is limited, the transferal of the biological concept of dysfunction is problematic. In what follows, I will define dysfunctions in terms of different accounts of normativity (“Normativity” section) and the exclusion of unwanted effects (“Exclusion of Unwanted Cases” section). Normativity As I argued in the introduction, we can view dysfunctions as bad effects. To identify bad effects, we must rely on some normative standard. In the case of functions, normativity is relative to a means-end relation or  – what I assume to be the same – a purpose.9 There are two main ways a purpose could be assumed to be established in the case of functions: intentions and existence. We can intentionally assign a purpose to entities or an entity persists or it is reproduced because it has a particular effect and therefore it exists for the purpose of bringing about that effect.

Social Dysfunctions  55 Sometimes it is assumed that entities that can maintain themselves have a good on their own (survival). Contributions to persistence or reproduction or to an intentionally assigned purpose are beneficial. Searle, for instance, argues that functions are intentionally assigned. As he puts it: Even natural phenomena, such as rivers and trees, can be assigned functions, and thus assessed as good or bad, depending on what function we choose to assign to them and how well they serve those function … The important thing to see at this point is that functions are never intrinsic to the physics of any phenomena but are assigned from outside by conscious observers and users. (Searle 1995, 14) In his view, these collectively assigned status functions are a building block of social reality.10 Dysfunctions will then occur if a trait that belongs to the entity to which we assign a function has an effect that hinders the fulfillment of that function. I would say that defining existence, reproductive success, survival, or persistence as a good also amounts to assigning a purpose intentionally, even if it is a special one. People who want to naturalize the normativity of functions like Ruth Millikan (1984) or Philippa Foot (2001) might disagree. In the case of etiological functions, the relevance of a purpose could also be denied in general since it is only selection history that justifies describing something as a function. This purpose-free account of functions would be very limited (it wouldn’t allow for theorizing artifacts and intended functions), and dysfunctions (as opposed to malfunctions) could hardly be defined since it is unclear how selection history without purposes could provide the normativity we need to determine dysfunctions. And I would not even know how to try to determine dysfunctions without any reference to some kind of normativity. Also in the case of functions it seems to me that it would not make much sense to claim that a function exists for something if that acclaim would rest on a merely causal story and not on a purpose. Why should we introduce ‘existing for’ in a purely causal story? Bedau (1991) argues that we then might have to describe effects of parts of (nonorganic, natural) reproducing entities as functions, as in the case of clay crystals and this would be a very unintuitive nonproductive way of talking about functions. One could argue that, in the case of functional analysis, we analyze only causally without appealing to any purpose. The difference between a simple causal effect and a function could then consist in the fact that we analyze the cause as a contribution to another effect, ability, or capacity. The gravitation of the sun would then be functional for Venus’s orbit around the sun. Using the term ‘function’ instead of ‘causation’ would then seem pointless to me. Cummins (1975) argues that contributions to a more complex effect of the containing system justify the use of functional terminology. I will discuss this account in “Exclusion of

56  Heiner Koch Unwanted Cases” section. However, I think that it is not convincing and we still need some purpose.11 Instead of a purpose, we could also appeal to an independent, external standard for judging whether an effect is good or bad. This is, for instance, sometimes the case in the definition of disease as dysfunction (e.g., Boorse (1977) argues for statistical deviance and Wakefield (1992) for some form of cultural normativity). But I agree with Radcliffe-Brown (1935, Footnote 7) and Merton (1938, 1968) that social dysfunctions are not inherently bad. They are bad in the same sense that functions are good. The normative standard at play in our evaluation of the dysfunction in question is relative to a purpose and not to morality. Dysfunctions do not bring with them their own normative standard. They do not have a purpose. They are bad for what the corresponding function is good for. Dysfunctions can be intentionally designed for a purpose or intentionally reproduced for their dysfunctional effect (some examples would be sabotage, debunking ideology, computer viruses, or corruption). But then, the same effect is a function relative to the purpose of the designer or user and a dysfunction relative to the entity for which it is bad. Functions can stand in different relations to a purpose. They can acquire their purpose directly (intentionally assigned or existing for it), or they can contribute to the fulfillment of a purpose that they do not have on their own. If they just contribute to a purpose, their own purpose is determined by their contribution to that purpose. This is typical for functional role analysis of entities. The normativity of functions is also necessary to distinguish functions from malfunctions or to judge function bearers relative to how well they do what they are supposed to do. Some people argue that a token that belongs to a functionally defined type cannot malfunction (e.g., Cummins 1975; Davies 2000). If a token malfunctions, it does not function and therefore does not belong to the functional type. Specific understandings of (proper) functions (e.g., Millikan 1989) and types (e.g., Sullivan-Bissett 2017) are proposed to solve this problem. But on the account of dysfunctions I have advanced here, they are not to be confused with malfunctions. Dysfunctions can be effects of entities that once had a function and are now malfunctional, but they do not have to be. Their malfunctions do not define their dysfunctionality. Malfunctions can also be bad for an entity; they can even lead to the death or demise of a system or organism. But the main difference between the negative effect of some malfunctions and that of dysfunctions (in the sense promoted here) is that malfunctions do not do what they are supposed to do and are therefore bad with regard to a purpose while dysfunctions are not bad because of what they are not doing but because of what they are actually doing.12 On this view, a malfunctional immune system would work only with 20% of its normal rate, whereas a disfunctional (and also malfunctional) immune system

Social Dysfunctions  57 would stop to work and have no functional effect at all. A dysfunctional immune system would have an adverse impact on the organism. It could for example attack parts of the body, as in the case of autoimmune diseases. A simple test to distinguish nondysfunctional malfunctions from dysfunctions is whether it would be better for the purpose in question if the effect of the entity would stop to exists or not. Even a malfunctional heart is better than none, but a removed inflamed appendix is better than having one. Things that do not make a difference, like a heart that is not beating, are disfunctional/nonfunctional (and possibly malfunctional). Dysfunctions are neither defined by being malfunctions, nor are they supposed to do something and can malfunction. So in my understanding of dysfunction it does not matter whether a dysfunction belongs to a functional type or not. Therefore, I will not address the question whether and how things can malfunction. Nonetheless, it is an interesting question how to define social malfunctions. Drawing the line between social functioning and social malfunctions could be even more complicated than in the case of organisms (Schwartz 2007). In Figure 3.1, the distinction between functional, malfunctional, disfunctional, and dysfunctional is summarized and related to hindrance and contribution to a purpose. Exclusion of Unwanted Cases There are cases in which, although something is bad relative to a purpose, we would not want to describe it as dysfunctional. Moreover, not everything that is good relative to a purpose is also functional. Therefore, we have to specify functions and dysfunctions further to rule out counterexamples. Effects we want to exclude usually are accidental contributions to a purpose, but we might apply other criteria, as well. In the case of dysfunctions, we also want to exclude some hindrances to fulfilling a purpose. It does not make a lot of sense to say that a rock which falls on someone’s head due to a natural disaster is dysfunctional for the human organism. Although an undemanding definition of both function and dysfunction according to which everything that is good or bad relative to a purpose constitutes a function or dysfunction might be

Figure 3.1 Functionality in relation to effects on a purpose.

58  Heiner Koch fully coherent, but I do not think that this would amount to an interesting construal of these notions. Especially if we are interested in explanations in terms of dysfunction, including such cases seems less fruitful to me. For this reason, I suggest that we regard such cases as instances of mere causes we are interested in (due to their good or bad effect), rather than instances of function or dysfunction. Even more undemanding is an understanding of functions and dysfunctions along the lines of functional analysis, which neither presupposes any purpose nor excludes any effect that causes or prevents another effect or capacity. On this account, rain would be regarded dysfunctional for a spreading fire. To be fair, matters do get more complicated on Cummins (1975) account. For instance, he rules out certain effects as being functional if they do not contribute to a more complex capacity.13 Moreover, he introduces a further criterion of functionality: belonging. Function bearers need to be contained in the system for which they are functional. Merton (1968) again is more liberal in this respect. Functions and dysfunctions can be external to the entity of the purpose they are affecting. In the interest of consistency, we should rule out certain dysfunctional effects by the same standards that we also apply to functions. Standards concerning functions will differ depending on whether they concern intentional function ascription, artifacts, functional analysis, self-maintaining systems, or selected effect functions. In highlighting these connections, I am, above all, concerned with consistency and explanatory value, rather than with arguing for any particular account. The effects many authors mainly aim to exclude are accidental effects. Typically, they exclude such effects by requiring nonaccidental existence (Ehring 1986). An example Ehring gives is of a bullet that is accidentally stopped by a coin in the pocket. The coin does not have the function to stop bullets (even if we might be inclined to say – and I am not – that the coin was functional for stopping the bullet).14 We can exclude such effects by including the condition that the function bearer exists or persists because it has beneficial effects for the system in question in our definition of functions. If the function bearer exists or persists because of its beneficial effects, it does not have this effect accidentally. This functional explanation comes in two major versions: functions can exist/persist because they contributed to fitness or functions can exist/persist because they contributed to the survival/ maintenance of the system of which they are part and would not exist without the system. On the organizational account (OA) advanced by Mossio, Saborido, and Moreno, an entity is not functional but only useful in the following case: An entity is useful if it contributes to the maintenance of an organizationally closed and differentiated system, without being produced under some constraint generated by the system. Accordingly, various

Social Dysfunctions  59 entities such as oxygen, food, and gravity are useful for a given class of self-maintaining systems, without being functional … useful entities are subject to the same kind of normativity as functional entities (one can answer the question ‘What is the oxygen for in an organism?’). (2009, 832) Neither of these versions can be applied to dysfunctions straightforwardly. Dysfunctions do not exist or persist because of their dysfunctional effect. They neither contribute to the survival of the system of which they are part nor do they contribute to the fitness of the entity of which they are part – quite the contrary. If we want to exclude accidental bad effects by reference to the same standard as in the case of functions, we will have a problem. But if we weaken the condition slightly, we will be able to accommodate these cases after all. On the modified account, dysfunctions can exist or persist not because of their dysfunctional effect but rather because of factors that are similar to those which explain the existence or persistence of function bearers. Thus, dysfunctions persist because they are maintained by the system of which they are part. They also persist because they are reproduced by the reproductive unit of which they are part. In these cases, they are not accidental negative effects because they are systematically maintained or reproduced by the system of which they are part. Being maintained or reproduced by a system normally involves being part of the system. We might even tend to say that in these cases belonging to a system depends on being maintained or reproduced by the system.15 But we can also imagine cases in which something is maintained by the system but is external to it. Pollution of the environment is maintained or reproduced by the social system but it is not part of it. Still, we might want to say that a polluted environment is dysfunctional for maintaining the social system. Being maintained or reproduced can at least be distinguished from belonging as a criterion for excluding accidental negative effects. Another criterion for excluding accidental negative effects consists in having a systematic effect or interacting systematically with the system in question. Systematic effects are tightly connected to maintenance and belonging since it is plausible to assume that things that are part of something systematically interact with each other, and in the case of maintenance or reproduction systematic interaction also seems necessary. Belonging without maintenance or reproduction is a criterion which Cummins (1975) uses and which is – contrary to his complexity criterion  – rarely discussed. In Cummins understanding of functions, they need to be contained in (belong to) the system to whose capacity they contribute. But it is unclear whether every contribution to the overall capacity of a more complex system already establishes belonging

60  Heiner Koch (Cummins 1975, Footnote 18), whether some explanatory interest and analytical account figure in selecting the relevant system, contributions and belonging, or whether further nonexplicated criteria define “belonging to a system”. Nonetheless, at least in the case of dysfunctions in the context of functional analysis, ‘belonging to the (relevant) system’ is a plausible candidate for a criterion that distinguishes mere negative effects (that are not contained in the system) from dysfunctions. In this case, of course, we do not distinguish accidental negative effects from dysfunctions since functional analysis allows for accidental effects. Again, belonging might just guarantee that we are dealing with systematic interactions of the function or dysfunction bearers within the entity of which they are part. In the case of artifacts, accidental contributions can be ruled out by appealing to some design, plan, or intended use. Contributions to the purpose of the artifact are accidental if they were not included in the original design, plan, or intended way of using the artifact. This allows for an interesting distinction of accidental negative effects from dysfunctions. A part of a computer that has a negative effect relative to the purpose of the computer (by causing malfunctions or even by destroying the computer) is just an accidental negative effect if it is part of the computer due to an unintended mistake in the production process. But if, for example, a condenser is built into the computer by design and has this negative unintended effect, it is a nonaccidental dysfunctional part of the computer. We could even argue that the entity which exists because of a mistake in the production process and not by design is not part of the computer. Again belonging, existence, design, and systematic effects are closely related. In the case of social artifacts (such as institutions) unforeseen frictions between parts that were intentionally designed and have intended effects that interact in an unintended way and therefore become dysfunctional are important subjects of research for Merton (1968). In the case of intentionally assigned functions (like making noise of the heart or being a border to a geographical line), we can exclude things to which we have not assigned a function. Therefore, parts of an entity that contribute to its purpose do not automatically have a function. Even contributions to an intentionally assigned function need to be assigned a function separately to have a function. Parts of the heart do not have the function to contribute to the noisemaking of the heart. They just happen to have these effects. Then, we could either say that, on this account, dysfunctions do not exist (since we do not intentionally assign dysfunctions to entities or if we do so their dysfunctionality is the function that we have assigned) or we exclude nothing and everything that is a hindrance to the fulfillment of a purpose is dysfunctional. We could also include everything that is part of the entity we assigned the function to and exclude every negative effect that does not belong to the entity

Social Dysfunctions  61 we assigned the function to. This can, for example, be important in the case of organizations. We can define conditions of membership in them. We can also define purposes for organizations or intentionally assign a function to them (the fire department, for example, has the intentionally assigned function to extinguish fires). The behavior of external persons could then be hindrances to the fulfillment of the fire department’s function, but only the behavior of persons who are members of the organization is dysfunctional (if we use the belonging criterion). Possible Definitions of Dysfunctions in Relation to Functions After having specified purpose/normativity and exclusion criteria as defining elements of functions and dysfunctions, we get several different concepts of dysfunction corresponding to the respective concepts of a function we have adopted as can be seen in Table 3.1. I added a version that I called “systematic” because the idea that functions or dysfunctions need to have some kind of systematic effect seems to be implied in several different concepts. Systematic effects are also of particular interest since they are of most interest for explanatory purposes since they occur frequently or reliably and might even allow lawlike statements Table 3.1 Overview on different concepts of dysfunction Explanatory context

Purpose

Exclusion

Dysfunction

Etiological

None or fitness

Existing for

Bad for fitness and reproduced by the system Bad for system maintenance and being maintained by the system None or hindrance to a more complex capacity and/or of a system it is part of Bad for intended purpose and existence/use according to design/ plan/use None or bad for intended purpose or bad for purpose and belonging Systematic hindrances of purpose

Organizational Self-maintenance Maintained

Functional analysis Artifact

None or capacity None or as purpose complexity and/or belonging Intention Design/plan/use

Intention

Intention

Intention, none or belonging

Systematic

Divers

Systematic effect

62  Heiner Koch about them. Based on their organisational account (OA), Mossio, Saborido, and Moreno explicitly argue against excluding accidental effects and against a systematic conception of functions and dysfunctions: A contribution to self-maintenance is accidental if it occurs only in relatively infrequent and occasional circumstances. In the OA, however, the frequency of a contribution to self-maintenance is irrelevant for determining its status. Thus, both functional and useful contributions might be systematic or accidental, relative to the specific context for which we have defined the relevant class of selfmaintaining systems. (2009, 833)

Dysfunctions in Explanations and Prospective Research We have seen that different concepts of dysfunction are relative to different explanatory contexts. But even if my main interest lies in the explanatory value of dysfunctions, I have not developed explanatory models for dysfunctional effects on society. Unfortunately, this will have to be done in further research. Still, I want to point out some interesting features of explanations that involve dysfunctions in the remainder of this paper. The role of functions seems always to be quite intuitive. Function bearers exist or persist because they contribute to a purpose, work well together, and lead to stability, and stability helps them to persist. We are not puzzled by functions. Functions help to dissolve our puzzlement over how something works. That is also a danger of functions in explanations. They make sense of things where no sense is to be found or where the intuitive solution is far from messy reality. Dysfunctions are different. Dysfunctions are puzzling. Where do they come from and why do they persist? What consequences do dysfunctions have for the fulfillment of a purpose? Of course, the answer to these questions depends on the account of dysfunctions you prefer. I will sketch some of the possible ideas: First of all, the question as to where dysfunctions come from is interesting in its own right. As in the case of unintended functions we can take them to be a variation (similar to mutations) that simply occurs without the need for further elaboration. Also similarly to some biological traits, we could regard them as maladaptations. They were once functional but become dysfunctional once the environment changes. This view provides us with one way to interpret Marx’s talk of “fetters.” On this interpretation, they are relations of production that were once functional but became dysfunctional as technology and the productive forces developed further. Merton (1968) often argues that strains and tensions lead to dysfunctions. This claim is often true for organizations or enterprises in

Social Dysfunctions  63 which several elements on their own have only functional effects but in combination produce effects that are dysfunctional. Of course intentions also play an important role in the creation of dysfunctions. Mistakes in design or use can lead to dysfunctional effects and dysfunctions can also be created on purpose as in the case of sabotage. While the existence of dysfunctions is not very puzzling, their persistence is. In etiological and organizational accounts, dysfunctions are a threat to survival. It would be plausible to assume that these threats are eliminated. But immune systems can fail, and getting rid of maladaptive traits is not that easy. In the case of social entities, it is even more plausible that dysfunctions persist once they appear. As Radcliffe-Brown mentioned, societies normally do not die. Dysfunctions are rarely such a threat to societies or subsystems of societies that they vanish if they do not lose the dysfunctional element. But in the cases of societies, it is still puzzling why dysfunctions persist because members of the society in question can get rid of dysfunctions intentionally. Furthermore, the maintenance and reproduction of dysfunctions will involve some kind of intentionality. People could just not care enough to remove dysfunctional elements. Also, it could be unclear who is responsible for removing dysfunctional elements. In light of the Marxist examples I discussed in “Early Social Theory” section, we can sketch some further ideas concerning the appearance and persistence of dysfunctions. Dysfunctions can (even necessarily) accompany functions. One entity may have functional and dysfunctional effects, or one functional entity cannot be maintained or reproduced without the dysfunctional entity. Capitalistic production relations might be functional for a capitalistic society, but at the same time, they cause economic crises that are dysfunctional for capitalistic societies. We could even assume that it is not possible to have capitalistic relations of production without these dysfunctional elements. Also there might be some path dependence that explains the persistence of dysfunctions. Once we designed an institution, it might be difficult to change certain features even if they are dysfunctional. People got used to the way they handle things; they have to establish new forms of coordinated behavior, physical structures they cannot replace might support dysfunctionality, and we might simply not know how to replace the old structure of the institution with a new better one. Institutions may also have some kind of failing regulatory mechanism to detect and remove dysfunctional elements. Furthermore, there might be a conflict of goals. In his contribution to this volume, Daniel Little describes how interests of individuals can conflict with goals of organizations, and how people lack control of organizational processes. Some people may even wish to stick to dysfunctional elements or intentionally maintain them because they foster social change, a change of social structure, or even social type.

64  Heiner Koch One relevant subcategory of dysfunctions comprises what Merton calls ‘latent dysfunctions’ (Merton 1968).16 They are dysfunctions that are either not detected or unintended. Functional explanations in the Marxist tradition often concern unintended latent functions. Elster (1983, 57) even reconstructs functional explanations in social sciences as generally including latent functions. They must either be unintended (in his case assumption (3)) “or at least the causal relation … is unrecognized” (assumption (4)). This criterion is too demanding but stresses the point that latent functions are the most interesting target of functional explanations. The latent character of dysfunctions can also explain why they persist. Before we can remove them, we have to reveal them. Examples of unintended and often but not necessarily unrecognized dysfunctional effects in Marxism are the tendency of the profit rate to fall or economic crises such as overproduction crises. One of the main arguments against functional explanations is the missing mechanism argument (see also Raphael van Riel’s contribution to this volume). Functional explanations often do not provide detailed mechanisms (such as specific feedback loops) for why the effects of functions explain their existence (Elster 1979). Therefore, they are inferior to explanations that provide such mechanisms. In this chapter, I changed the focus in the reconstruction of Marx’s argument from functions to dysfunctions. This shift might also be helpful with regard for the missing mechanism debate since it points towards a mechanism that might be easier to identify: social change by virtue of dysfunctions. Dysfunctions do not need some analogy to natural selection in the social domain. Of course, this has the price that the existence of dysfunctions does need separate explanations since it is hardly imaginable how a dysfunctional effect could explain the existence of the dysfunction (Cohen 1978, 264). The most interesting feature of many dysfunctions is their tendency to promote social change. The nonpuzzling intuitive aspects of dysfunctions (that has to be taken with a pinch of salt) are that they cause society “to struggle toward some sort of eunomia, some kind of social health” (Radcliffe-Brown 1935, 398) or that “there develops a strong and insistent pressure for change” (Merton 1968, 113). Mossio, Saborido, and Moreno (2009, 833) also introduce an interesting distinction that is also relevant for our understanding of dysfunctions. Functions can be dispensable or indispensable for a system. Indispensable functions are necessary for every organization of the system such that a malfunction leads to the demise of the system. Malfunctions of dispensable functions are only necessary for a specific organization of the systems so that a change of the organization is theoretically possible. Analogue dysfunctions may have a negative impact on the maintenance of a system in such a way that it can change its organization to avoid the negative impact or in such a way that the dysfunction will occur under every organization of

Social Dysfunctions  65 the system. As in the case of functional explanations, a specific mechanism is needed to make this claim plausible and provide an interesting explanation of social change.17 Along these lines, Merton stresses the following point: The concept of dysfunction, which implies the concept of strain, stress and tension on the structural level, provides an analytical approach to the study of dynamics and change. How are observed dysfunctions contained within a particular structure, so that they do not produce instability? Does the accumulation of stresses and strains produce pressure for change in such directions as are likely to lead to their reduction? (1968, 107) Answers to these questions will most likely not be possible on a very abstract level. Rather, we will have to analyze specific dysfunctions in their specific context. If similar dysfunctions in a similar context lead to similar forms of social transformation, generalizations might be possible. But it is more likely that we will only achieve a deeper insight into some social mechanisms (or maybe even types of mechanisms) if we analyze dysfunctions (Schweitzer 2015). Finally, I want to make some remarks on a topic that motivated my examination of dysfunctions: stable, emancipatory transformations of society. This motivation might be in line with Marxists’ interest in functional explanations. Such transformative mechanisms would have to be specific and directed (Carling 1993). They cannot be generalized and are specific to a historical situation, and social change is not undirected but directed towards emancipation. The underlying type of mechanism might still be transferable to other situations. The main task would be to identify effects that are (i) dysfunctional for oppressive elements of society, (ii) persist and spread in the system they are part of, (iii) foster social change due to their dysfunctional effect, and (iv) are functional for an emancipated society. Without the last aspect, we would face the danger that the collapsing old structure or organization is not replaced by an emancipatory structure or organization. It does not guarantee the directedness we search for but might support it. Moreover, without (iv) dysfunctions that might promote social change but have a high cost come into play like a crisis, destruction of environment, or extreme suffering. Dysfunctional elements that are functional for an emancipatory society also allow for a smooth transition but do not exclude revolutionary changes. Dysfunctions as elements of a stable, emancipatory transformation of society fit well into Eric Olin Wright’s (2010) real utopias project, into a social science of emancipatory transformations. Certainly a worthwhile prospective research project.

66  Heiner Koch

Acknowledgment I would like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

Notes











12 Since Cummins (1975) argues that functions are not defined by what they are supposed to do but what their actual disposition in relation to the capacities of a system are, he would see dysfunctions and functions in functional analysis as being more similar than dysfunctions and etiological functions.

Social Dysfunctions  67







References Bedau, Mark. 1991. “Can Biological Teleology be Naturalized?” Journal of Philosophy 88: 647–655. Bigelow, John, and Robert Pargetter. 1987. “Functions.” Journal of Philosophy 84: 181–196. Boorse, Christopher. 1977. “Health as a Theoretical Concept.” Philosophy of Science 44, no. 4: 542–573. ———. 2002. “A Rebuttal on Functions.” In Functions, edited by A. R. Ariew and R. Cummins, 63–112. Oxford: Oxford University Press. Burman, Åsa. 2018. “A Critique of the Status Function Account of Human Rights.” Philosophy of the Social Sciences 48, no. 5: 463–473. Carling, Alan. 1993. “Analytical Marxism and Historical Materialism: The Debate on Social Evolution.” Science & Society 57, no. 1: 31–65. Cohen, G. A. 1978. Karl Marx’s Theory of History: A Defence. Princeton, NJ: Princeton University Press. Cummins, Robert. 1975. “Functional Analysis.” Journal of Philosophy 72: 741–765. Daniels, M. J. 1952. “Latent and Manifest Function of the Theory and Research of Bronislaw Malinowski.” Journal of the Arkansas Academy of Science 5, no. 1: 143–148. Davies, Paul. 2000. “Malfunctions.” Biology and Philosophy 15: 19–38. Durkheim, Émile. 1893. De la division du travail social: Étude sur l’organisation des sociétés supérieures. Paris: Félix Alcan.

68  Heiner Koch Ehring, Douglas. 1986. “Accidental Functions.” Dialogue XXV: 291–302. Elster, Jon. 1979. Ulysses and the Sirens: Studies in Rationality and Irrationality. Cambridge: Cambridge University Press. ———. 1983. Explaining Technical Change: A Case Study in the Philosophy of Science. Oslo: Universitetsforlaget. Foot, Philippa. 2001. Natural Goodness. Oxford: Clarendon Press. Griffiths, Paul. 1993. “Functional Analysis and Proper Functions.” British Journal for the Philosophy of Science 44: 409–422. Guyau, Jean-Marie. 1884. Esquisse d’une morale sans obligation ni sanction. Paris: Félix Alcan. Hindriks, Frank. 2013. “Restructuring Searle’s Making the Social World.” Philosophy of the Social Sciences 43, no. 3: 373–389. Iorio, Marco. 2003. Karl Marx – Geschichte, Gesellschaft, Politik: eine Einund Weiterführung. Berlin: De Gruyter. Kitcher, Philip. 1993. “Function and Design.” Midwest Studies in Philosophy 18: 379–397. Marx, Karl. 1904. A Contribution to the Critique of Political Economy. Chicago: Charles H. Kerr & Company. Marx, Karl, and Friedrich Engels. 1969. Selected Works, vol. 1. Moscow: Progress Publishers. Maasen, Sabine, Everett Mendelsohn, and Peter Weingart, eds. 1995. Biology as Society, Society as Biology: Metaphors, vol. 18. Dordrecht: Springer Science & Business Media. McLaughlin, Peter. 2009. “Functions and Norms.” In Functions in Biological and Artificial Worlds: Comparative Philosophical Perspectives, edited by U. Krohs, and P. Kroes, 93–102. Cambridge, MA: MIT Press. Merton, Robert K. 1938. “Social Structure and Anomie.” American Sociological Review 3, no. 5: 672–682. ———. 1968. Social Theory and Social Structure. New York: Simon and Schuster. Millikan, Ruth Garrett. 1984. Language, Thought, and Other Biological Categories: New Foundations for Realism. Cambridge, MA: MIT Press. ———. 1989. “In Defense of Proper Functions.” Philosophy of Science 56: 288–302. Mossio, Matteo, Christian Saborido, and Alvaro Moreno. 2009. “An Organizational Account of Biological Functions.” The British Journal for the Philosophy of Science 60, no. 4: 813–841. Nolan, Paul. 2006. “Why GA Cohen Can’t Appeal to Charles Darwin to Help Him Defend Karl Marx (But Why Others Can).” Science & Society 70, no. 2: 155–179. Orru, Marco. 1983. “The Ethics of Anomie. Jean Marie Guyau and Émile Durkheim.” The British Journal of Sociology 34, no. 4: 499–518. Radcliffe-Brown, Alfred R. 1935. “On the Concept of Function in Social Science.” American Anthropologist 37, no. 3: 394–402. Saborido, Christian, Alvaro Moreno, Matteo González-Moreno, and J. C. H. Clemente. 2016. “Organizational Malfunctions and the Notions of Health and Disease.” In Naturalism in the Philosophy of Health, edited by Élodie Giroux, 101–120. Cham: Springer. Sanderson, Stephen. 1990. Social Evolutionism – A Critical History. Cambridge: Blackwell.

Social Dysfunctions  69 Schwartz, Peter. 2007. “Defining Dysfunction: Natural Selection, Design, and Drawing a Line.” Philosophy of Science 74, no. 3: 364–385. Schweitzer, Bertold. 2015. “From Malfunction to Mechanism.” Philosophia Scientiæ. Travaux d’histoire et de philosophie des sciences 19, no. 1: 21–34. Searle, John. 1995. The Construction of Social Reality. New York: Free Press. Sullivan-Bissett, Ema. 2017. “Malfunction Defended.” Synthese 194, no. 7: 2501–2522. van Riel, Raphael 2016. “Mental Disorder and the Indirect Construction of Social Facts.” Journal of Social Ontology 3, no. 1: 27–48. Wakefield, Jerome. 1992. “The Concept of Mental Disorder: On the Boundary between Biological Facts and Social Values.” American Psychologist 47: 373–388. Wray, K. Brad. 2002. “Social Selection, Agents’ Intentions, and Functional Explanation.” Analyse & Kritik 24, no. 1: 72–86. Wright, Erik. O. 2010. Envisioning Real Utopias. London: Verso.

4

In Search of the Missing Mechanism. Functional Explanation in Social Science Raphael van Riel

Introduction Functionalism, once a dominant view in sociology, social anthropology, and Marxist thought, has largely been abandoned as an approach to understanding social structure (see, however, Harold Kincaid’s contribution to this volume for an overview of current functionalist theorizing in the social sciences). Arguments against functionalism vary depending on the intended target. They range from criticisms of a focus on social stability (as in Parsons 1937, 1951) and of a more or less overt ethnocentrism (as in Spencer 1873) to worries raised in connection with metaphysical and epistemological commitments of functionalist theories of social reality. A worry of this latter type underlies the prominent missing mechanism argument against functionalism in the social sciences (henceforth MMA), which is due to Jon Elster (1982, 1983). Elster’s primary target is a particular form of functionalist view: sparked by Jerry Cohen’s reconstruction of Marxist functional explanation (1978), it aims at what Elster took to be a general trait of almost1 all functional explanation in the social sciences, namely, that a feedback loop between function bearer and functional effect is hypothesized without any evidence that such a feedback loop exists. To begin with, it will be helpful to take a quick look at an example of a potentially problematic functional explanation, an example which is also used by Elster (1982) to illustrate the problem. Elster comments on the alleged beneficial consequences of the Bonapartiste regime for the capitalist class. The Bonapartiste state is “anticapitalist” in that it does not grant the capitalists political power, and is in this respect disadvantageous for capitalism, while being, at the same time, beneficial in that it preserves capitalist economic structure in a context where this structure is in fact in need of support. The existence of the regime is then explained in terms of these beneficial effects. The problem with this sort of explanation is that it seems to commit the theory to some mechanism at the individual level, which underlies the feedback loop connecting the functional effect of the Bonapartiste

In Search for Missing Mechanisms  71 regime (support of capitalist economic system) with its existence. Since the function cannot be traced to an intention (it is not the case that anybody set up the Bonapartiste regime in order to promote long-term capitalist interests), it must be some other sort of mechanism, like, for instance, a mechanism of selection. It is, however, unclear what this mechanism might consist in. And without evidence that such a mechanism exists, we should refrain from accepting an explanation of this sort. Yet, not all functional explanation is problematic in this respect. Here is another example: The fact that doing so maximized expected returns explains why the company reduced its labour force. (Jackson and Pettit 1992, 101) This explanation may be true because it seems that people may have decided to reduce labor force because they anticipated that it would maximize returns, and they intended to achieve this end by reducing labor force. The function can then be traced to an intention, which causes the functional effect. The gist of the antifunctionalist argument I will be concerned with is this: unless we can trace an alleged function to an intention, there is (apart from very few exceptions) no reason to accept functional explanations of social facts; and the main question I will pursue here is whether this line of reasoning, properly understood, generalizes to all sorts of functional explanation in the social sciences. It will turn out that several types of functional explanation once dominant in the social sciences, and, in particular, in social anthropology, remain unaffected. Elster (1990) was well aware of at least some of the limitations of his argument. Based on a distinction he drew between functional explanation and allegedly nonexplanatory functional description, or functional analysis, Elster maintains that MMA only bears on the former, whereas the latter is perfectly innocent, as long as one does not feel tempted to derive explanatory claims from mere functional description or analysis. It is this latter claim with which I take issue here. The point of this chapter is, thus, not so much that critics of functional explanation have to be persuaded that the type of functionalism endorsed in social anthropology and some parts of sociology is not affected by MMA; rather, they have to be persuaded that it is explanatory, without falling prey to MMA. I will introduce several types of functional explanation, only one of which is a proper target for MMA. In the section “MMA and Non-Intentionalist Functional Explanation” I will briefly introduce a general characterization of a functional explanation and comment on MMA. I will then turn to the distinction between descriptive and explanatory functionalism, as it is introduced by Elster and Harold Kincaid (“Two Types of Functionalism”). In “The Missing

72  Raphael van Riel Mechanism Argument Proves Too Much”, I will argue that MMA has limited scope because there are various types of explanation which have been offered in connection with so-called descriptive functionalism, and which do not conform to the type of functional explanation MMA targets. Moreover, there are some cases of functional explanations which do conform to this type of explanation, but which remain unaffected by MMA as well. I will conclude with a general hypothesis about the conditions a functional explanation has to meet so that MMA becomes relevant as a counterargument. In a nutshell, these include identity of the function bearer and the object whose existence is explained in the functional explanation (as is officially recognized by MMA) and absence of an innocent causal feedback loop. Before we start, let me add one comment to place these considerations within the broader philosophical literature on functionalism and functional explanation. This literature has focused on various related though ultimately different questions: –

– –

what a correct analysis of ‘in order to’ or related function ascriptions amounts to, in the spirit of conceptual analysis or descriptive metaphysics (see, for a discussion, McLaughlin 2001), often with an eye on the alleged goal directedness of function talk, whether any such analysis could be achieved, and, if not, what a fruitful alternative, potentially perspectivalist (Craver 2013) explication might look like (Millikan 1984, 1989), or whether paradigmatic functional explanations fit some general model of explanation (Nagel 1956, 1977; Hempel 1959).

Different conceptions of functional explanation have been proposed, inspired by both the aim to successfully reconstruct scientific discourse and to do justice to linguistic intuitions (von Wright 1963, 1971; Wright 1973; Cummins 1975; Neander 1988, 1991; Griffiths 1993). The question I will pick up here is independent of a decision on these issues. The types of explanation I will be concerned with in what follows do not rest on an irreducible notion of goal-directedness, the question of whether they are affected by MMA can be answered independently of the question of whether they offer an analysis of ordinary or scientific ‘in order to’ ascriptions (or related ascriptions), and we need not commit to a general theory of explanation in order to settle the question of whether these explanations are really explanations. Luckily, we are at least sometimes able to recognize an explanation without having a full-blown theory of explanation in the background (otherwise, it would be hard to see how we could come up with any such theory in the first place). For instance, sentences whose main operator is an explanatory connective, such as ‘because,’ ‘in order to,’ ‘due to,’ or ‘in virtue of,’ express explanations. And, similarly, true answers to ‘why-’ and many2 ‘how-’-questions

In Search for Missing Mechanisms  73 constitute explanations. These may paradigmatically represent causal or mechanistic dependencies. So, to rephrase my main claim: it is these things many functionalists came up with, based on functional analyses, without thereby committing, in any way, to some mechanism of selection or intentionality. When directed against these forms of functional explanation, MMA fails. I will, later on, turn to one distinction familiar from the philosophical debate – the distinction between proper functions (Millikan 1984) and Cummins functions (Cummins 1975). If you are familiar with the distinction, you may find the following sketch of my argument helpful: it will turn out that MMA targets proper function ascriptions only, although a dominant part of functionalism in the social sciences is concerned with explanations which are based on functions that are, to some extent, similar to Cummins functions. Elster and others have failed to see that the latter can be explanatory without requiring a mechanism of selection in the background, even if these functions come with a causal feedback loop.

MMA and Non-Intentionalist Functional Explanation In this section, I will introduce the concept of a functional explanation, as it underlies MMA, and introduce a general version of this argument. (Later on, I will spend much ink on explanations which are not functional in the sense defined here, although they are tied to functionalism, before turning back to functional explanations which accord with the definition.) At a high level of abstraction, the general idea behind a functional explanation is described by Cohen, the first target of Elster’s critique. According to Cohen, functional explanations are “explanations … in which … consequences are used to explain causes” (1982, 483).3 Elster gives a more thorough characterization: X is explained by its function Y for group Z iff (1) Y is an effect of X; (2) Y is beneficial for Z; (3) Y is unintended by the actors producing X; (4) Y – or at least the causal relation between X and Y – is unrecognized by the actors in Z; (5) Y maintains X by a causal feedback loop passing through Z. (Elster 1983, 57) Intuitively, this definition captures the explanation of the existence of the Bonapartiste regime, but not that of the reduction of labor force. We have thus here a notion where the source of the problem of paradigmatic functional explanation in the social science, namely: absence of an intention, is built into the definition – or, perhaps, supposed to be built into the definition. McLaughlin, who also aims at an analysis of function

74  Raphael van Riel talk which is independent of intentionality (2001, 42), has pointed out that conditions (3) and (4) are misleading because, first, the assumption that people intend the desired effects is perfectly compatible with the idea that these intentions do not play any role at all in the causal feedback loop. Only the latter is relevant, whereas only the former is ruled out by (3) and (4). Moreover, (3) and (4) do not rule out that an external source could have induced the feedback loop, thereby making functionality dependent on intentionality. Thus, (3) and (4) should be replaced. I suggest we replace the condition that the causal effect has to be beneficial for some group. Of course, we do find reference to some group as the beneficiary in the Marxist example discussed above, but, as will turn out later on, the definition thereby sets up a restriction which does not do justice to how functional explanation is construed in the social sciences – it is not necessarily a group of people with respect to which some phenomenon is said to be functional. It may also be some social structure, institution, or composite whole. Finally, I suggest we drop use of the term ‘benefit.’ Hempel (1959) suggested that a benefit patently consists in support of continued existence. Replacing reference to a benefit by reference to support of continued existence, we get rid of normative or teleological overtones, which are clearly not intended by the explanations we will be concerned with.4 Slightly adjusting the definiendum, so that we now define the predicate ‘_is a social functional explanation,’ we arrive at: [Functional Explanation] p is a social functional explanation iff there is some social object x, some social object y, some group z, 5 such that i p is an explanation of x’s existence in terms of x’s causal effect y

Based on this definition, we can hint at where the mechanism is supposed to come into play: it concerns the causal feedback loop mentioned in condition (ii). Condition (iii) rules out that intentionality sets up this feedback loop. So, if there is something to MMA, it will concern a missing account of the causal feedback loop by virtue of which the bearer of the function continues to exist. Elster describes MMA as follows: it rests on the assumptions that … a functional explanation can succeed only if there are reasons for believing in a feedback loop from the consequence to the phenomenon to be explained. Secondly, these reasons can only be the exhibition of a specific feedback mechanism in each particular case.

In Search for Missing Mechanisms  75 The second premise is not needed in the case of functional explanation in biology, for here we have general knowledge – the theory of evolution through natural selection – that ensures the existence of some feedback mechanism, even though in a given case we may be unable to exhibit it. But there is no social-science analogue to the theory of evolution, and therefore social scientists are constrained to show, in each particular case, how the feedback operates. (Elster 1983, 61) Only if a mechanism in virtue of which the feedback loop occurs can be identified, can an explanation which makes reference to such a feedback loop succeed. Two types of feedback mechanism seem available: intentionality and a mechanism of selection. In this sense, all those who fail to identify a feedback loop and assume that the feedback loop is nonintentional are committed to some kind of selective feedback mechanism. We cannot, however, just help ourselves to the assumption that such a mechanism exists.7 As a consequence, the explanations social functionalists hope to be able to come up with fail, and systematically so.8 Let us thus turn to those limits of MMA that are officially recognized in the debate.

Two Types of Functionalism In his survey and defense of functionalism in the social sciences, Harold Kincaid offers a helpful, tentative characterization of an important distinction between two types of functionalism, which he says is overlooked all too easily: [F]unctionalism vacillates, often unconsciously, between identifying functions in the sense of describing social structure vs. explaining social practices as existing because of their effects. (Kincaid 2007, 221) Later in his paper, Kincaid suggests that the distinction may not be as neat as it may seem at first sight, and context suggests that he means that we cannot use this distinction to neatly distribute individual scientists among these two camps of functionalism. Some try to do both, describe social structure and, in addition, explain social phenomena. Yet, these attempts, presupposing law-based explanations, are bound to fail. The only viable alternatives are, according to him as well as Elster (see above) and Little (1998), mechanisms of selection (Kincaid 2007, 244). A similar thought seems to underlie Elster’s distinction between a functional description and a functional explanation in his criticism of Robert K. Merton’s work on manifest (roughly: recognized or intended) and latent (roughly: unrecognized and unintended) functions (Merton 1957). Here, Elster argues that identifying a “beneficial” effect

76  Raphael van Riel (supporting continued existence) of a social practice within a social context is one thing, explaining the presence of the practice in terms of its effects is quite another: … Merton never states in so many words whether the task of functional analysis is to explain social phenomena or, more modestly, to identify and describe phenomena. (Elster 1990, 130) This appears to resemble Kincaid’s description above. These passages suggest a distinction between explanatory functionalism, according to which the consequences of a phenomenon are invoked in the explanation of the presence of this phenomenon, and descriptive functionalism, which is said to aim at a description of a system in terms of how the component parts of the system interact or contribute to some overall feature of the system. These two characterizations describe their targets in different respects – the notion of explanatory functionalism is described in terms of a particular claim about explanatory relationships which follows from this view, whereas the notion of descriptive functionalism is defined in terms of a goal of an approach, or a research program. As such, there is no logical reason to believe that the two are necessarily distinct. Yet, I believe that this does justice to how the two are described in the literature. We merely learn that descriptive functionalists (Elster) or those who want to functionally analyze a system (Kincaid) aim at a theory of a particular type, whereas explanatory functionalists are committed to the truth of a particular type of explanation, or perhaps even to some generic statement about how the presence of practices and institutions can patently be explained (they are committed to, but typically do not explicitly endorse this claim). Moreover, it is tentatively suggested, by the choice of the terms ‘descriptive’ vs. ‘explanatory’ and Elster’s claim that descriptive functionalism is more modest than explanatory functionalism, unless the two are conflated, that descriptive functionalism, properly construed, is not concerned with explanations. That this is in fact Elster’s view comes out nicely in his discussion of Merton, where he comes up with two readings of Merton’s use of the notion of a latent function – they are either a tool for description and, thus, primarily psychologically or intellectually satisfying, or for explanation. As one would expect, Elster dismisses Merton’s explanatory aspirations, claiming that he may have mistaken “the thrill of discovery for the thrill of explanation” (Elster 1990, 131), and that to the extent that Merton took latent functions to be explanatory, he was misguided. Kincaid is a bit more cautious: he contends that some descriptive functionalists believed to provide law-based explanations (which he rejects, see above), and that others may provide explanations in terms of the

In Search for Missing Mechanisms  77 causal contribution of a subsystem or trait to some other social unit, which he suggests is best conceived of as ultimately selectionist in nature (Kincaid 2007, 244). But how do we get from these descriptions to the claim that functional explanations fail due to the lack of a mechanism which would support an unintended feedback-mechanism? On another occasion, Kincaid employs a slightly different formulation: [Functionalism has been] ambivalent between two different claims […] namely, that society is composed of component parts with specific functions as opposed to the claim that there is some kind of selective mechanism ensuring that those social parts exist in order to serve their functions. (Kincaid 2007, 214f.) Here, the distinction is drawn in terms of composition of component parts on the one hand, and selective mechanisms on the other. As we have seen above, Elster invokes a similar idea when criticizing explanatory functionalism, for according to him, explanatory functionalism fails to the extent that it does because it fails to come up with an alternative to a selective mechanism which would undergird the functional explanation. On the assumption that this makes transparent what both Elster and Kincaid have in mind when they speak of explanatory functionalism, they appear to be committed to the following claim (here for reasons of simplicity restricted to the social sciences): [Selective Functionalism] All canonical explanatory functionalism in the social sciences is committed to some kind of selective functionalism (i.e. aims at explanations which can only be backed by selectionist mechanisms). That this is in fact something like the consensus view among those who take MMA to be successful is also indicated by Pettit’s (1996) defense of functional explanation as an explanation of the resilience of a trait (see also Jackson and Pettit 1992, 121f.). Here, the suggestion is not that functionalists did in fact provide resilience explanations, but, rather, that there are explanations which lie in the vicinity of those explanations functionalists came up with, which may, unlike the latter, be correct. We just have to change the explanandum: we do not explain why a trait persists, but, rather, why it would persist even when it came under pressure, i.e. why it is resilient.9 Since Pettit and Jackson seem to suggest that this is the best we can do about functional explanation in the social sciences, they seem to support something like [Selective Functionalism]. Based on this reconstruction, I can now restate the main thesis I wish to defend in the next section: I will argue that the sort of functionalism

78  Raphael van Riel characterized as “descriptive” by Elster and Kincaid typically aims at explanations, and that it may do so in an interesting way, without collapsing into selectionist or intentionalist functionalism. MMA proves too much if it is conjoined with the assumption of [Selective Functionalism], because [Selective Functionalism] is just false. Moreover, since many functionalists were concerned with descriptive functionalism, it will turn out that MMA, to the extent that it is successful, has fairly limited scope. There are two problems here. One is that not all explanations functionalists came up with are functional in the sense defined above, but were, nevertheless, explanations. Very often, functionalists simply did not aim at an explanation of why the bearer of the function itself exists. Instead, they offered an explanation in terms of causal relations between a function bearer and the object it is functional for. Thus, critics of functionalism overshot the target when assuming that there is no room for explanation other than the one defined above, within functionalist approaches. Moreover, they failed to recognize that even explanations which accord with their definition need not rest on dubious mechanistic presuppositions.

The Missing Mechanism Argument Proves Too Much Types of Functional Explanation The difference between explanatory and descriptive functionalism is more subtle than the characterizations offered by Kincaid and Elster suggest. To give you a first idea, consider an early informal though highly influential description of functionalist explanation which is due to Hempel: The kind of phenomenon that a functional analysis is invoked to explain is typically some recurrent activity … in an individual or a group, such as a physiological mechanism … a culture pattern or a social institution. And the principal objective of the analysis is to exhibit the contribution which the behavior pattern makes to the preservation or the development of the individual or the group in which it occurs. (Hempel 1959, 304) Here, Hempel maintains that functional analysis leads to explanation. And, since Hempel makes no mention here of a feedback loop which would tie the functional effect of the recurrent activity to the recurrent activity itself, it should become clear that there might be functional explanations (in some sense of ‘functional’) which are not covered by the definition proposed above (of course, this observation is not supposed to replace an argument – Hempel might be mistaken).

In Search for Missing Mechanisms  79 There are, I submit, several types of functional explanations worth distinguishing in connection with the functional explanation of social phenomena, only one of which is a potential target for MMA. Here, I will focus on three types of distinctions, all of which are based on actual examples (thus, I do not propose to offer a taxonomy of functional explanation based on a priori grounds). I will talk of types of functionalism in the very undemanding sense that acceptance of a functional explanation of a given type makes one a functionalist of this type. The first distinction, (i) persistence functionalism and (ii) etiological functionalism, concerns the explananda of a functional explanation. This distinction is nicely captured by social anthropologist Alfred Radcliffe-Brown, in the following passage: [O]ne ‘explanation’ of a social system will be its history, where we know it – the detailed account of how it came to be what it is and where it is. Another ‘explanation’ of the same system is obtained by showing (as the functionalist attempts to do) that it is a special exemplification of laws of social physiology or social functioning. (Radcliffe-Brown 1952, 186) Only one of these, the nonhistorical explanation, Radcliffe-Brown regarded as functional in the sense he intended. He thus rejects what I call here etiological functionalism. Persistence functionalism, however, is the view that a composed object (a “system,” in Radcliffe-Brown’s terminology) persists in virtue of its functional organization. Unlike etiological functionalism, persistence functionalism is not concerned with the explanation of the etiology of social phenomena. The next type of distinction concerns the “beneficiary” of the functional effect, i.e. patently the object whose existence is supported by the effect of the function bearer (the distinction is orthogonal to the distinction between persistence and etiological functionalism): (iii) composed object functionalism, (iv) constituent functionalism, and (v) external functionalism. Composed object functionalism is the view that components contribute to the existence of some whole they are a proper part of. Again, in Radcliffe-Brown’s (1952) terminology, who suggests it as an interpretation of Durkheim’s ‘besoins’ (‘needs,’ Radcliffe-Brown 1952, Chapter 9): a constituent has a function for the whole of which it is a part if it ensures that some of the necessary conditions of existence of the whole are met. Radcliffe-Brown’s composed object functionalism thus coincides with his persistence functionalism – the continued existence, rather than the etiology, of the composed object is explained. This view has been associated with demanding notions of social systems, which heavily rely on the idea that there is some degree of integration, functional harmony,

80  Raphael van Riel and absence of conflict (in particular, perhaps, in connection with the work of Durkheim (see, for instance, his 1912 work), Radcliffe-Brown and Parsons). Merton, who was highly skeptical of these assumptions, appears to aim at what I call here constituent functionalism, where some practices or institutions are said to contribute to the persistence of some other constituent of a society (often unrecognized and, thus, latent), while being possibly dysfunctional for the composed whole (Merton 1957).10 Using an example of Berger to illustrate the point: antigambling legislation may be functional for the creation of “an illegal empire for the gambling syndicates,” and, thus, dysfunctional for some system of which these syndicates are a part (Berger 1963, 40f). In Merton’s case, we seem to find an etiological perspective: antigambling legislation did promote the growth, if not the occurrence of an empire of gambling syndicates. As such, however, there is no reason to believe that latent functions may not also merely explain the persistence of some social phenomenon that was in place even before the function bearer occurred.11 Finally, Bronisław Malinowski’s later functionalism can be described as external in the sense that he, unlike Radcliffe-Brown, invokes universal and basic human needs social practices are said to serve, thus supporting the persistence or, perhaps, the thriving of human beings (1944). Now, prima facie, all of these types of functionalism hypothesize causal dependencies as well as corresponding explanations; and, as will become apparent, they are all independent of mechanisms of selection and intentionality. These become relevant only when a further condition is met: the function bearer is identical to the object whose existence we intend to explain (as per the definition of a functional explanation introduced above). This is the last distinction which I will draw here: the object whose existence we intend to explain is, or is not, identical to the function bearer. I will refer to this last distinction as (vi) identity vs. (vii) nonidentity functionalism. In some, but not all cases, identity functionalism may pose a problem – it introduces the feedback loop alluded to in MMA. Trivially, composed object-, constituent-, and external functionalism are all types of nonidentity functionalism. A Helpful Distinction: Cummins and Proper Functions At this point, a quick look at a distinction familiar from the philosophy of science might help. What is typically described as “descriptive” functionalism identifies something like Cummins functions (Cummins 1975), what is typically described as “explanatory” functionalism is concerned with something like Millikanian proper functions. Cummins functions come cheap. As Krohs puts it, Cummins’ approach “eliminates teleology by giving an account of function that is analytical and physicalistic” (2009, 70).

In Search for Missing Mechanisms  81 A helpful way to introduce the distinction (though not illuminate it) is to hint, as is commonly done, at a linguistic distinction between having a function and functioning as. Artifacts have functions, bestowed upon them by their creators or users. And some would argue that also biological traits may have functions, due to facts about a mechanism of selection. These are proper functions. But to have a function is not to be conflated with (actually) functioning as. As Millikan has repeatedly argued (1984, 1989), having a function is compatible with failure to function as. A knife may have the function to cut things, but may fail to function in that way, perhaps due to being rusty and blunt. The color patterns of the fur of an animal may have the function to enable the animal to hide from a predator (thus functioning as camouflage) but also fail to do so, due to changes in the environment, without losing their function. In contrast, intermolecular hydrogen bonds may function as (part of) the basis for the high boiling temperature of water, or the double helical structure of DNA, but they do not have this function in a way that would allow for malfunctioning of hydrogen bonds. Thus, there is a difference between two types of function talk. Yet, as Griffiths rightly observes (1993, 412), at least in the biological case, the functions of paradigmatic biological traits are Cummins functions with respect to survival and fitness (provided that they do not malfunction). Based on a notion of explanations which, in turn, are based on something like Cummins functions, we can first argue that composed object functionalism is explanatory – although explanations of this type do not, as such, constitute functional explanations as the term has been defined above. Here, no feedback loop is invoked, the object whose existence we explain is different from the function bearer. Composed Object Functionalism Is Explanatory The presence of a component may be explanatorily relevant for the persistence of some object of which it is a part (i.e. for which it is functional, on Radcliffe-Brown’s interpretation). Griffiths clearly indicates that these functional analyses are explanatory: “the fitness of a type of organism can be explained by Cumminsesque functional analysis” (Griffiths 1993, 412 [my emphasis]; and Cummins (1975) himself takes the analysis to be explanatory). Clearly, organisms of a type survive or are fit (if they are) because they exhibit particular traits; the presence of these traits in these animals explains (at least partly) survival and fitness. Consider a common fictional scenario,12 which reveals that we can have an explanation exclusively based on Cummins functions, without a mechanism of selection in the background. Heart valves in a spontaneously occurring swamp “human” being (lacking the paradigmatic phylogenetic history, but resembling a human being in all other respects) will be functional for that being, in this sense: they will contribute to its

82  Raphael van Riel persistence. Then, the being persists partly because its heart possesses heart valves. This is more than a description. And in fact, it does involve a mechanism which backs up this explanation – we can explain how the presence of heart valves contributes to the survival of the object it is a part of. Heart valves do so by ensuring that blood passes through these valves in one direction only. If these valves do not function in this particular way and allow blood to flow in both directions, the efficiency of blood flow is reduced (as in tricuspid insufficiency), which reduces chances of survival of the organism. This explanation is true independent of any theory about the evolution of heart valves. And it does indeed seem to be an explanation, even by the lights of Elster’s demanding account, according to which all causal explanation has to indicate some mechanism (Elster 1989, Chapter 1). This perfectly matches paradigmatic explanations provided in, for instance, cultural anthropology (on which Merton also draws when introducing his notion of a latent function, Merton 1957). Here, it is assumed that a society’s components perform some function which contributes to the continued existence of some composed social object they are a part of, and thereby have some “beneficial” effect. As Radcliffe-Brown put it (making explicit that he intends this to be a hypothesis): Every custom and belief of a primitive society plays some determinate part in the social life of the community, just as every organ of a living body plays some part in the general life of the organism. (Radcliffe-Brown 1964, 229) Merton (1957), following an idea of Radcliffe-Brown, claims that the Hopi rain dance contributes to the persistence of the social system in which it occurs. For this to be true, we merely need some kind of Cummins-style functional explanation. Radcliffe-Brown even hints at a mechanism at the psychological level, claiming “that the usages of a society work or ‘function’ only through their effects in the life, i.e. in the thoughts, sentiments and actions of individuals” (Radcliffe-Brown 1952, 185).13 I believe that explanations à la Durkheim and Parsons proceed in a similar spirit. These are instances of composed object functionalism – a causal feature of a trait partly explains how a feature of a composed whole (here: persistence of a system) occurs. This is widely acknowledged; what has been overlooked is that this kind of functionalism is explanatory. It is not, however, a functional explanation in the sense specified above: we here do not yet explain the presence of the function bearer itself (that is: the constituent part, such as heart valves, the rain dance ritual) in terms of its effects on the system it is a part of. We will turn back to this later on. Before that, we should briefly reflect on explanations corresponding to constituent and external functionalism.

In Search for Missing Mechanisms  83 Constituent and External Functionalism Are Explanatory Malinowski’s later functionalism differs from composed object functionalism in that it does not focus on the explanation of a feature of a system or social whole, but, rather, on explanations of how particular practices serve universal human needs (see, for a clear exposition, Mandelbaum 1969). I call this external functionalism, thereby highlighting what I take to be the main difference between the anthropological approaches of Radcliffe-Brown and Malinowski: something like a Cummins-style explanation is provided, not with respect to a feature of a composed object of which the function bearer is a part (and here, we deviate from Cummins), but, rather, with respect to some external effect of the practice – it is said to serve particular human needs. (Just as an aside: application of Elster’s idea that some group is the beneficiary of the function is most straightforwardly applicable to this case.) A third variant of explanation is defended by Merton. Merton is not so much interested in examining the contribution of particular causal features of social practices to the ‘necessary conditions of existence’ of a well-integrated social system, nor does he refer to basic human needs, but, rather, to other institutions or social practices in some context, thereby often highlighting that what appears to be functional for the persistence of an alleged system is in fact deeply problematic, for its latent functions differ from its manifest functions. In this spirit, constituent functionalism provides explanations in which a causal feature of a particular constituent partially explains why another constituent (a practice or institution) occurs or persists. We came across an example: the explanation of the occurrence of empires of gambling syndicates by reference to antigambling legislation. This is a simple causal explanation; but a causal explanation in which we have a cause of an occurrence of a trait or a cause of persistence of a trait. We have thus distinguished three kinds of functional explanation in terms of the relation between the function bearer and the entity which is served by that function – the composed object of which the function bearer is a part, an object external to that system (like basic and universal human needs), and another institution or practice, which does not (or does not have to) stand in some mereological relation to the function bearer. All of these explanations are independent of selection mechanisms or intentionality concerning the function; and they do not, as such, explain the presence of the function bearer, as the characterization of a functional explanation described above would require. Yet, they are explanatory, just like an explanation of the survival of an organism in terms of the contribution of its heart valves to its survival is explanatory. And just like this explanation does not, as such, account for the presence of heart valves, the explanation of the persistence of a social system, or of the

84  Raphael van Riel persistence of humans whose basic needs are satisfied, or of the presence of another practice or institution, does not explain why the function bearer exists. MMA does not apply to these types of explanations. Perhaps, one might want to complain that my whole argument is beside the point because these explanations are not functional ones. But this seems to be a terminological point. If use of the term “function” by Radcliffe-Brown, Merton, Hempel, and others was misleading, we may complain about that. But on what grounds? No one owns the term, and these theorists did not intend to capture some pretheoretic teleological notion, aiming at something like conceptual analysis. Thus, quarrel about the appropriate use of ‘function’ in this context should not amount to more than a terminological footnote. Feedback Loops and Identity Functionalism As I have just indicated, these types of functional explanation are not functional in the sense defined above – it is not the case that here, the existence of the function bearer is explained in terms of its effects. That is, they do not fall in category (vi), identity functionalism. As a consequence, there is simply no feedback loop presupposed by these explanations. In this final subsection, I will argue that there are in fact cases of identity functionalism, and, thus, types of functionalism which accord with the definition of a functional explanation, which remain unaffected by MMA.14 Radcliffe-Brown offers an idea of how identity functionalism may go together with persistence functionalism. Consider, first, Radcliffe-Brown’s idea that by examining complex patterns of social usages (as he labels the function bearers), we may arrive at social laws, i.e. patterns that can be found across all societies, and that this is, all by itself, explanatory: [A functionalist] explanation of [a social] system is obtained by showing … that it is a special exemplification of laws of social … functioning. (Radcliffe-Brown 1952, 186) For our present purpose, we may safely ignore the question of whether discovering such laws would be explanatory, if successful at all. Rather, a particular type of investigation Radcliffe-Brown seems to be committed to when searching for these laws is of interest. In order to support the idea that some alleged laws were really laws, Radcliffe-Brown had to show how these alleged laws were realized in different practices across different groups of people. For instance, he had (and tried) to explain how particular practices in different societies instantiate the institution of taboo (1952, chapter 7), or a type of institution which promotes social cohesion. These investigations amount to explanations, namely,

In Search for Missing Mechanisms  85 mechanistic explanations, in which we explain how particular cultural patterns, including roles, kinship relations, rites, etc. contribute to the persistence of a social composite object (what he took to be a social system). The hope was that once we have explanations of this type, and find similar explanations across different societies, we would be able to back up a hypothesized law – that, for instance, any social system requires institutions which increase social cohesion, regulate conflict, procreation, etc. This is admittedly a strong assumption (as is the complementary hypothesis that every social practice contributes to the persistence of the system of which it is a part, see above). Assume, for the moment, that Radcliffe-Brown is right: there are these laws which determine whether a social system continues to exist, depending on whether it contains the required institutions. Now, the social system would turn out to be a pattern of interlocking institutions or social usages. It seems that then, the existence of the whole contributes to the existence of its constituents. This does not commit us to circularity; rather, we have here commitment to an asynchronic relation of causal contribution. Some practices do persist because they fulfill a particular function for a composite social object. To illustrate the basic causal pattern, let me use a simplified sketch. Assume that the rain dance ceremony is repeated whenever two things come together: shortage of water and a pattern of beliefs among members of the group, which include belief in the authority of the Gods, a theory about how to influence the will of the Gods, willingness and perhaps feeling of duty to contribute to the ceremony, etc. Assume, moreover, that engaging in this practice increases social cohesion. Social cohesion, in turn, causally contributes to the absence of fundamental conflict, and, thereby, to the persistence of patterns of belief (basic principles of the ideology are not questioned, etc.), including those which trigger the willingness to take part in the rain dance ceremony when there is a shortage of water. Thus, there is a causal feedback loop, based solely on psychological mechanisms, where social object functionalism can be used to back up persistence functionalism in combination with identity functionalism. Interestingly, we do not need, for this type of explanation to be correct, the strong assumption that there are laws concerning the needs of a social system. All we need is some actual causal feedback loop. If past rain dance ceremonies, or, for that matter, regular after-work meetings, contribute to the continued existence of a set of norms of which expectations concerning future rain dance ceremonies or after-work meetings are a part, we can explain the persistence of these practices by reference to their “beneficial” effects on the system of norms. We do not need the additional assumption that the presence of some such practices is a necessary condition for the existence of social or normative systems as such (whatever these are supposed to be). Moreover, it is clear that here, the feedback loop need not be backed up by a mechanism of selection (it is, to me at least, totally unclear how reference to a mechanism of selection

86  Raphael van Riel could possibly relate to this kind of explanation, in an interesting way; of course, there may be other explanations where a selectionist story becomes relevant, say, concerning the spread of a particular pattern of beliefs in a particular environment, while, at the same time, other patterns of beliefs vanish). A quick look at an analogous case may help. Consider again heart valves. They do not only contribute to the persistence of the object of which they are a part; they also contribute to their own persistence by ensuring that the organism, upon which their existence causally depends, continues to exist.15 This is not mysterious at all. Often, organisms continue to exist for an extended period of time because heart valves ensure that blood flows in one direction only. Often (though not always) the long-term consequences of heart valves which let blood flow in both directions include death of the organism, which will result in decomposition (or burning) of heart valves. Thus, in these cases, it is true that heart valves contribute to their own persistence not due to some mechanism of selection from which their presence results, but simply because they perform a Cummins function for the survival of the organism, and do themselves causally depend on survival of the organism. This is a very simple causal story which backs up a causal explanation. There is a straightforward path from Cummins functions to a combination of persistence and identity functionalism, via composed object functionalism – as long as the persistence of the function bearer is sustained by the existence of the composed object it is functional for. There is a similar path from Malinowski’s external functionalism to persistence plus identity functionalism: by serving some basic human needs, social practices cause humans to survive and, perhaps, thrive; institutions that ensure hygiene (Malinowski 1944), for example, contribute to the persistence of groups of humans; once we have this practice in place, the persistence of the group may contribute to the persistence of particular practices which ensure hygiene, via some psychological mechanism, habit, or the like, or merely by contributing to the survival of the people involved. Again, there is a feedback loop, involving neither selection, nor intention. These explanations are functional in the sense defined above. Yet, they remain unaffected by MMA. In fact, we may hypothesize that MMA becomes relevant when either, we combine identity functionalism with etiological functionalism (as in the case of proper functions of biological traits, in Spencer’s functionalism and as in some genealogical accounts, which I have not touched upon here, cf. Matthieu Queloz’s contribution to this volume), or when we combine persistence functionalism with identity functionalism without there being any evidence that a mere causal feedback loop ties the object whose persistence is supported by the functional effect, to the function bearer (as, perhaps, per the assumption that the Bonapartiste régime continues to exist because it is beneficial for the capitalist class).

In Search for Missing Mechanisms  87

Conclusion One type of functionalism is committed to the presence of proper functions, and, thus, either to functional explanations in Elster’s sense, or to some analogue in terms of intentions of designers or users. MMA is officially directed against, and successful only when applied to this type of functionalism. Unfortunately, there was some collateral damage to the attack on this type of functionalism, not recognized as such by some of the attackers – a paradigmatic instance of a “guilt by association” fallacy. This is mainly due to the fact, as I have argued, that descriptive functionalism is typically mischaracterized as being nonexplanatory, and all nonintentionalist functional explanation is taken to be based on a mechanism of selection. Based on this misconception, it would follow that if MMA succeeds, the only version of functionalism that is not doomed to fail is nonexplanatory or merely descriptive in nature. In general, apart from some Marxist and, perhaps, Hegelian (I am told) and Spencer-style theories of society, there are very few, if any versions of functionalism, which defend explanations that presuppose (either implicitly or explicitly) a problematic type of identity functionalism as a general view on social institutions or practices. Most functionalists, like Durkheim, Malinowski, Radcliffe-Brown, and Merton, and, as I believe but have not argued here, Parsons, have pursued research programs which are, at least by the lights of MMA, innocent, even when they claim to provide functional explanations of a particular sort, which may or may not be functional in the sense defined above. Let me close with a conjecture concerning the source of the confusion. Perhaps, intentional and selective mechanisms are the only candidates which could possibly render true teleological claims in ordinary discourse.16 If we assume, in addition, that all functional explanation lies in the vicinity of teleological explanation in ordinary discourse, the general conclusion that functionalism in the social sciences is either nonexplanatory or (typically) mistaken would make perfect sense. Then, the confusion stems from the false assumption that what goes by the name ‘functionalism’ in the social sciences is typically concerned with function ascriptions which have teleological overtones, in this demanding sense.

Acknowledgments I am grateful to Dan Little, the members of the DFG-Network “Social Functions in Philosophy”, and the group in my research seminar at the University Duisburg-Essen for comments on a draft of this chapter. I would also like to thank Brian Epstein and Hans-Bernhard Schmid for the discussion following a presentation of this paper, and Kai Ploemacher and Christoph Thies for their support during the editing process. Finally, I would like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

88  Raphael van Riel

Notes







Note that this definition defines a concept which appears to be an instance of the first two conditions mentioned by Kincaid (2007, 222). Roughly, these conditions ensure the presence of a feedback loop. It is, thus, also similar to Wright’s definition (1973) – to a limited extent. It should be noted, as Nagel pointed out, that Wright’s claim seems to be that the existence of hearts in general (that is, the existence of that type of organ) is explained by the fact that organisms with hearts that circulate the blood are more likely to survive and reproduce than organisms with hearts that do not do so. (1977, 283 [italics added])

In Search for Missing Mechanisms  89 I am not sure what to make of Kincaid’s third condition, according to which the function bearer must be prior to the “beneficiary,” which is explained as the latter causing the persistence of the former only if the function bearer causes the beneficiary.



the demonstration that a phenomenon has unintended, unperceived and beneficial consequences seems to bestow some kind of meaning on it, and since to bestow meaning is to explain, the sociologist tends to assume that his job is over when the first four criteria are shown to be satisfied. (1983, 58–59)



12 A similar case is discussed, for instance, in McLaughlin (2001), with a slightly different goal; see also endnote 15.

90  Raphael van Riel the background, but need not contain counterparts to inheritance and replicators. If so, what I argue here differs substantially from Kincaid’s view – I will argue that no selection mechanism whatsoever has to be alluded to. Perhaps, the difference stems from Kincaid’s general dismissal of “organicist” thinking in the social sciences; here, I ignore problems which stem from organicist assumptions in the work of Radcliffe-Brown and others.



References Berger, Peter. 1963. Invitation to Sociology. A Humanistic Perspective. Garden City, NY: Doubleday & Company. Cohen, G. A. 1978. Karl Marx’s Theory of History: A Defense. Oxford: Oxford University Press. ———. 1982. “Reply to Elster on ‘Marxism, Functionalism, and Game Theory.’” Theory and Society 11, no. 4: 483–495. Craver, Carl. 2013. “Functions and Mechanisms: A Perspectivalist View.” In: Functions: Selection and Mechanisms, edited by Philippe Huneman, 135–158. Dordrecht: Springer. Cummins, Robert. 1975: “Functional Analysis.” Journal of Philosophy 72: 741–764. Durkheim, Èmile. 1912: Les formes elementaires de la vie religieuse. Paris: Presses Universitaires de France. Elster, Jon. 1982. “Marxism, Functionalism and Game Theory. The Case for Methodological Individualism.” Theory and Society 11: 453–482.

In Search for Missing Mechanisms  91 ———. 1983. Explaining Technical Change: A Case Study in the Philosophy of Science. Cambridge: Cambridge University Press. ———. 1989. Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press. ———. 1990. “Merton’s Functionalism and the Unintended Consequences of Action.” In: Robert Merton: Consensus and Controversy, edited by Jon Clark, Celia Modgil and Sohan Modgil, 129–135. London; New York: Falmer Press. ———. 2015. Explaining Social Behavior: More Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press. Giddens, Anthony. 1984. The Constitution of Society. Outline of a Theory of Structuration. Berkeley: University of California Press. Griffiths, Paul. 1993. “Functional Analysis and Proper Functions.” British Journal for the Philosophy of Science 44: 409–422. Hempel, Carl G. 1959. “The Logic of Functional Analysis.” In Symposium on Sociological Theory, edited by Llewellyn Gross, 271–307. New York: Harper and Row. Jackson, Frank, and Philip Pettit. 1992. “Structural Explanation in Social Theory.” In Reducion, Explanation, and Realism, edited by K. Lennon and D. Charles, 97–131. Oxford: Oxford University Press. Kincaid, Harold. 2007. “Functional Explanation.” In Handbook of the Philosophy of Science. Volume 15: Philosophy of Anthropology and Sociology, edited by Stephen Turner and Mark Risjord, 213–247. Dordrecht: Elsevier. Krohs, Ulrich. 2009. “Functions as Based on a Concept of General Design.” Synthese 166: 69–89. Little, Daniel. 1998. Microfoundations, Method and Causation. New Brunswick, NJ: Transaction Publishers. Malinowski, Bronislav. 1944. A Scientific Theory of Culture. Chapel Hill: The University of North Carolina Press. Mandelbaum, Maurice. 1969. “Functionalism in Social Anthropology.” In Philosophy, Science, and Method. Essays in Honor of Ernest Nagel, edited by Sidney Morgenbesser, Patrick Suppes, and Morton White, 306–332. New York: St. Martins Press. Merton, Robert K. 1957. Social Theory and Social Structure. Glencoe, IL: The Free Press. Millikan, Ruth. 1984. Language, Thought, and Other Biological Categories: New Foundations for Realism. Cambridge, MA: MIT Press. ———. 1989. “In Defense of Proper Functions.” Philosophy of Science 56, no. 2: 288–302. McLaughlin, Peter. 2001. What Functions Explain: Functional Explanation and Self-Reproducing Systems. Cambridge: Cambridge University Press. Nagel, Ernest. 1956. “A Formalization of Functionalism.” In Logic without Metaphysics and Other Essays in the Philosophy of Science, edited by Ernest Nagel, 247–287. Glencoe, IL: Free Press. ———. 1977. “Teleology Revisited.” Journal of Philosophy 74: 261–301. Neander, Karen. 1988. “Discussion: What Does Natural Selection Explain? Correction to Sober.” Philosophy of Science 55: 422–426. ———. 1991. “The Teleological Notion of ‘Function.’” Australasian Journal of Philosophy 69: 454–468.

92  Raphael van Riel Parsons, Talcott. 1937. The Structure of Social Action. New York: McGraw Hill. ———. 1951: The Social System. London: Routledge. Pettit, Philip. 1996. “Functional Explanation and Virtual Selection.” The British Journal for the Philosophy of Science 47, no. 2: 291–302. Radcliffe-Brown, Alfred. R. 1952. Structure and Function in Primitive Society. Glencoe, IL: The Free Press. ———. 1964. The Andaman Islanders. Glencoe, IL: The Free Press. (Originally published: 1922). Spencer, Herbert. 1873. The Study of Sociology. New York: D. Appleton and Company. von Wright, Georg. H. 1963. The Varieties of Goodness. London: Routledge. ———. 1971. Explanation and Understanding. London: Routledge. Wright, Larry. 1973. “Functions”, Philosophical Review 82: 139–168.

5

From Natural Hierarchy Signals to Social NormEnforcers. What Good Are Functional Explanations of Shame and Pride? Rebekka Hufendiek

Introduction Shame and pride are interesting examples of cognitive and behavioral features that have invited a variety of functional explanations in recent research. They are widely considered to be complex emotional responses that only occur in social beings, but that might be all there is in terms of consensus about them. Within the psychological and anthropological literature, there are different functional explanations of shame and pride to be found, ranging from claims about adaptive biological functions of regulating dominance and subordination within a hierarchically structured group via embodied signals, to the specific social functions that shame and pride can come to play in the stabilization of social norms, e.g., via prestige or public shaming. I will show in this chapter that functional explanations of pride and shame, as they are offered by psychologists and anthropologists, entail holistic assumptions about what kinds of beings we are, why we do what we do, what is beneficial for us, and how we structure our social surroundings.1 I take it that this is a typical – if not u navoidable  – feature of functional explanations of cognitive and behavioral features. The psychologist Jessica Tracy, for example, in her recent book sums up a discussion about the adaptive function of pride thus: “We’ve now seen how pride makes people who they are: how it motivates them to achieve and be moral, or aggressive and demanding, and to seek power, status, and climb the social ladder” (Tracy 2016, 143). Tracy’s theory of the function of pride includes the claim that pride is an evolved fundamental part of the self, that it can be felt and expressed as authentic pride or hubristic pride, where authentic pride motivates us to “achieve and be moral” (examples are Bill Gates and Barack Obama) and hubristic pride motivates us to be aggressive and dominant (examples are Lance Armstrong and Donald Trump). Both facets of pride function to increase our social status within a social hierarchy, such that hierarchical organization is taken to be an essential part of human social life.

94  Rebekka Hufendiek Where such claims are presented in the form of functional explanations, we can reasonably expect them to be bolstered by evidence pointing to an underlying causal mechanism. On closer look, however, there is an obvious tension between the holistic aim of the theory and the evidence that is actually available. Where claims about the function of emotions such as pride and shame are offered as if they were bolstered by solid evidence for the kind of causal mechanism needed for a functional explanation, they tend to make overgeneralizing and reifying claims about the human condition, giving such claims an ideological flavor. I suggest that the functional explanations of pride and shame (and presumably of cognitive and behavioral features in general) are eclectic explanations that entail elements of rational and interpretive modeling. Accordingly, oftentimes we should not expect functional explanations to result in successful prediction or manipulation (or other typical criteria for measuring the quality of causal explanations). Claims about the human condition in the background of functional explanations should not be understood as causal claims about our very essence and what we can and cannot change about it. Rather, the quality of explanations involving causal, rational, and interpretive elements depends on criteria such as coherence, increasing self-understanding, and inspiration of future research. The plan of this chapter is as follows. In a first step, I will give an overview of current empirical work on the functions of pride and shame. Second, I will distinguish etiological and causal-role explanations as the two main kinds of explanations offered within research on pride and shame, and I will argue that, while these are both seen as types of causal explanations, functional explanations do in fact often entail elements of rational and interpretive modeling and where they do, they should be described as eclectic explanations. In the third and fourth sections, I look at the scientific literature and argue that claims about the functions of pride and shame fail to offer enough evidence for both of the mechanisms that are supposed to underlie etiological and causal-role functional explanations. Given this finding, the final section reevaluates functional explanations of shame and pride to show that they often end up on a slippery slope towards ideological claims that overgeneralize and reify very particular empirical findings into general claims about the human condition. I suggest that authors should make explicit the eclectic nature of their explanations and point out which parts are the result of rational and interpretive, rather than causal, modeling. That avoids the reification and overgeneralization that often gives current functional explanations of shame and pride an ideological spin. Instead, it invites readers to use the theories on offer as contributions to our self-understanding and inspirations for future research.

Functional Explanations of Shame and Pride  95

Empirical Research on the Function of Shame and Pride There is an entire spectrum of claims about the possible functions of shame and pride to be found in the current psychological and anthropological literature. Shame and pride are complex emotional responses that according to theories prominent in psychology and anthropology are adaptations with roots in evolution that play a role in signaling high and low statuses within a group and thereby help to regulate social hierarchies (Fessler 2004, 2007; Tracy and Robins 2007b; Tracy, Shariff, and Cheng 2010; Tracy et al. 2013; Steckler and Tracy 2014; Beall and Tracy in press). This view rests on a growing amount of research suggesting that pride and shame occur cross-culturally and have preforms in nonhuman animals. When experiencing shame, people tend to lower their faces, avoid the gaze of others, slump their shoulders, and adopt a stooped posture and bent-knee gait. Pride involves the inverse pattern of behavior, namely an elevated face, direct gaze, squared shoulders, erect posture, and stiff-legged gait. The functionality of the two responses is understood as an antithetical relation in which the core components of the two displays are employed by nonhuman primates (and other mammals) during the negotiation or affirmation of relative rank. 2 Accordingly, a first claim about shame and pride is that they have status regulating functions. Pride has the adaptive function of motivating us to seek high status and is assumed to be part of a suite of crossculturally present, evolved cognitive mechanisms for negotiating status relationships. The pride expression has the function of signaling high status, while the aversive shame-like emotion experienced by subordinate individuals is part of a motivational system that leads some actors to fight for higher rank and others to signal submission and appeasement. Status differences among individuals, as is further suggested, emerge in all known human societies, even fiercely egalitarian foraging societies, and these differences influence patterns of conflict, resource allocation, cultural transmission, and mating and often facilitate coordination on group tasks (Tracy et al. 2013; Tracy 2016). There are further claims, however, about how shame and pride can come to perform different functions in different social contexts. Status regulation via shame and pride, according to many, can come in different forms: shame and pride motivate status-oriented behavior that can be dominance-oriented or prestige-oriented. Examples of prestige-oriented pride and shame range from winning or losing sporting competitions to feeling pride at being employee of the year, or shame at being the worst drummer in the village (Fessler 2004, 2007). Some would have it that these are two facets of the same emotions that both have their roots in biological evolution (Tracy and Robins 2007a; Beall and Tracy in press). Others argue that pride and shame originally played dominance-oriented

96  Rebekka Hufendiek functions in regulating group hierarchies, but that these were largely replaced through a process of cultural evolution in which cooperative organization and the permanent exchange of information became more important. In this context, shame and pride acquired the new function of signaling prestige. A further claim is that shame and pride have a norm-enforcing function. Shame and pride, according to this view, are ways to self-monitor the evaluations of others and thereby conform to social norms (Scheff 1988). Shame in particular sanctions what is socially undesired and thereby plays an important role in social regulation and cohesion. Shame, in this context, is seen as a response with a heteronomous character: triggering shame in somebody who violated a norm does not address the person as an autonomous subject but rather as a member of a group with a deep desire for conformity or harmony. Understood in this vein, shame is usually contrasted with guilt (rather than with pride), where guilt is understood as a response that is closely connected to moral conscience and motivates comparative actions. A far-reaching suggestion in this context is that cultures can be divided into those that maintain parts of social order mainly through triggering shame and those that appeal mainly to guilt for the same purpose (Benedict 1946). That difference is often understood in the context of the distinction between individualistic and collectivistic cultures (Fessler 2004; Shweder et al. 2008). In line with this, there is also cross-cultural evidence that shame is a socializing emotion that parents instill in their children in response to norm-inconsistent behavior, in order to support the transmission and inculcation of social norms (Mesquita, Leersnyder, and Boiger 2016). Cross-cultural studies also suggest that in some cultures shame is generally a highly valued emotion, and therefore permanently encouraged in infants, while in other cultures shame tends to be discarded (Trommsdorff and Konrad 2003). Status regulation (including dominance and prestige signaling) and norm enforcing are rough categories under which the large majority of psychological and anthropological functional claims can be subsumed. These accounts differ not only with regard to the functions they ascribe to shame and pride but also with regard to the kind of functional explanations they offer. In the next section, I will introduce the distinction between etiological and causal-role functional explanations and argue that they need to be bolstered by different kinds of empirical evidence.

Functional Explanations Functional explanations in all shades have been traditionally understood as a subspecies of causal explanations (Cummins 1975; Little 1991; Neander 1991; Kincaid 2007; Craver 2013). Functional explanations, as we will see in detail below, are holistic in the sense that they have to place

Functional Explanations of Shame and Pride  97 their explanandum within an interactive system for which it is supposed to have beneficial effects. That makes it particularly difficult to offer empirical evidence for the assumed underlying mechanism. Two theoretical conceptions of functional explanation have been particularly prominent in philosophy of science, but also in the relevant research on pride and shame: the etiological theory and the causal-role theory. The etiological theory of functions claims that the proper function of an entity is what explains why it is there or why it has the form that it has (Wright 1973). Etiological theories highlight the fact that functions are primarily ascribed to kinds and only secondarily to single members of these kinds. History matters for the development of such functional properties: the function of an item is to do what items of that type were selected to do in the past (Millikan 1989; Neander 1991). According to causal-role theory, on the other hand, functions can be ascribed to parts or components of complex systems that realize certain activities (Cummins 1975; Craver 2007, 2013). Where the etiological theory assumes that a functional explanation has to single out the mechanism that led to the historical selection and persistence of a feature A, causal-role theory assumes that a functional explanation unpacks the causal mechanisms that enable A to do B in system S. 3 I take the etiological account and the causal-role functional account not to be mutually exclusive but rather to be asking different questions (Huneman 2013). The question an etiological account is set up to answer is “Why do we have pride?” where the answer is a causalhistorical reconstruction, while a causal-role functional account is trying to get an answer to the question “Why is pride useful for creatures like us?” where the answer is a description of what pride does in the relevant system. An etiological explanation entails a commitment to what the function bearer is and what it does. Minimally, it can be unpacked in terms of a causal relation (A causes B). But usually the causal claim is part of a more comprehensive mechanist explanation accounting for the conditions under which and the system within which A normally causes B. An etiological explanation further entails the claim that function bearer A persists because it causes B. Finally, an etiological explanation entails the claim that the effect, B, is in some sense beneficial to the system it belongs to. The general form of the etiological account can capture social and biological functions alike. It is, however, far more frequently applied to supposed biological features, since natural selection is widely accepted as a feedback mechanism that can cause the persistence of functional items. In the case of shame or pride, the relevant feature would be an inheritance system that resulted from evolutionary pressures due to natural selection at some point in human evolution or a cultural inheritance system that transmits certain forms of behavior over generations.

98  Rebekka Hufendiek For cultural or social items, it is sometimes generally doubted that there are such selection mechanisms at all (Elster 1982), and most of the suggested mechanisms (e.g. Sperber 2007) are controversial. It has often been remarked that an etiological explanation “seems to commit us to a rather mysterious causal process that is nonetheless seductive” (Kincaid 2007, 213). To demystify the relevant causal process, it needs to be unpacked as a certain kind of mechanism and it needs to be clear what kind of evidence speaks in favor of the actual presence of such a mechanism. All kinds of random causal processes that might have had an effect on why A persists need to be distinguished from the particular claim that it is A causing B that results in A’s persistence. In principle, various kinds of evidence might be advanced in favor of such a functional explanation. The evidence makes up a continuum running from direct evidence that the causal claims entailed by an etiological explanation hold to indirect plausibility arguments to infer that they do. Direct evidence shows that a function bearer had a certain causal effect and that having that effect caused it to persist. Decisive data on the actual historical process that led to persistence and on comparable scenarios are often not available, and scientists then must resort to more indirect and contentious sorts of evidence. That can take the form of data measuring the persistence of the relevant trait, along with data about the presence and absence of similar traits and their alleged effects in similar environments, which are sufficient to allow for causal modeling (Kincaid 2007, 217). There are many sources for indirect information in the case of human emotions like pride and shame, ranging from the study of primates (as a model for our ancestors) to cross-cultural comparisons (as proof that a feature is present cross-culturally) and game-theoretical models (as proof that certain features are likely to stabilize under certain conditions), but as indirect sources the degree of plausibility that they can offer depends on various contextual factors (Griffiths 1997). These contextual factors need to be addressed in the types of arguments frequently offered to bolster etiological explanations, which include the following: a.

Trait environment correlation arguments present evidence that a trait or a practice occurs in one environment but not in another. From that, it is inferred that the trait serves a social function in that particular environment where it does occur. The observation that a trait occurs in one environment but not in the other, however, does not give us any idea of how that feature was caused and certainly not that the trait persists because of its effects. Only combined with the observation of regularities under which the feature occurs can correlation arguments be considered strong arguments. b. Optimality arguments suggest that a trait persists because it was the best solution to a problem. Ecological anthropologists often argue that some foraging practices maximize food supply and infer that

Functional Explanations of Shame and Pride  99

c.

they exist in order to do so. Such arguments, however, are only as strong as the input that goes into them: (i) Do we have measurements of all the relevant costs and benefits of the practice? (ii) Do we have solid evidence about which practices are actually feasible? (iii) Can we exclude alternative explanations? Stability arguments are a common alternative to optimality arguments. Stability arguments claim that once a trait comes to dominate for whatever contingent reason, it is hard to replace. Is shame really an optimal solution to a fundamental problem of hierarchically structured group life? Could we not explain it just as well as a trait that under highly contingent circumstances became dominant and was never selected against? Taking optimality arguments for granted that cannot prove competing stability arguments wrong has been called Panglossian functionalism (Little 1991). I will come back to optimality and stability arguments when discussing the evidence for the different accounts of shame and pride below (see Kincaid (2007) for further discussion).

Much functionalism, with regard to biological and social phenomena (and also with regard to emotions such as shame and pride), is not concerned with the selection history of the trait in question but rather with identifying the beneficial causal role that an item plays within a system. Such an explanation entails a commitment to a function bearer A and spells out what A does within a complex system S. By doing that, an interpretation of how B, which is caused by A, can be beneficial for the entire system is offered. Different causal dispositionalist accounts have defended different views of how to account for causal relations, of how to explain that A normally causes B, and finally of the explanatory status of functional descriptions. At first glance, it might look as if the evidence needed to bolster a causal dispositionalist explanation is easier to get than the evidence for a selectionist explanation. Causal-role functional explanations, however, are very demanding explanations as well. Emotions like pride and shame, for example, are complex responses that are part of even more complex organisms. They span several levels of description from endocrinal and neuronal, to behavioral and motivational, and finally cognitive and evaluative. They seem to be biologically determined responses to a certain degree, but strongly influenced by the social context as well. To bring enough order into such a “causal stew,” to be able to make claims about a causal-role function, it needs to be defined what the relevant system is: Is it a biological organism? An individual? A group? A society? Furthermore, it needs to be defined what the relevant components are, and how they interact. Causal-role functionalists have highlighted that there are always different ways in which one can order the evidence for causal connections

100  Rebekka Hufendiek into different mechanisms, with different functions. Claims about functions and mechanisms therefore remain a perspectivalist endeavor, since there are always alternative ways of arranging the data (Craver 2013). The explanatory power of the resulting explanations can be measured, however, in terms of prediction and manipulation. Are we able to predict the responses that the described mechanism will display under certain conditions? Do we understand the causal relations well enough to be able to intervene and manipulate responses? The more successful we are in manipulating a mechanism the more likely it is that our explanation is correct. It is, however, a fact about current neuroscience (Craver’s main source for examples) and even more about current psychology and anthropology that mechanistic and particularly functional explanations are complex explanations that assemble data within a certain theoretical framework, where the circumstances are so difficult to control that experiments often cannot clearly show which of several explanations is more successful with regard to factors such as prediction and explanation. Perspectivalism appears to be a reasonable consequence. Perspectivalists hold that both etiological and causal-role functional explanations are causal explanations that demand a particular kind of empirical evidence to be plausible. That does not mean, however, that physiologists remain confused about whether the pancreas of a particular patient or an NMDA receptor is functioning well or not. It rather means that ascribing functions is the result of a complex explanatory process that involves not only discoveries of facts but also decisions about what is considered to be normal or abnormal: we divide a system into parts in part by deciding first what needs to be done in order for the mechanism as a whole to behave as it normally does … From the perspectivalist view adopted here, these judgments of normality continue upward until they are grounded ultimately in the judgment or interests of an observer. Functions, on this view, are roles within mechanisms, defined ultimately in terms of a topping-off point selected for its relevance to observer interests and perspectives. (Craver 2013, 141–142) Perspectivalism includes the idea that where functional explanations make claims about what is beneficial to a system or an organism, they involve assumptions that are related to the researchers’ interests. What I want to add to the perspectivalist view is partly an elaboration and partly a correction of what it means that judgments about normality are ultimately grounded in the interests of the observer. The interests of the observer are not just personal preferences but rather theoretical assumptions that need to be unpacked. I suggest that this means that functional explanations include more than just causal modeling, namely rational

Functional Explanations of Shame and Pride  101 and interpretive modeling. After all, functional explanations explain the presence of a feature in terms of its beneficial effects for a complex system. That cannot be done without some kind of holistic approach: the feature needs to be placed within the system, and we will need to say something about what the relevant system is and why the feature in question is beneficial for it. But what kinds of methods do we have to do that, particularly where causal modeling alone does not give us the answers we are looking for? One place to look is the social sciences, which have a long tradition in discussing what causal explanations can and cannot explain. Many authors have drawn on rational choice theory and interpretation to answer questions about the social world. We can ignore the difficult question whether and to what extend rational choice accounts or interpretations can be understood or be translated into causal explanations. The relevant point here is that rational and interpretive modeling do not rely on the same kind of evidence and do not seek to illustrate the same kinds of relations as do causal models. Starting from the claim that human beings are intentional creatures that act on the basis of reasons, rational choice theory describes the nature of rational decision making on the basis of a set of beliefs and goals and encompasses probability theory and game theory. Where we reconstruct the intentions and rational interests of agents within certain circumstances what we do is rational modeling. Variants of that theory have also been used to explain evolutionary development, particularly of functional features, assuming that natural selection is a causal process that can be to some extent rationally reconstructed. In philosophy we can find similar attempts in fictional states of nature or genealogies. Here, again, the aim is not to unveil a particular causal structure, but rather the function of a trait, where the ascription of a function at the same time offers an idea of what a trait does, why it might be there, why it makes sense to think of certain traits as more fundamental then the trait in question and of others as dependent on that trait and so on. We can come to understand why the development of a new feature for creatures with certain needs in certain environments was useful (Craig 2005, see also the contribution of Matthieu Queloz to this volume). I will refer to the whole bunch of methods that underlie this kind of reasoning as rational modeling. Rational modeling can ask for dependency relations between certain traits and assume, for instance, that humans (or other living beings) must have been social to some degree, before the evolution of traits like pride and shame could have been functional at all. Various kinds of rational modeling have been criticized (and rightly so) for being too abstract, highly conjectural, methodologically misguided, and oftentimes biased in their assumptions about what humans are like. In particular, the rational choice approach is controversial because it presupposes a particular model of what a rational agent is that has often

102  Rebekka Hufendiek been criticized as one-sided and inadequate (Dupré 2001). Harold Kincaid brings up a similar objection with regard to the use of game theory as a part of functional explanations: Those explanations describe preferences, strategies of individuals, individual endowments, and so on. Yet those explanations take as given those preferences, strategies, initial endowments, payoffs, amounts of interaction, etc. – in short, they presuppose as unexplained the rules of the game, institutions, social structure (e.g. who owns what), and so on. This means that even if these functional explanations succeed, they leave much unexplained. (Kincaid, this volume) I take it, however, that part of this critique rests on the supposition that we expect functional explanations to be either well-grounded causal explanations or else dubious as explanations. But this supposition constitutes a false dilemma, and I suggest viewing rational modeling as a valid part of functional explanations, where it is made explicit as such and where it is judged by its own standards. Rational modeling needs to be coherent with the causal evidence available, and it needs to ask for alternative explanations. The tendency for just-so stories and Panglossian functionalism can be avoided if rational modeling considers evidence regarding trait – environment, optimality, and stability arguments alike. Where such evidence is not (or only to some degree) available rational modeling has to carefully weigh different alternatives, looking for the most reasonable dependence relations and the most coherent approach. Moreover, oftentimes rational modeling can point to possible solutions and thereby inspire future research to verify or falsify these claims. A second reason for critique is that rational choice theory seems to abstract away a lot of what might be contingent or culture-specific about the behavior of human beings. Within the social sciences, it has been suggested that social inquiry should be concerned with the interpretation of meaningful social practices and that these can only be understood in their specific context. Interpretive modeling then is the attempt to understand particular social practices within a specific cultural context, where meaningful human actions are understood as being functional or dysfunctional. What we can expect from interpretive models is that they make sense of human behavior and thereby increase our self-understanding (Geertz 1972; Taylor 1985). I assume that interpretation and rational choice theory can in many cases offer explanations in the sense of answering why-questions in a scientifically respectable manner, particularly where we do not have a well backed up causal model (Little 1991). I will therefore speak about causal, rational, and interpretive modeling in the following and suggest

Functional Explanations of Shame and Pride  103 that all three can play a vital role in constructing functional explanations. We should simply adjust the criteria we use to measure the quality of these explanations. Where we might expect that a good causal explanation enables prediction or manipulation, that is not what we can reasonably demand from rational or interpretive modeling to the same extent. In the next two sections, I will show that current functional explanations of pride and shame cannot reasonably be seen as offering functional explanations understood as causal explanations either of the etiological or of the causal-role kind. I will also gesture towards parts of the relevant accounts that entail rational and interpretive modeling. In the final section, I will show how presenting one’s own approach as if it were backed up by causal evidence gives some of the recent work an ideological appearance, and I suggest that making the role of rational and interpretive modeling explicit can be a remedy against that.

The Evidence for Etiological Explanations of Pride and Shame Pride and shame are uncontroversially social responses in the sense that we only ascribe them to creatures with complex social abilities. While there is a controversy over whether we find prototypes of shame and pride in other primates, or whether they are human-specific reactions demanding certain cognitive capacities, such as having a notion of the self or the ability to evaluate oneself in relation to others, none of these accounts suggests ascribing shame or pride to creatures who do not happen to live and interact in a structured group with conspecifics. Whatever the functions of shame and pride are supposed to be (status regulation in terms of dominance or prestige signaling, or norm enforcement), their functionality depends on the presence of a social group, i.e., none of these responses would be functional in a nonsocial organism that did not need to interact with a structured group of conspecifics as a fundamental part of its existence. There is no description of what shame and pride are doing that does not include some sort of social group or social structure as the maintaining system. Still, there are etiological explanations that take pride and shame to be the results of biological selection that mainly help the survival of the individual organism within the social group. The currently most prominent etiological explanation for pride and shame suggests that they function to regulate social status (Tracy, Shariff, and Cheng 2010; Tracy et al. 2013; Steckler and Tracy 2014). An etiologist must define the function of pride and shame with reference to the history of selection. The most prominent defender of the claim that pride is a universal part of human nature with the adaptive function of regulating status, Jessica Tracy, assumes that emotions have social functions

104  Rebekka Hufendiek when they occur in social creatures to facilitate in-group interaction. Tracy and colleagues presuppose, however, that these emotions evolved through natural selection because facilitating in-group interaction increases the biological fitness of individuals. According to their account, pride and shame are more intermediate in their fitness contribution compared to emotions like disgust or fear because they serve immediate social goals that increase biological fitness in social creatures like us. Pride and shame can be described as emotions that evolved because they were functional for an individual who was a member of a hierarchically structured social group. Therefore, strictly speaking, this is a biological function insofar as it is part of an etiological explanation that assumes a biological history of selection for a disposition that is functional for an organism within a social group. This general claim about the functionality of emotions comes with certain related claims about the functionality of elements that help to realize pride and shame responses. The most prominent of these claims is that the pride expression is a particularly central element in realizing the function of pride: the expanded posture is assumed to attract attention while the accompanying smile is assumed to communicate “I’m dominant, but I’m still your friend!” (Tracy and Robins 2007a, 276). Furthermore, the functionality of the pride expression depends on the cognitive capacity of others to receive the relevant information about social status from seeing the emotions expressed. Tracy and colleagues suggest that social status is, to some extent, inferred from pride and shame expressions through an automatic unconscious cognitive process. This cognitive capacity is functional in its own right, since it informs judgments and social decisions (Shariff, Tracy, and Markusoff 2012). As we have seen, etiological explanations of pride and shame differ when it comes to the details. While Tracy and some of her colleagues hold that pride is an emotion with a hubristic and an authentic facet that motivates dominance- and prestige-oriented behavior but that both pride and shame have their origins in evolution, another explanation on offer includes cultural selection. On this view, emotional reactions were shaped during a long history of humans becoming responsive to certain kinds of norms, such as altruistic and cooperative norms, to such a degree that the original status functions of prototypes of shame and pride came to play a minor role. Thus, both the prestige-signaling function and norm-enforcer functions of emotions only evolved when humans acquired a theory of mind and created cooperative structures (Fessler 2004). The degree to which the status function still matters depends on the organization of the relevant society. Status functions will matter more in societies that are hierarchically organized, and prestige and norm-enforcer functions will matter more in egalitarian contexts and individualistic cultures. This explanation is an etiological explanation

Functional Explanations of Shame and Pride  105 that requires a social selection mechanism. Within a cooperative cultural context, a new motivational system evolved, one oriented not towards relationships between superiors and inferiors, but rather towards relationships among prospective cooperative partners. In a cooperative context, shame mainly functions to enhance conformity to cultural standards (Fessler 2007). Fessler assumes that what he calls the primordial hierarchy-oriented functions of shame and pride still play a role in human behavior. It largely depends on the organization of society which function comes to be dominant. I will suggest in the last section that the details in which the two approaches differ rely on several general assumptions about the human condition that appear to be largely a matter of interpretive and rational modeling and not causal modeling alone. How well are these etiological explanations backed up by evidence? As for the first claim about the biological etiological function of emotions in general, many current researchers doubt that there are universally occurring emotion types with adaptive functions far back in evolution (Barrett and Russell 2015; Barrett 2017; Crivelli and Fridlund 2019). Psychological constructivists assume that the bodily arousal, expressions, and behavioral tendencies involved in emotions are anything but fixed through a selection mechanism like a genetic inheritance system. Rather, within a process of experiences in a certain social setting, the brain comes to interpret certain kinds of bodily arousal as belonging to a lexical emotion category. This history of experiences and predictions by the brain about the body is what sets up emotion types within a social context (Barrett 2012; Barrett and Russell 2015). Both versions of the etiological account I introduced assume a complex evolutionarily acquired mechanism to be the normal cause of an emotion, and that claim is already highly controversial within current debates. Let us assume for the sake of the argument, however, that there were stable and uncontroversial evidence in favor of universal types of emotions. The selection mechanism (A persists because it does or enables B) is most relevant for a functional explanation. Suppose that there were relatively uncontroversial evidence that all human beings show emotional responses such as shame and pride and that we had good reasons to think that these responses are at least partly innate and hardwired and that they steadily do the same thing: motivating rank-related behavior. That makes it a reasonable assumption that there is some sort of selection mechanism responsible for the persistence of these emotions. But it does not offer a grain of evidence for the claim that the emotional response persists because of the relevant selection mechanism. The studies of Tracy and colleagues mainly offer evidence for the universality of pride expressions and suggest an interpretation of how B is beneficial to system S on that basis (Tracy and Robins 2007a; Tracy 2016; Beall and Tracy in press). Their interpretation apparently relies on an optimality argument: it assumes that pride and shame are optimal solutions to some

106  Rebekka Hufendiek kind of problem, namely peaceful cooperation within a hierarchically structured group. Even if they are right about the biological origin of the relevant emotions, we lack evidence as to which were the relevant and which the random causal processes involved in the development of shame and pride. We have no clue about the presence and absence of pride in different environments such that we could compare scenarios and come close to something like causal modeling. What are different kinds of appeasing cooperation within a hierarchy? What are the relevant costs and benefits of different practices? Which other practices would be feasible? How can we exclude alternative explanations relying on stability arguments such that pride and shame were never optimal responses but side effects of a highly dysfunctional hierarchical system that just happened to be not dysfunctional enough to lead to extinction? An account that seeks to avoid Panglossian functionalism and embraces rational and interpretive modeling could develop answers to these questions on the basis of what little data we have. The claim that the pride expression functions to communicate (friendly) dominance deserves special attention. Direct evidence for a feedback mechanism in the form of an evolved inheritance system is not available, so the evidence that the expression is universal needs to be combined with indirect evidence to allow for an inference to the best explanation that the pride expression has an adaptive function. Besides the above mentioned material on the universality of the expression, the main piece of evidence offered in favor of the claim that pride expressions evolved because they signal high status is evidence showing that human beings cross-culturally automatically and implicitly recognize pride expressions as indicators of high status (Shariff, Tracy, and Markusoff 2012). This is at the same time an argument for the claim, that there is an adaptive ability to recognize social status that informs social judgments. The evidence that Tracy and colleagues gathered suggests that people automatically take a pride expression to signal high status, even where the context (of, e.g., a presented picture) might suggest otherwise. These studies, however, will need to be replicated in many contexts and across more than two cultures to support the inference that that cognitive mechanism might indeed exist. Furthermore, that a cognitive mechanism is implicit and automatic does not imply that the mechanism is hardwired or adaptive. An implicit association can be learned and still trump contextual information, e.g., because we learn to interpret facial expressions from early on because they strongly matter in our environment, and so on. Again, the present evidence does not justify treating the functional description as a causal explanation. Very similar arguments can be raised against further studies that Tracy and colleagues have conducted. The anthropologist Daniel Fessler conducted comparative studies in Bengkulu (Indonesia) and California (US) that he relates to the

Functional Explanations of Shame and Pride  107 theoretical background assumptions that status-signaling shame is a homology that likely motivates appeasement displays in nonhuman primates and “plausibly constitutes the ancestral aspect of shame, while shame in contexts of defection, prestige failure, and nonconformity presumably constitute the derived, uniquely human aspects of this trait” (Fessler 2004, 249). Unsurprisingly, Fessler does not offer any direct evidence for the causal-historical evolution of the respective functions, nor does he offer the kind of data that could be compared in a way that would allow for causal modeling. Fessler’s work, however, exemplifies the central merits of interpretation and rational reconstruction. Fessler discusses which cognitive abilities need to be in place before certain forms of shame and pride can evolve, and he discusses the effects of different social practices and social structures on shame and pride, ranging from hierarchical vs. egalitarian organization, to cooperation vs. competition and autonomy vs. heteronomy. He thereby delivers an insightful analysis of the role that shame and pride can in principle come to play within different social settings and considers some to be more basic than others.

The Evidence for Causal Dispositional Explanations Causal dispositional explanations do not need to deny that emotions might have origins in evolution and that some emotional traits might have been adaptive. They can remain neutral on the selection history. A causal-role functional explanation does, however, presuppose a clarification of what the function bearer is, whether it occurs universally or only in certain circumstances, what it normally does, and in which relation it stands to the containing system. That description has to be such that it offers a ground for the assumption of a beneficial effect of the feature on the overall system. If we look into the empirical literature, most of the causal dispositionalist accounts on offer in psychology and anthropology assume that the functions of pride and shame have something to do with social status and social norm enforcement. Most of these accounts, however, do not spell out in detail what the function bearer and the system it is part of are supposed to be. The most developed framework with regard to the functional claims is the interpersonal approach. This approach claims that emotions can only be individuated and understood in terms of their functioning when considered in a social context of interpersonal and group interaction. Defenders tend to claim that no complex cognitive appraisals need to be in place for individuals to be able to show shame or pride within a social structure in which they make sense. Emotions signal needs and aims of social beings and thereby enable “social survival” – that is, the human capacity to build social bonds and address and overcome problems such as social exclusion or loss of power (Parkinson, Fischer, and Manstead

108  Rebekka Hufendiek 2005; Fischer and Manstead 2016).4 Within that framework, shame and pride can come to be functional on various levels. Shame (contrary to the status regulating account) is not explicitly related to subordination within a dominance relationship, but rather is seen as serving an affiliation function in acknowledging one’s own wrongdoing on the interpersonal as well as on the group level (the norm-enforcer function). The aversive nature of group-based emotions such as shame is also supposed to be functional in preventing group members from engaging in actions that might give rise to these emotions, since anticipated group-based shame under certain conditions results in less in-group favoritism and motivates group members to engage in collective action to stop the shame-evoking behavior (Shepherd, Spears, and Manstead 2013). Shame induction, according to the authors, is used in various cultures as a key mode of socialization that does not necessarily presuppose higher cognitive abilities: “Children are often made aware that they ought to be ashamed of themselves by a nonverbal presentation of contempt as well as by explicit verbalizations” (Parkinson, Fischer, and Manstead 2005, 193). The interpersonal theory takes it for granted that certain emotions are universally present and play different roles in different social contexts (Fischer and Manstead 2016). They therefore face the same objections as does the etiological approach. Psychological constructivists, on the other hand, assume that emotions as kinds are constructed by the brain in a learning process that takes place in a particular social environment. They embrace an ascriptive view with regard to functions, assuming that the functions of emotion types depend on acts of collective intentionality in which we ascribe functions to a bunch of bodily processes and feelings (Barrett 2012). If, again, for the sake of the argument, we grant the interpersonalist that pride and shame are universally reoccurring emotion types, there is another problem lurking: to be bolstered with severe evidence, such an account needs a comprehensive description of the relevant components and interacting levels within the individuals having the emotions and the different institutions imposing norms on them, standing in a complex relationship of mutual causation. This poses a variety of difficulties: interpersonalists take social practices to be a main force that shapes the nature of emotions, while psychological constructivists take emotions to be a result of neural guesses within a particular social context. That is just one of many levels where disagreement with regard to the relevant mechanisms can take place. Another concerns the cognitive complexity that shame and pride presuppose. While some suggest that simple forms of pride and shame can respond to perceptions of status within a social hierarchy (Clark 2010) or to perceptions of social rules and norms (Hufendiek 2015), others argue that self-evaluative emotions only occur once children have gained an explicit notion of the self and others and an explicit understanding of social norms (Lewis 2016).

Functional Explanations of Shame and Pride  109 In the context of the interpersonalist theory, it tends to be assumed that pride and shame can be present early, and that human infants can be normatively “shaped” from early on, e.g., by evoking shame through a contempt-expression (Fischer and Manstead 2016). It is also assumed that shame experiences play a role in the development of conscience and a reflective notion of the self in relation to others, while many psychologists assume that it is the other way around: an explicit notion of the self in relation to others and an understanding of social rules and norms must be present before infants show shame, pride, or guilt in any interesting sense (Lewis 2016). Such background assumptions have different consequences with regard to the assumed causal relations within shame and guilt cultures as social systems, in which these emotions are functional components. These accounts have different implications for how the mechanisms that underlie the social function – as it is understood on a macrolevel – look in detail and consequently also different implications with regard to how predict or manipulate pride and shame responses. The main problem for the causal dispositionalist is, however, the relation between part and whole. The quality of the functionalist explanation hinges on how well that relation can be established. A first problem, when it comes to evaluating causal dispositionalist accounts, is that they are often unclear with regard to what the maintaining system is (individual? family? community? society?), and where they are clear, the system is too complex to allow for a comprehensive mechanist analysis of all the components and different levels standing in complex relations of interaction. Defenders of the interpersonal view, for instance, not only assume that emotions unfold in a certain way during infancy and within social interaction, but they also assume that certain emotion types develop cross-culturally but can vary to a certain degree in different cultures. Since shame is assumed to have the social function of motivating people to restore harmony, shame is assumed to be more relevant in collectivistic cultures or in social settings with high power distance in status (Parkinson, Fischer, and Manstead 2005, 74–75). In a similar vein, the historian Ute Frevert analyzes the decline of public shaming in the context of the European Enlightenment and its persistence in the Chinese education system (Frevert 2016). The claim that cultures can be divided into shame and guilt cultures is a very strong assumption with far-reaching implications. It is assumed that shame and guilt reliably come to play a crucial causal-role function in the overall hierarchical and normative order of a social group or society. These claims are very demanding to unpack with regard to function bearer and system: Frevert, for example, assumes that emotions need to be analyzed as social practices, where whether and how they are experienced depends on their relative status in a given society. To analyze this status, the weight and impact of all kinds of institutions need to

110  Rebekka Hufendiek be taken into account. What men felt, for instance, when they evoked notions of pride and honor in the 19th century needs to be studied by examining institutions such as the education system or the military and their social ramifications (Frevert 2016, 56). Like the interpersonalists, Frevert takes shame to be a response with universal behavioral tendencies such as shrinking and gaze aversion that can come to be felt, enacted, and evaluated differently in different contexts. Shaming practices, according to her, developed differently in the cultural contexts of post-Enlightenment Europe and present-day China. While triggering shame still plays a central role in social and moral education in China, Frevert claims that Such tactics would no longer be endorsed in the Western hemisphere where shaming has lost currency, both in schools and in the wider public. … [S]haming is generally held to be a violation of human dignity and the duty of respect. In highly individualistic societies that put personal autonomy first and community values second, public shaming seems no longer acceptable. (Frevert 2016, 59) While Frevert offers many sources to bolster and develop the claim about the Western world (and has developed that claim in other works; see Frevert, 2011), the only reference for the further claim is a BBC documentary on the Chinese education system from 2008. That is amazingly little evidence, particularly because a European historian would certainly not come up with comparably little evidence to argue for any such claim about “the Western hemisphere.” But apart from that, the more general problem is that the claim that emotions such as shame are universal in the sense that Frevert and many interpersonalists assume, and the claim that they vary with regard to how they are embedded into social practices, are both very demanding claims that cannot be backed up by evidence that would come close to causal modeling. On the other hand, historians in particular might be the first ones to object that they are not aiming for pure causal explanations, since that does not seem to be an adequate tool to capture all the dimensions of complex human interactions, including their rational motivation and their cultural meaning. When psychologists and anthropologists as well as historians offer functional explanations of pride and shame, they offer explanations that frequently draw on rational and interpretive modeling at least as much as on causal evidence. These eclectic explanations can still be evaluated with regard to empirical adequacy: one might find Frevert’s analysis of the decline of shame in Europe a compelling interpretation, given the available data, but still insist that her claims with regard to China overgeneralize on the basis of a single source in a way that is insufficient for all kinds of scientific explanations.

Functional Explanations of Shame and Pride  111

Functional Explanations, Ideology, and Rational and Interpretive Modeling I have argued that functional claims about shame and pride cannot be unpacked in the form of a causal explanation and it does not seem as if the relevant empirical data will be available any time soon. I have also suggested that this is because functional explanations have to situate their explanandum within a holistic framework and in doing so authors have to use an eclectic strategy and bring in elements of rational and interpretive modeling. Ideology I will now demonstrate that presenting such eclectic functional explanations as if they rested on solid evidence of a causal mechanism leads to overgeneralization and reification, which often leads the relevant works down a slippery slope towards ideological claims. Reification is one way through which ideologies can present misleading claims with the effect of stabilizing oppressive social relationships (Geuss 1981; Shelby 2003). To be clear, I am not claiming that any of the accounts of pride and shame discussed so far is clearly serving a hegemonic function or stabilizing oppressive social relations. What I have shown so far, however, is that the approaches of pride and shame discussed here deserve epistemic critique, where they present functional explanations and make it seem as if they had the evidence for an underlying causal mechanism. I will argue further in the following that because of the holistic structure of functional explanations many of the accounts discussed so far show strong tendencies to overgeneralize and reify their findings, or adopt background assumptions about the human condition that are themselves highly contested. I will further suggest that there are tendencies of reification to be found in some of the works discussed that deserve closer attention. The status-signaling theory, in the variant that Tracy defends, suggests that shame and pride expressions signal dominance and submission, or more or less prestige, in a hierarchically structured group to smoothen organization and cooperation. That suggests that hierarchical organization is a part of human evolutionary history that left deep traces in the way we emotionally communicate with each other. Such an account makes strong assumptions about the influence of evolution on the current expression of human nature. Status differences are supposed to emerge in all human societies and these differences supposedly influence patterns of conflict, facilitate coordination on group tasks, and so on (Tracy et al. 2013, 163). In her recent book, Tracy develops the idea that hierarchical organization of some form is an inescapable part of human societies. She thereby embraces a strong modal claim: given the structure

112  Rebekka Hufendiek of human nature, including shame and pride as motivating forces, human beings cannot form societies that are entirely free of hierarchies. According to Tracy’s further claim that pride comes in either an authentic or a hubristic variety, she offers the idea that hierarchical structure can be organized in two different ways: prestige- or dominance-oriented hierarchies. These hierarchies relate to the two forms of pride: authentic pride and hubristic pride. Where Tracy reamins neutral on these findings in her studies, in her popular book she combines these findings with normative advice: after having identified Obama as a typical prestige seeker of the authentic pride variant and Trump as the typical dominance seeker of the hubristic type, she goes on to analyze the moral differences of the two kinds of leadership to be observed in the two types: Authentic pride makes people work hard to achieve and to behave morally – to strive to be a good person. This is the kind of pride that should make people leaders. Hubristic pride has a completely different set of consequences. It makes people un-empathic, disagreeable, and even aggressive. (Tracy 2016, 117) It would be misleading to say that Tracy is committing a naturalistic fallacy here, since she takes both kinds of pride to be equally natural, universally present in human beings, and adaptive. It is rather that Tracy introduces normative claims about how we should and should not behave that she does not justify further, apparently because she takes them to be too obviously true for further justification. She takes one type of pride-motivated behavior to be moral and the other immoral and suggests on these grounds that we should seek to be Obama-like achievers and avoid being Trump-like bullies. One might agree with that statement or not, but the point here is that it relies on an overgeneralization and reification of empirical findings. It basically presupposes that in all possible social circumstances there will be hierarchies and a need for leaders, and that leaders come in the Obama and the Trump variety of which one is morally good and the other morally bad. So, if we want to be moral, the recipe for that is to seek the Obama type within ourselves. These claims are made in the context of a popular book that keeps on repeating that pride is a human universal without mentioning that this is a theory that is highly controversial in current emotion research. We have seen that Tracy’s account can be critized epistemically for offering a theory that sells its result as based on scientifically solid evidence that can offer general insights about the human condition, where the evidence is in fact far less clear than Tracy suggests. Tracy apparently suggests that all human societies will necessarily establish hierarchies of one form or the other. In this general form, the claim is so vague that it seems hard to falsify. What is a hierarchy, and what would a

Functional Explanations of Shame and Pride  113 nonhierarchical organization look like? In its generality, however, the claim seems to invite being used as a basis for the justification of all kinds of social inequalities. Social inequalities, however, have often been justified with the claim that they are natural and thereby inescapable. The claim that hierarchical organization is natural in this sense has very recently been defended by authors such as Jordan Peterson, according to whom hierarchical organization goes back at least 350 million years in evolution to creatures like the infamous lobster who in its fight for a dominant position in a hierarchy already shows a slumped or expanded posture. Peterson is also explicit regarding the normative conclusions that follow from his amazingly overgeneralized findings: for humans, just as for lobsters, The bottom of the dominance hierarchy is a terrible, dangerous place to be. The ancient part of your brain specialized for assessing dominance watches how you are treated by other people. On that evidence, it renders a determination of your value and assigns you a status. (Peterson 2018, 45) These claims, on a closer analysis, would probably show all the features of an ideology in the narrow sense in pure form, and that is certainly a notable difference from authors such as Tracy who are mainly known for conducting interesting yet controversial studies on emotions and emotion-related behavior. The observed tendency towards ideological claims, however, can be described as a slippery slope towards Peterson’s lobster. This is a long slope that starts with selectionist explanations, suggesting that pride and shame have an adaptive function, a long evolutionary history, and form a fundamental part of human nature without making clear that all these claims remain interpretive and controversial. Explicit essentializing conclusions about how living creatures must organize in hierarchies and how terrible it is to be at the bottom of such hierarchies, and that some unlucky creatures will always be there so it better not be you, mark the end of the slope. The social norm-enforcement theory, on the other hand, refers to social practices that constitute emotion types within our permanent interactions. According to those theories, emotions can have social functions on different levels: within small groups such as the family but also within societies. The most elaborated example of such a complex social emotional dynamic is the distinction between shame and guilt cultures. Shame cultures are cultures that make extensive use of public shaming with the aim of adapting individual behavior to the structure of society. Guilt cultures are those that value autonomy and the freedom and responsibility of the individual. Shaming is significantly less prominent in these cultures. Again, this has strong instrumental practical implications

114  Rebekka Hufendiek (which lead to the claim that cultures that value autonomy should reduce the role of shame in social and moral education and certainly in the justice system, as is indeed argued by some authors in moral and political philosophy; see Nussbaum 2004). A quick look back to the origins of the distinction between shame and guilt cultures makes it seem reasonable to ask for a comprehensive critique. The anthropologist Ruth Benedict popularized the distinction in her book The Chrysanthemum and the Sword, the main purpose of which was to inform the US public and particularly the military during the war against Japan about what to expect from the enemy, relying on the assumption that the enemy would behave significantly differently due to cultural differences (Benedict 1946). Those differences particularly concern hierarchical organization and the burden of shaming sanctions that Benedict describes as standing in sharp contrast to US culture and values: The Japanese … order their world with constant reference to hierarchy. … As long as ‘proper station’ is maintained, the Japanese carry on without protest. They feel safe. They are of course often not ‘safe’ in the way that their best good is protected but they are ‘safe’ because they have accepted hierarchy as legitimate. It is as characteristic of their judgment on life as trust in equality and free enterprise is of the American way of life. (Benedict 1946, 95–96) Benedict’s work has been criticized within anthropology for the little evidence it offers (the book was written during the war and Benedict could not travel to Japan), its vast overgeneralizations, and its presentation of Japanese culture as entirely alien to US citizens (Geertz 1988). It is therefore surprizing that it remains a prominent reference in psychology and even more so, since it does not even establish the claim that it is frequently cited for, namely that all cultures divide into those that cultivate guilt and those that cultivate shame. The slippery slope encountered here is one that begins with overgeneralizations of the role of shame in countries such as Japan or China, runs down the slope to even more overgeneralized claims about how shame is present and not present in Western and non-Western cultures (sometimes based on incredibly little evidence), and ends sometimes with a plea for mutual acceptance and relativism, and sometimes with claims that one of these cultures is superior. Interpretation We have seen that functional explanations of shame and pride cannot be established by pointing to solid evidence for the relevant underlying causal mechanisms. We have also seen that, where authors present their findings as if that were the case, they run the danger of sliding down a slippery slope towards ideological claims about what we can and cannot

Functional Explanations of Shame and Pride  115 change about human nature and social structures. To suggest on the basis of that tension that we should simply refrain from all kinds of functional explanations does not seem like a very promising suggestion, since functional explanations are attempts to answer questions that are of great interest. We have also seen that the trouble with functional explanations stems from their holisitic nature. They can be described as the topping-off point of a theory that describes complex multilevel mechanisms and then tells us what parts of these mechanisms are supposed to do, what they do under normal circumstances, or why what they do is beneficial for the system. Perspectivalists suggest that these judgments about normality lie in the eye of the beholder. I tried to elaborate this claim further by suggesting that the judgments about what is normal do not need to rely on purely individual interests, but can be grounded in rational and interpretive modeling instead. What would that mean for the explanations available? We have seen in the last sections that there is a broad consensus on the claim that the functionality of pride and shame depends on the presence of a social group or structure. Shame and pride would not be functional in a nonsocial individual that did not need to interact with a structured group of conspecifics as a fundamental part of its existence. That in itself is an interesting finding, but it hardly seems to be an empirical finding (although it might be empirically falsified). It rather seems to be the result of rational modeling, where one strategy to find out what pride is would be to ask what a world without pride and shame would look like, what kinds of needs pride and shame satisfy, what kinds of problems they solve, and so on. Reasoning about the functions of cognitive and behavioral features in this vein seems to be a promising way to avoid Panglossian functionalism and replace it with rational reconstructions that can clarify what needs to be the case for a cognitive or behavioral feature to be functional, as well as illuminating the dependency relations among various cognitive and behavioral features (pride and shame can only occur in beings that already are social to a certain degree) and offer interpretations of the needs that such features might fulfill or the problems they might solve. Particularly with regard to those needs, we have seen that the suggestions for possible functions of shame and pride range from status signaling to prestige signaling and social norm enforcement. That there is such a broad spectrum of possible functions is, again, an interesting finding in itself that enhances self-understanding and can elucidate further research in psychology, anthropology, and history alike. It is important to recognize that the theories discussed above make different assumptions with regard to the question of whether one of these is more fundamental than the other, whether all are the result of biological evolution or only some, whether all are universally present or not, and not surprisingly with regard to the question of how desirable they are, and to what extent we can manipulate and inculcate these responses and in which

116  Rebekka Hufendiek directions. That makes it all the more relevant to present such claims as what they are: as (at least in part) interpretations that entail judgments about what is normal and abnormal and what we can and should do. I have suggested that rational modeling can make sense of the broad patterns we observe in human cognitive and behavioral traits, and that interpretive modeling is concerned with the subtleties of the meaning of particular social practices. Two final examples might illustrate how I think that both rational and interpretive modeling can play a positive role in reasoning about the functions of pride and shame. First, the anthropologist Daniel Fessler suggests that status signaling does not pressupose collective intentionality or a theory of mind and can therefore predate human evolution. This is an interesting case of considering dependency relations that obviously needs to be empirically informed, but it can also test which assumptions are rational when asking when and how shame and pride could have evolved. These kinds of “how possibly” considerations figure prominently in reasoning about the evolution of functional traits, and I do think that they deliver valuable results in and of themselves, that they can be tested for coherence, and inspire future research. Second, the historians David Nash and Anne-Marie Kilday have offered rich studies with comprehensive interpretations of how public shaming lost its dominant role in Scotland from the 18th century onwards (Nash and Kilday 2010). Within such an interpretive framework, general ideas of status and prestige functioning and more than anything norm enforcement do play a role in making sense of pride and shame. Such analyses, however, can bring into focus the details of how public shaming became dysfunctional in certain political contexts but regained power in other social domains, instead of simply vanishing during a process of Enlightenment (Nash and Kilday 2010). Such interpretations, furthermore, refrain from overgeneralizations by pointing to the fact that functional mechanisms (if one wants to talk about mechanisms here at all) that are part of a social structure are not stable in their functioning but can be permanently reshaped through social practices. The interpretive framework can capture the fine-tuning of functioning and dysfunctioning to be found in the history of public shaming. An amazing example discussed by the authors of how social practices can influence these functions in different directions is Daniel Defoe’s “Hymn to the Pillory,” written in 1703 aiming to revoke the shame the sentence of three days at the pillory was meant to bring about in him, and addressed to those who put him there: Shame, like the exhalations of the sun, Falls back where first the motion was begun; And they who for no crime shall on thy brows appear, Bear less reproach than they who placed them there. (Nash and Kilday 2010, 63)

Functional Explanations of Shame and Pride  117 It is certainly a complex question whether and to what degree such an attempted change in practices surrounding who is and is not to feel shame can be succesful. These are relevant questions with regard to the social functions of pride and shame and they cannot be addressed by causal modeling alone.

Conclusion I have argued that functional explanations of cognitive and behavioral features are holistic in their aim and therefore need to combine causal, rational, and interpretive modeling. I have demonstrated this discussing the examples of pride and shame. The discussion of current research on the function of these emotions demonstrates that we do not have the empirical evidence for the causal mechanisms that are supposed to underlie both etiological and causal-role explanations of the functions of pride and shame. I have also demonstrated that where researchers present their findings about these functions as if they had evidence for the complex causal mechanisms in question, their works show ideological tendencies of reification and overgeneralization. As a remedy for research on cognitive and behavioral features in general, I suggest treating functional explanations as eclectic explanations that entail elements of causal, rational, and interpretive modeling. Making the rational and interpretive elements explicit and measuring them with regard to their own standards can avoid overgeneralization and reification, without having to discard all attempts to offer functional explanations of cognitive and behavioral features for lack of mechanisms altogether.

Acknowledgments It took a while to write and rewrite this chapter, and several people have been reading it during that time. In particular, I  would like to thank Joshua Crabill, Daniel James, Raphael van Riel, Christine Sievers, and Markus Wild for their feedback and comments. I  would also like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

Notes



118  Rebekka Hufendiek



References Barrett, Lisa Feldman. 2012. “Emotions Are Real.” Emotion 12, no. 3: 413–429. ———. 2017. How Emotions Are Made: The Secret Life of the Brain. Boston, MA: Houghton Mifflin Harcourt. Barrett, Lisa Feldman, and James Russell, eds. 2015. The Psychological Construction of Emotion. New York; London: The Guilford Press. Beall, Alec, and Jessica Tracy. forthcoming. “The Evolution of Pride and Shame.” In Cambridge Handbook of Evolutionary Perspectives on Human Behavior, edited by Lance Workman, Will Reader, and Jerome H. Barkow. Cambridge: Cambridge University Press. Benedict, Ruth. 1946. The Chrysanthemum and the Sword: Patterns of Japanese Culture. Boston, MA: Houghton Mifflin. Clark, Jason. 2010. “Relations of Homology between Higher Cognitive Emotions and Basic Emotions.” Biology & Philosophy 25, no. 1: 75–94. Craig, Edward. 2005. Knowledge and the State of Nature: An Essay in Conceptual Synthesis. Repr. Oxford: Clarendon. Craver, Carl. 2007. Explaining the Brain: Mechanisms and the Mosaic Unity of Neuroscience. Oxford: Oxford University Press. ———. 2013. “Functions and Mechanisms: A Perspectivalist View.” In Functions: Selection and Mechanisms, edited by Philippe Huneman, 133–158. Dordrecht: Springer. Crivelli, Carlos, and Alan Fridlund. 2019. “Inside-Out: From Basic Emotions Theory to the Behavioral Ecology View.” Journal of Nonverbal Behavior 43, no. 2 (January): 161–194. Cummins, Robert. 1975. “Functional Analysis.” Journal of Philosophy 72 (November): 741–764. Darwin, Charles. 2009. The Expression of the Emotions in Man and Animals. Cambridge: Cambridge University Press. Dupré, John. 2001. Human Nature and the Limits of Science. Oxford: Clarendon Press; Oxford: Oxford University Press.

Functional Explanations of Shame and Pride  119 Elster, John. 1982. “Marxism, Functionalism, and Game Theory: The Case for Methodological Individualism.” Theory and Society 11, no. 4: 453–482. Fessler, Daniel. 2004. “Shame in Two Cultures: Implications for Evolutionary Approaches.” Journal of Cognition and Culture 4, no. 2: 207–262. ———. 2007. “From Appeasement to Conformity. Evolutionary and Cultural Perspectives on Shame, Competition, and Cooperation.” In The SelfConscious Emotions. Theory and Research, edited by Jessica Tracy, Richard Robins, and June Tangney, 174–194. New York; London: The Guilford Press. Fischer, Agneta, and Antony Manstead. 2016. “Social Functions of Emotion and Emotion Regulation.” In Handbook of Emotions, edited by Lisa Feldman Barrett, Jeannette Haviland-Jones, and Michael Lewis, 4th ed., 424– 440. New York; London: The Guilford Press. Frevert, Ute. 2011. Emotions in History: Lost and Found. The Natalie Zemon Davis Annual Lecture Series at Central European University, Budapest. Budapest; New York: Central European University Press. ———. 2016. “The History of Emotions.” In Handbook of Emotions, edited by Lisa Feldman Barrett, Jeannette Haviland-Jones, and Michael Lewis, 4th ed., 49–65. New York; London: The Guilford Press. Geertz, Clifford. 1972. “Deep Play: Notes on the Balinese Cockfight.” Daedalus 101, no. 1: 1–37. ———. 1988. Works and Lives: The Anthropologist as Author. Stanford, CA: Stanford University Press. Geuss, Raymond. 1981. The Idea of a Critical Theory: Habermas and the Frankfurt School. Modern European Philosophy. Cambridge: Cambridge University Press. Griffiths, Paul. 1997. What Emotions Really Are: The Problem of Psychological Categories. Chicago, IL: University of Chicago Press. Hufendiek, Rebekka. 2015. Embodied Emotions: A Naturalist Approach to a Normative Phenomenon. New York; London: Routledge. Huneman, Philippe, ed. 2013. Functions: Selection and Mechanisms. New York: Springer. Kincaid, Harold. 2007. “Functional Explanation and Evolutionary Social Science.” In Philosophy of Anthropology and Sociology, edited by Stephen Turner and Mark Risjord, 213–247. Amsterdam: Elsevier. Lewis, Michael. 2016. “Self-Conscious Emotions: Embarrassment, Pride, Shame, Guilt, and Hubris.” In Handbook of Emotions, edited by Lisa Feldman Barrett, Jeannette M. Haviland-Jones, and Michael Lewis, 4th ed., 792–815. New York; London: The Guilford Press. Little, Daniel. 1991. Varieties of Social Explanation: An Introduction to the Philosophy of Social Science. Boulder, CO: Westview Press. Mesquita, Batja, Jozefine Leersnyder, and Michael Boiger. 2016. “The Cultural Psychology of Emotions.” In Handbook of Emotions, edited by Lisa Feldman Barrett, Michael Lewis, and Jeannette Haviland-Jones, 4th edn., 393–411. New York; London: The Guilford Press. Millikan, Ruth. 1989. “In Defense of Proper Functions.” Philosophy of Science 56 (June): 288–302. Nash, David, and Anne-Marie Kilday. 2010. Cultures of Shame: Exploring Crime and Morality in Britain 1600–1900. London: Palgrave Macmillan.

120  Rebekka Hufendiek Neander, Karen. 1991. “The Teleological Notion of ‘Function.’” Australasian Journal of Philosophy 69, no. 4: 454–468. Nussbaum, Martha. 2004. Hiding from Humanity: Disgust, Shame, and the Law. Princeton, NJ: Princeton University Press. Parkinson, Brian, Agneta Fischer, and Antony Manstead. 2005. Emotion in Social Relations: Cultural, Group, and Interpersonal Perspectives. London: Psychology Press. Peterson, Jordan B. 2018. 12 Rules for Life: An Antidote to Chaos. Toronto, ON: Random House Canada. Saborido, Cristian, Matteo Mossio, and Alvaro Moreno. 2010. “The Teleological Dimension of the Concept of Biological Function from the Organizational Perspective.” Teorema: International Journal of Philosophy 29, no. 3: 31–56. Scarantino, Andrea. 2016. “The Philosophy of Emotions and Its Impact on Affective Science.” In Handbook of Emotions, edited by Lisa Feldman Barrett, Michael Lewis, and Jeannette Haviland-Jones, 4th edn., 3–49. New York; London: The Guilford Press. Scheff, Thomas. 1988. “Shame and Conformity: The Deference-Emotion System.” American Sociological Review 53: 395–406. Searle, John. 1995. The Construction of Social Reality. New York: Free Press. ———. 2010. Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press. Shariff, Azim, Jessica Tracy, and Jeffrey Markusoff. 2012. “(Implicitly) Judging a Book by Its Cover: The Power of Pride and Shame Expressions in Shaping Judgments of Social Status.” Personality and Social Psychology Bulletin 38, no. 9: 1178–1193. Shelby, Tommie. 2003. “Ideology, Racism, and Critical Social Theory.” The Philosophical Forum 34, no. 2: 153–188. Shepherd, Lee, Russell Spears, and Antony Manstead. 2013. “When Does Anticipating Group-Based Shame Lead to Lower Ingroup Favoritism? The Role of Status and Status Stability.” Journal of Experimental Social Psychology 49, no. 3: 334–343. Shweder, Richard, Jonathan Haidt, Randall Horton, and Craig Joseph. 2008. “The Cultural Psychology of the Emotions: Ancient and Renewed.” In Handbook of Emotions, edited by Michael Lewis, Jeannette Haviland-Jones, and Lisa Feldman Barrett, 3rd ed., 409–427. New York; London: The Guilford Press. Sperber, Dan. 2007. “Seedless Grapes: Nature and Culture.” In Creations of the Mind: Theories of Artifacts and Their Representation, edited by Stephen Laurence and Edouard Margolis, 124–137. Oxford: Oxford University Press. Steckler, Conor, and Jessica Tracy. 2014. “The Emotional Underpinnings of Social Status.” In Psychology of Social Status, edited by Joey Cheng and Jessica Tracy, 201–224. New York: Springer. Taylor, Charles. 1985. Philosophy and the Human Sciences. Philosophical Papers 2. Cambridge: Cambridge University Press. Tracy, Jessica. 2016. Take Pride: Why the Deadliest Sin Holds the Secret to Human Success. Boston, MA; New York: Houghton Mifflin Harcourt. Tracy, Jessica, and Richard Robins. 2007a. “The Nature of Pride.” In SelfConscious Emotions: Theory and Research, edited by Jessica Tracy, Richard Robins, and June Tangney, 263–279. New York; London: The Guilford Press.

Functional Explanations of Shame and Pride  121 ———. 2007b. “Emerging Insights into the Nature and Function of Pride.” Current Directions in Psychological Science 16, no. 3: 147–150. Tracy, Jessica, Azim Shariff, and Joey Cheng. 2010. “A Naturalist’s View of Pride.” Emotion Review 2, no. 2: 163–177. Tracy, Jessica, Azim Shariff, Wanying Zhao, and Joseph Henrich. 2013. “Cross-Cultural Evidence That the Nonverbal Expression of Pride Is an Automatic Status Signal.” Journal of Experimental Psychology: General 142, no. 1: 163–180. Trommsdorff, Gisela, and Hans-Joachim Konrad. 2003. “Parent-Child Relations in Cross-Cultural Perspective.” In Handbook of Dynamics in ParentChild Relations, edited by Leon Kuczinsky. Thousand Oaks, CA: Sage Publications. 271–306. Wright, Larry. 1973. “Functions.” Philosophical Review 82, no. 2: 139–168.

6

What Explains Social Role Normativity? Charlotte Witt

The excellence of a thing is relative to its proper function. (NE 6.1 1139a 16–17)

Introduction There are many dimensions to the topic of social normativity and many approaches one could take. My work focuses on social role normativity, which is the normativity that attaches to social positions and to their occupants. One reason for my focus is that all human societies include social positions and associated roles in their ontologies, while differing in the particular social roles available.1 So, if we could understand the normativity of social roles, we would understand something central to the normative structure of human social worlds. In this chapter, I explain the normativity of social roles; that is, I propose an explanation of their character as normative.

What Is a Social Role? Social roles demarcate ways of being a person or a social agent. Examples of social roles include being a carpenter or a baker, being a mother or a father, being a President or a professor, being a woman, or being Asian American. 2 I begin by emphasizing that social roles are ways of being a person or being human that specify what we ought to do or how we ought to act rather than characterizing them primarily as a vehicle that establishes kinds of persons because I want to highlight the idea that these roles normatively structure our social agency. It is useful to differentiate an inquiry into social roles that focuses on their causal status and the way in which they explain what an individual will do, from my inquiry into social roles, which focuses on their normativity and the way in which they specify what an agent ought to do. So, even if social roles are central mechanisms of category construction (Mallon 2016, 57) and thereby directly relevant to the possibility of social scientific knowledge, my interest is in exploring their ontology,

What Explains Social Role Normativity?  123 on what it is to be a social role rather than approaching social roles in relation to the possibility of social knowledge.3 If one views social roles primarily in relation to social kind construction, then the central issue will be how these kinds can achieve sufficient stability to be objects of social scientific knowledge in a fashion parallel to the way in which natural kinds are objects of scientific knowledge. This is, of course, an important question. But it is not the central question driving my inquiry. I am interested in social roles as normative ways of being human – of being in the social world – and in the puzzle of the source or nature of their normativity. I propose to explain their normativity in terms of the work or function associated with the social role. Rather than endorse contemporary accounts of the normativity of functional ascriptions, however, I borrow the idea of an Afunction from Aristotle – for whom artisans and artists are paradigmatic occupants of social roles, and their artisanal or artistic activities are comprised of a set of normatively evaluable techniques or expertise. Being a carpenter is a social role that entails a set of techniques (involving tools and materials) and expertise that are evaluable in relation to the function of making things out of wood. But being a mother is also a way of being in the world that entails a set of techniques (involving tools and materials) and expertise that are evaluable in relation to the activity of mothering. A bit more controversially, I propose that we think of gender and race or ethnicity along the same lines, as social roles, as ways of being a person that are normatively evaluable forms of expertise. Think of the ways in which a young man might fail to fulfill the norms of masculinity by wearing the “wrong” clothes, for example, and another might live an academically driven existence in college, and thereby engage the norms and expectations that define being a hard-working and high-achieving Asian American.

What Is Social Role Normativity? Consider President Trump. His decisions, statements, and actions are criticized from many perspectives and for many reasons. However, a  central line of criticism focuses on his current social position. Yes, it is wrong to make racist statements, but it is particularly wrong for a President. Yes, it is wrong to bend the truth, but it is particularly wrong for a President. Yes, it is wrong to exploit one’s profession financially, but it is particularly wrong for a President. You get the idea. Even though its exact content is debatable, there is a social role, a set of norms that attaches to Trump because he is President, not because he endorses them or has promised to keep them, but because he occupies this very social position. The normativity is just part of what it is to occupy a social position. Every social position occupancy draws a social role in its wake. Notice that Trump’s behavior as President has not been predictable in

124  Charlotte Witt relation to the cluster of properties that define the social kind President; the social role can fix what the occupant ought to do, but it cannot reliably predict what an individual will do. Of course, the norms that attach to President Trump by virtue of his social position as President are also publicly undertaken. Being a president, like being a priest or being a soldier, is a social position whose norms are voluntarily, publicly, and explicitly undertaken by the individual in question. Nonetheless, there is a difference between what it is that grounds or explains Trump’s obligation to protect and defend the US Constitution, which is that Trump is President, and the ceremonial and public acceptance of that responsibility. When we criticize an action of Trump’s as not presidential, we are not primarily focused on his oath of office but rather on the norms that govern his actions as President. To put the point another way, we might also and in addition criticize President Trump because he broke his oath of office to protect and defend the United States by consorting with Putin and the Russian oligarchs. While the existence of social positions and roles in human societies is obvious, the character of their normativity is not. How do we explain the idea that obligations or “ought tos” are somehow implanted in the social world, attached to its positions, practices, and structures? In addition, and adding to the puzzle, we can all think of aspects of social roles that are politically and morally unacceptable, and it seems hard to square this idea with the idea of their normativity. So, social role normativity seems to be particularly in need of explanation both because it is a widespread form of social obligation and because it seems to be centrally implicated in maintaining adherence to oppressive aspects of social roles. There are several questions that arise for our understanding of social normativity in the context of social roles and social position occupancy. First, we might wonder how and why the norms that define particular social roles attach to individual social agents. I have developed the distinction between voluntarism and ascriptivism as a frame for discussing this topic (Witt 2011). A voluntarist thinks that an individual is responsive to and evaluable under a social norm because she prefers, endorses, or identifies with that norm. For the voluntarist the fact that an individual occupies a given social position might make her a candidate for responsiveness and evaluation but it requires, in addition, that she prefers, endorses, or identifies with the norm.4 In contrast, an ascriptivist thinks that a norm attaches to an individual just because she occupies a social position where that norm is part of the social role governing the position.5 For the ascriptivist, a mother or a father is responsive to and evaluable under maternal and paternal social norms (which of course vary both from society to society and within a single society) simply because the individual is a mother or is a father. In contrast, the voluntarist

What Explains Social Role Normativity?  125 thinks that the source of normativity lies in the subject, in the individual’s preference, explicit or implicit endorsement or self-legislation of the norm, or their identification with it. A second important question about social role normativity concerns the origin of the normativity of social norms, especially for the ascriptivist. For the voluntarist about social norms the normativity originates in the preference, endorsement, or self-conception of the individual. But, for the ascriptivist this kind of answer is not available. It is a puzzle how the norms associated with a social role could exert normative authority in relation to an individual if the authority in question is not a matter of subjective preference, self-legislation, or self-conception. After all social roles vary from society to society and are embedded in different constellations of practices and different kinds of social structures. And while we can tell a descriptive or anthropological story about the social norms in a given society or in human societies in general, such a story does not establish or require either that the norms in question are irreducibly normative (collectively) and also (individually) open to assessment as legitimate or illegitimate. These two requirements on social roles – that they are irreducibly normative and also open to assessment – might seem to pull in opposite directions, but I take it that an adequate explanation of social role normativity should do justice to both of them.

Social Role Normativity and Afunctions My suggestion is that social role normativity should be understood in terms of an individual’s social position occupancy.6 It nails the social role to the social position occupancy itself rather than to the agent’s attitudes towards those norms. But where does the normativity reside in the ascriptivist picture? Following Aristotle, I think it is useful to think about social roles as constituted by a function or by work that is to be realized in an activity or in actions: For just as for a flute-player, a sculptor or any artist, and, in general for all things that have a function (ergon) and an action (praxis), the good and the well is thought to reside in the function, so would it seem to be for a human being, if he has a function. Have the carpenter, then, and the shoemaker certain functions and actions, and a human being none – is he by nature idle (argon)? (NE 1.7 1097b 25–30) This is, of course, Aristotle’s famous function argument. I am not interested here in exploring the argument itself in any depth.7 Rather, I want to focus on certain ideas about social roles, functions, and normativity, and the relationship among them, that it presupposes, and that comprise being an Afunction.

126  Charlotte Witt But, first, a little background will be useful. In the preceding chapters of the Nicomachean Ethics Aristotle has been pursuing a normative question concerning what the final – i.e., most complete, self-sufficient – good is for a human being. Aristotle thinks that by identifying the final good he will be able to fill in some content to the notion of eudaimonia or human flourishing, which all agree is the end or goal of human life. His strategy here is to argue from the existence of a function (and a good) for a human being occupying a social role, to the existence of a function (and a good) for a human being simpliciter. Rachel Barney notes: The builder and the shoemaker are human beings, identified qua practitioners of a certain craft [techne]; they are socially constructed kinds of human beings, or roles or identities that a human being may take on. (Barney 2008, 297) Now for Aristotle (and for Plato), crafts, like building and making shoes, and arts, like flute playing, are intrinsically normative activities; each activity realizes a function that is associated with an excellence in relation to which the function is performed well (or badly). So, for Aristotle function or work is a normative concept tied to the excellence of a thing and its good, and not merely a descriptive term attaching to a person’s job or occupation. My suggestion is that if we think about a social role in relation to a function realized in activities or actions, then the normativity of the social role flows from, and is grounded in, a normatively infused notion of function. Aristotle points to the conceptual connection between function (ergon), end, and activity (energeia) in the Metaphysics (9.8 1050a 21–23). To be a carpenter is to perform a set of actions and activities that are normatively assessable in relation to the function or work of being a carpenter, of being what one is. To be a mother is to perform a set of actions and activities that are normatively assessable in relation to the work of being a mother, of being what one is. Aristotle and Plato differ both in their central examples of functions and in their understanding of functions, and these differences are important for my purposes.8 For Plato, a tool is the paradigmatic case of a functional being, and a tool requires both a designer and a user in order to have and to realize its function. The function of a tool is extrinsic in these regards; both the origin and the realization of the function refer to entities external to the tool itself. For this reason, the normativity of a tool can be given an instrumental explanation. A hammer is to be used to drive nails; it was created with, and is used by a carpenter for that purpose.9 Tools are paradigm instances of functional entities with instrumental value.

What Explains Social Role Normativity?  127 In contrast, Aristotle’s central examples in the function argument (and elsewhere) are artisans. Both the source and the realization of the function obtain just insofar as that individual is an artisan of a certain kind. Function and identity go hand in hand; what something does is what it is. “This relation to end and identity makes ergon a powerful normative concept closely linked to the good of a thing and determining what counts as its excellence or virtue (arête)” (Barney 2008, 301). For a person to be an excellent flute player is to function well qua flute player, and neither the origin nor the activation of the function has an external source. Artisans, unlike tools, are not paradigm examples of functional entities with instrumental value; their norms are rooted in their identity and the standards of excellence that define the relevant function/activity. Some social roles, like being a flautist, are primarily comprised of actions, like practicing and performing, and their excellence is measured by the actions themselves.10 However, Aristotle distinguishes between acting (praxis) and making (poiesis) – between activities like flute playing that have no product beyond the activity itself, and arts like housebuilding which are activities that can be done well or badly, but also have a product that is an independent standard by which to judge both the agent (the builder) and the activity (the building). As Sarah Broadie puts it, the craftsman’s good productive activity, is in one sense, his end: the perfection of him as a craftsman. The product, which is other than the activity, is his end in a different sense: it is that by which one judges his success, hence judges him to have achieved his end in the first sense. We assess the excellence of the making, hence of the maker, by the quality of what is made. (1994, 207) In the case of actions, the evaluations are internal to the actions themselves, but in the case of makings, there is an external standard as well – the product. Although the normativity in the case of making (poiesis) is more complex than in the case of action (praxis), in both instances the values are intrinsic in the sense that they are realizations of the craftsman’s end or function, of what it is to be an artisan of this kind. Insofar as the individual is an artisan they realize their functions, their excellence as artisans, in their artistic or productive activities. This excellence or goodness, this expertise, is intrinsic to being an artisan, to what the artisan is, and not merely instrumental as in the function of a tool. Of course, what makes someone a good flute player in Brussels today might differ from what made someone a good flute player in Aristotle’s Athens in that both the techniques for playing well and the particular aesthetic properties that constitute playing well might differ, and so the abilities and expertise of a good flautist might differ as well. But this is entirely

128  Charlotte Witt consistent with there being objective standards for what counts as playing well and being a good flautist in each context. It is not up to the individual flautist to decide how to play well or what counts as playing well around here. In this regard being a flautist is like being a mother or being a President. The activities and attitudes that constitute being a good mother or being a good President might vary among different societies. Nonetheless – and even in cases where they are contested – there is an underlying presupposition that there are shared and objective local norms for being a good mother, or a good President, or good flautist. That is what the contestation is all about. There would be no reason to contest the norms surrounding breast-feeding in the United States in 2018, for example, without the assumption that there is a way to be an excellent mother around here. It is important to be clear that in proposing that we use the notion of an Afunction to explain social role normativity, I am not using the term either as it appears in the discussions of functions in the philosophy of science and the philosophy of biology, or the way it is used in the social sciences. I return to this point below. Rather, I explore a more commonplace idea of function or work, exemplified by arts and crafts, that we find presupposed by arguments in Aristotle and Plato that are intelligible to us – even if they are not ultimately persuasive.11 Note also the OED definition of “function” as “the special kind of activity proper to anything. The mode of activity by which it fulfils its purpose” (OED 1095). In light of the OED definition, it is important to clarify that an Afunction is not a purpose or an intentional goal. When Aristotle says that the good resides in the function or in the action, he might be taken to be referring to the intention or purpose of the flautist; for example, the intention to play well that day. And, indeed, playing beautifully might be the goal of the performer that day. But this is not what Aristotle means when he says that the good resides in the function. Rather, he means that the activity itself, as an expression of the flute-playing function is also an expression of the virtue or goodness of the artist. It is normatively evaluable. And that it is so is a fact independent of the intention of the musician, although often the intention to play well and the activity of playing well will line up. But, for Aristotle, the player’s intention is not the source of the normativity of the activity; that is intrinsic to the activity itself and grounded in the function and its standards of excellence. With his idea of a practice, Alasdair McIntyre captures something of the Aristotelian and commonsense notion of a function, or Afunction, realized in activities guided by expertise or technique: By a practice I am going to mean any coherent and complex form of socially established cooperative human activity, through which goods internal to that form of activity are realized in the course of trying to achieve those standards of excellence which are appropriate to, and partially constitutive of, that form of activity… (MacIntyre 1981, 187)

What Explains Social Role Normativity?  129 MacIntyre lists arts, sciences, games, and politics as examples of practices. The important point, for my purposes, is that with the idea of a practice MacIntyre singles out a kind of social activity that has an internal normativity or, in other words, goods or norms to be realized that are noninstrumental. Further, because practices are defined as cooperative human activities, they presuppose cooperation, and so their normativity could not be derived from their value as instruments of social coordination, on pain of circularity. Of course, practices might also have instrumental value because of their contribution to a stable social organization; but this instrumental value does not ground the normativity of practices for MacIntyre. Unlike MacIntyre, my focus is on the agents who are embedded in his practices, and the social roles they enact in engaging with the practices. Nonetheless, his idea of a practice is useful in describing a normatively rich social context within which agents occupy social roles that are expressed in activities characterized by noninstrumental normativity, the standards of excellence that are “appropriate to and partially constitutive of that form of activity.” MacIntyre’s standards of excellence are what I refer to as the expertise and techniques associated with social role functions. For Aristotle, the expertise and techniques characteristic of the activities of artisans and artists, the standard by which they are evaluable, are both acquired dispositions in individuals (expertise) and bodies of shared and teachable social knowledge (techniques). They thread the needle between subjective and objective; between individual and social world. Expertise is not merely a subjective disposition existing in an individual; it encodes a shared social knowledge, which is not up to the individual to determine. There is no private expertise. And, although techniques are shared and teachable social knowledge, they are also practical and embodied because they are realized in and by the activities of individual artisans pursuing their crafts. Finally, MacIntyre’s discussion of practices is useful in the picture it paints of a social world as rich with shared human activities that have intrinsic and not merely instrumental value. It is also worth pointing out that the norms in this social world are not depicted as originating in the preference, endorsement, or self-identification of rational subjects, but rather in the standards of excellence that are constitutive of the activities and practices themselves. In other words, MacIntyre’s practices are fully compatible with, and indeed an attractive expansion of, the ascriptive approach to social role normativity that I endorse.

Functions in Social, Artificial, and Biological Worlds It may be useful at this point to distinguish the notion of an Afunction that I am using to explain social role normativity from structural functionalism as a concept in the social sciences. Structural functionalism, as

130  Charlotte Witt I understand it, is a view of society as a stabile arrangement of parts that function together to achieve harmony, or some other social good.12 And, the normativity of a functional part is instrumental and derives from the end or goal to be achieved or the larger whole to which the part contributes. This is not the idea of functional normativity that I borrow from Aristotle. As I mentioned above, the Afunction’s possible contribution to a larger social good – the function’s instrumental value – cannot be used to explain its intrinsic normativity, i.e. the local standards or expertise that makes one a good practitioner.13 And this is for two reasons. First, these forms of activity presuppose the existence of social cooperation, and so their normativity cannot be explained simply in terms of their contribution to a larger social good, like social cooperation and stability – on pain of circularity. And, second, the local standards or norms that partially constitute the activity may, in fact, not promote larger goals like social cooperation and stability. The two sources of value come apart in some cases. To call someone “an excellent cook” refers to local standards pertaining to the activity and product of cooking – tastes good, nutritious, economical – rather than to the instrumental value that the activity of cooking might have (or might not have) as it contributes to a larger societal good like the economy or the environment. Recall Aristotle’s words “the good and the well resides in the function.” These local norms, or expertise/techniques, are independent of whether or not, in addition, the function or work has instrumental value and contributes to the stability of society as a whole or some other larger social good. There is an extensive philosophical literature on the normativity of biological and artifactual functions (Krohs and Kroes 2009). In the philosophy of biology, the central issue is how to understand the normativity of the parts of organisms like the heart. What does it mean to say that the heart has a function, something it ought to do? Note that the heart’s function is neither what is actually does (some hearts malfunction) nor what it could do (in fact, some hearts could never circulate blood) but rather what it is to do or ought to do (circulate blood). A similar issue arises for artifacts, which also seem to have functional normativity that is difficult to explain in causal terms. An artificial heart has the function of circulating blood; that is, what it ought to do, not simply what it can do or is doing. While the question addressed here with regard to organisms and artifacts is the same as the one I pose concerning social role normativity, namely what is the source of the normativity, the range and nature of the entities clearly differs. Just as there are important and relevant differences between organisms and artifacts, so too, social roles differ from both in significant ways. One key difference concerns the fact that social roles are ways of being human; they are ways of enacting our social agency. And so social role normativity raises a host of issues concerning what it is for an agent to respond to a norm and what it is for an agent to be evaluated by others under a norm that I group together

What Explains Social Role Normativity?  131 under the dichotomy of voluntarism and ascriptivism. Corresponding to the complexity attaching to social role normativity is a rich set of resources to explain or to ground that normativity (the notions of expertise and technique) and these resources are not directly relevant to either organisms or artifacts, and hence are not available to ground or explain their functional normativity.

Afunctions and Ascriptivism In addition to providing a source of normativity for the ascriptivist, Afunctions, as exemplified by arts like flute playing and crafts like housebuilding, have a number of features that are useful for the goal of understanding social role normativity. First, notice that whether the housebuilder or the flautist performs their respective function well or badly is not up to them to determine. And I mean this in two ways. First, they do not (directly) create the standards (the expertise and techniques) for their arts or crafts, and second, they are not the final arbiters of whether or not, on a given occasion, they have performed well or badly in relation to those standards, although they may, of course, have an opinion on both issues, and there may be disagreement about both as well. Second, like other social roles, arts and crafts exist in a complex social context – of tools, practices, institutions, and the like – and are systematically related to these entities and contexts – and to one another. The butcher, the baker, the candlestick maker are artisans whose activities are intermingled with one another, with material objects, with practices, with economies, and with a web of larger social institutions. Third, arts and crafts, like other social roles, vary from culture to culture and are open to change over time while at the same time incorporating the intrinsic normativity captured by notions of expertise and technique. In other words, artisanal social roles embody and display the tension between the social role’s intrinsic normativity (after all there is expertise and technique) and the possibility of critique, revision, and change (the norms that currently shape expertise and technique are open to revision). And, as I mentioned earlier, articulating and explaining this tension is a requirement on an adequate understanding of social role normativity.

Conclusion Our thinking about social role normativity is facilitated by borrowing the notion of an Afunction from Aristotle. The Afunction is particularly helpful as a model for explaining noninstrumental social normativity because Aristotle did not center his concept of function on the functionality of tools whose normativity is clearly instrumental. Rather, Aristotle centers his exposition of what a function is on the examples of artisans

132  Charlotte Witt and artists, who are the occupants of social roles and whose actions and activities are normatively evaluable in relation to their identities as artisans and as artists. Importantly, this normativity is not primarily instrumental, but rather it is rooted in the artisan’s identity, in their expertise, and in their excellence in being what they are.

Notes





Michael Hardimon develops an ascriptive view of some social role obligations that differentiates between contractual and noncontractual social roles: “It is perhaps worth stating explicitly that I am assuming here that the roles of family member and citizen exhaust the class of noncontractual social roles” (1994, 347).



What Explains Social Role Normativity?  133

12 Along the same lines Rachel Barney points to the “architectonic” aspects of Aristotle’s view of human society as reflected in the opening of the Nicomachean Ethics as a possible interpretation of the function argument. As Barney notes, the architectonic interpretation of the function argument is compatible with what she calls the “realization” interpretation (2008, 318).

References “Function.” In The Compact Edition of the Oxford English Dictionary, 1971 ed. Oxford: Oxford University Press. Anderson, Elizabeth. 2000. “Beyond Homo Economicus: New Developments in Theories of Social Norms.” Philosophy and Public Affairs 29, no. 2 (Spring): 170–200. Barnes, Jonathan (ed.). 1984. Aristotle: The Revised Oxford Translation (2 vols). Princeton, NJ: Princeton University Press. Barney, Rachel. 2008. “Aristotle’s Argument for a Human Function.” Oxford Studies in Ancient Philosophy 34: 293–322. Bicchieri, Cristina, Ryan Muldoon, and Alessandro Sontuoso. “Social Norms.” The Stanford Encyclopedia of Philosophy (Winter 2018 Edition), edited by Edward N. Zalta. https://plato.stanford.edu/archives/win2018/entries/ social-norms/. Broadie, Sarah Waterlow. 1994. Ethics with Aristotle. Oxford: Oxford University Press. Hardcastle, Valerie. 2002. “On the Normativity of Functions.” In Functions: New Essays in the Philosophy of Psychology and Biology, edited by Andre Ariew, Robert Cummins, and Mark Pearlman, 144–155. Oxford: Oxford University Press. Hardimon, Michael O. 1994. “Role Obligations.” The Journal of Philosophy 91, no. 7 (July): 333–363. Korsgaard, Christine. 2003. “Personal Identity and the Unity of Agency.” In Martin and Baressi. Krohs, Ulrich, and Peter Kroes. 2009. Functions in Biological and Artificial Worlds. Cambridge, MA: MIT Press. MacIntyre, Alasdair. 1981. After Virtue. Notre Dame, IL: Notre Dame University Press.

134  Charlotte Witt Mallon, Ron. 2003. “Social Construction, Social Roles and Stability.” In Socializing Metaphysics, edited by Frederick F. Schmitt, 327–355. Lanham, MD: Rowman and Littlefield. ———. 2016. The Construction of Human Kinds. Oxford: Oxford University Press. Martin, Raymond and John Baressi. 2003. Personal Identity. Oxford: Blackwell. McLaughlin, Peter. 2000. What Functions Explain. Cambridge: Cambridge University Press. ———. 2009. “Functions and Norms.” In Functions in Biological and Artificial Worlds, edited by Ulrich Krohs and Peter Kroes. Cambridge, MA: MIT Press: 93–103. Okrent, Mark. 2018. Nature and Normativity: Biology, Teleology and Meaning. New York: Routledge. Rouse, Joseph. 2017. The Routledge Handbook of Philosophy of the Social Mind. New York: Routledge. Witt, Charlotte. 2011. The Metaphysics of Gender. Oxford: Oxford University Press. ———. 2015. “In Defense of the Craft Analogy: Artifacts and Natural Teleology.” In Aristotle’s Physics: A Critical Guide, edited by Mariska Leunissen, 107–121. Cambridge: Cambridge University Press.

7

The Social Function of Morality Andreas Müller

Introduction This chapter discusses various attempts at deriving metaethical conclusions from claims about the function of morality. In doing so, I will, for the most part, grant the truth of such function claims and focus on what metaethical theses they do and do not support.1 Metaethicists do not (primarily) seek to defend specific ethical theories, such as consequentialism, or to justify particular ethical judgments, e.g. about the moral status of abortion. Instead, they are concerned with the nature and possibility of ethical truth and knowledge in general: what does it take for an ethical judgment to be true, are there any ethical truths, and if so, how can we gain knowledge of them? My question here is to what extent insights into the function of morality can help us answer these questions. Recently, various authors have argued that understanding how our moral beliefs and practices are shaped towards performing certain biological or social functions should lead us to accept skeptical or negative answers to these questions. According to Sharon Street (2006), for example, such insights show that we have to reject a realist conception of moral truth; according to Richard Joyce (2006), they even show that moral knowledge is entirely unattainable. Such debunking arguments have been the subject of extensive debate (Sauer 2018). Other authors, most notably Philip Kitcher (2011) and Philip Pettit (2018), try to put insights into morality’s function to a more constructive use. They argue that a proper understanding of morality’s function not only supports the idea that there are moral truths that we can gain knowledge of but also provides us with a better understanding of the grounds of those truths, an understanding, moreover, that reconciles the existence of such truths with a naturalistic worldview. In what follows, I will first survey these and similar positions. My main focus in doing so will be on the constructive, vindicatory project, rather than the debunking one. Thus, the second half of the chapter zooms in on Pettit’s and Kitcher’s proposals. A critical discussion particularly of the latter one will reveal that while a promising vindicatory role for a function claim in metaethics is in elaborating a nonrepresentationalist

136  Andreas Müller conception of the truth of ethical judgments, grounding ethical truth in morality’s function in that way allows for naturalizing ethics in a modest sense at the most.

The Variety of Function Claims As I said, I will not be questioning the claims about the function of morality that those various authors accept, nor will I discuss whether and how they support those claims. Nevertheless, it will be instructive to have a look at what exactly they have in mind when they ascribe a function to morality. That is because very different things can be asserted by making such a claim, and the claims that we find in the literature are often somewhat vague about or differ on important aspects. A first dimension along which function claims can differ is the very concept of function that is being employed. What do we mean when we talk about something’s function? According to the causal conception, ascribing a function to an entity amounts to ascribing a certain causal role to it, a role that it plays within a larger system of which it is a part (Cummins 1975). For example, the function of a bike’s chain is to transmit the force from the cranks to the rear wheel. According to the etiological conception, on the other hand, the function of an entity is a feature that explains why that entity exists or persists (Wright 1973). For example, the human heart’s function is to pump blood through the circulatory system. It also makes a thumping sound that indicates excitement or exertion, but that is not a function of the heart because that is not part of the explanation why humans have one. Whereas the ascription of a causal function requires the identification of a causal system of which the function bearer is a part, the ascription of an etiological function requires the identification of a feedback mechanism that connects the presence of the relevant feature (e.g., the pumping of blood) in previous instances of the function bearer to the existence or persistence of current instances. Candidates for the former systems include machines and organisms. The best-known feedback mechanism underlying etiological functions is Darwinian natural selection, but there are others, such as those involved in cultural evolution. 2 The claim that our moral beliefs have a certain function can be understood in terms of either of these conceptions.3 It might point to these beliefs’ role in causing certain types of behavior, for example, or it might amount to claiming that the fact that we have those beliefs can be explained as the result of a process of natural or cultural selection. Most of the authors who seek to draw metaethical conclusions from insights into the function of morality focus on those insights that they take to be provided by the evolutionary or social-history genealogy of morality. Even though they are not always explicit about this,4 they are thus employing the concept of function according to the etiological conception.

The Social Function of Morality  137 Another respect in which claims about the function of morality can differ, and in which the claims endorsed in the literature do in fact differ, is what is considered to be the relevant function bearer. What entity exactly is the function ascribed to? One obvious candidate are types of behavior, such as altruistic behavior, or our motivational dispositions to engage in such behavior. Another candidate for the relevant function bearer, which some recent authors have focused on, are our moral and evaluative judgments, which Street claims are “thoroughly saturated with evolutionary influence” (Street 2006, 114). Other authors are concerned with the function of our moral concepts or, relatedly, of our capacity to make moral judgments (Joyce 2006, Chapter 3; Pettit 2018, Chapter 5). Kitcher, too, is concerned with the function of a specific capacity. But the one he focusses on – the capacity for “normative guidance” (Kitcher 2011, 74) – is not, at least not in the first instance, a capacity to make certain judgments. Instead, it enables a subject to obey the commands encoded in social rules. Importantly, when it comes to the function that Kitcher takes to be distinctive of morality – the remedying of failures to engage in altruistic behavior (2011, 103) – the bearer of that function is the capacity for normative guidance in combination with certain social rules (viz. those that promote altruistic behavior). On Kitcher’s view, the bearer of the relevant function is thus more complex than on many other views: it includes both a capacity and certain rules that serve as the input to that capacity. Finally, claims about the function of morality can differ in the exact purpose that they ascribe to morality. What end is promoted by morality when it satisfies its function? One possible answer here is, of course, the increasing of reproductive fitness by improving one’s chances of survival and reproduction, which is the end that ultimately drives natural selection. Most authors, however, do not focus on this end. Instead, they take the (or at least a) function of morality to be the avoidance of weakness of will, the enabling of interpersonal commitment and reliance (e.g., Joyce 2006, Chapter 4; see also Pettit 2018), the furthering of mutually advantageous cooperation in groups, or the remedying of failures to engage in altruistic behavior via normative guidance (Kitcher 2011). However, not everyone identifies such an optimistic purpose. Some argue instead that the function of social norms and normative guidance has been to create or stabilize relationships of power and subordination, to enable exploitation and oppression (Sterelny 2012). As this brief survey already indicates, different authors can and do have rather different things in mind when they appeal to the function of morality. One further distinction is worth highlighting at this point. Some authors, such as Street (2006), ascribe what we might call a biological function to morality. That is, they take morality to serve a biological purpose, such as reproduction and survival. And, insofar as they adopt an etiological conception of function (as Street does), they take the

138  Andreas Müller relevant feedback mechanism to be a biological one, too, viz. Darwinian natural selection. In contrast, many other authors have a social function in mind. They take the purpose that is furthered by morality to be a social one, such as cooperation or remedying altruism failures, and they emphasize the importance of social feedback mechanisms, in particular those involved in cultural evolution (Joyce 2006, 40–43; Kitcher 2011, Chapter 3). As we will see below, this focus on the social function of morality allows metaethicists, particularly those seeking to ground the possibility of moral truth in that function, to avoid certain problems.

Debunking, Vindicating, and the Naturalistic Challenge Claims about the function of morality can be used to support a variety of specific metaethical theses and positions. Nevertheless, the views of those who appeal to such claims are typically aligned with one of two broader projects. The first of these projects is to debunk morality, or at least a conception of morality according to which morality is a domain in which truth and knowledge can and should be sought. Those pursuing the debunking project seek to undermine the idea that our moral judgments can be justified or true by uncovering their origin. More specifically, they attempt to show that the truth of these judgments does not play any role in their genealogy, that is, in the explanation of how we came to have them, and that their genealogy does not support the idea that those judgments stem from a process that reliably produces judgments that are true. Debunkers, in other words, provide a “truth-mooting genealogy” for our moral judgments (Mason 2010, 773). Their argument for the conclusion that we are never entitled to regard our moral judgments as justified thus involves two premises (see Kahane 2011, 106). The first premise states that our moral judgments can be explained as the result of a certain process. The second premise classifies this process as epistemically defective because it is not truth-tracking. Richard Joyce (2006) is a prominent proponent of this debunking project. Sharon Street (2006), too, argues that the origin of our moral judgments entails that they are never justified, on the assumption that their truth is understood along robustly realist lines, that is, as correspondence with a mind-independent realm of facts sui generis. Unlike Joyce, however, she does not advocate that we embrace moral skepticism, but turns this into an argument against those forms of moral realism. Claims about the function of morality play an important role in the case that debunkers offer for their first premise because the explanation they take our moral judgments to be amenable to is typically a functional explanation: we have those judgments, not because they are (likely to be) true, but because they are useful in promoting an end that has nothing to do with their truth. This will be most plausible if the function

The Social Function of Morality  139 that is ascribed to morality is a biological one. If our moral judgments are, at least to a large extent, the result of Darwinian processes that are selected for them because they provide subjects with a reproductive advantage – if they are “thoroughly saturated with evolutionary influence” (Street 2006, 114) – then it would be an unlikely accident if a significant number of them turned out to be true, especially if that required correspondence to certain mind-independent facts. The second project that might be pursued by those who appeal to the metaethical significance of function claims is to vindicate morality. The aim of those pursuing the vindicating project is the opposite of the debunker’s aim. Vindicators seek to show that the ideas of moral truth and moral knowledge are not undermined by a proper understanding of morality’s function. Rather, such an understanding allows us to better explain how moral truth and moral knowledge are possible; it thus supports those ideas. We can further distinguish between weak and strong vindication. Weak vindication involves showing that, contrary to what the debunker claims, the function of the target beliefs does not stand in the way of truth and knowledge but is, at least in principle, compatible with it. Strong vindication goes beyond the negative project of refuting the debunker. It involves showing that morality’s function helps us to explain what moral truth consists in. According to strong vindicators, claims about the function of morality thus positively supports the idea that some moral judgments are true. Rejecting the debunker’s skeptical challenge to the possibility of moral knowledge only requires weak vindication. Nevertheless, many proponents of the vindicating project who focus on the function of morality seek to provide a strong vindication of moral truth in terms of that function. That is because they are also – and perhaps primarily – concerned with a different challenge, one that is faced by anyone who accepts philosophical naturalism. According to that view, the natural world – that is, roughly, the world as it is described by the natural sciences – is all there is, there are not super- or nonnatural properties or facts. But it seems that the truth of moral judgments requires the existence of facts that are not part of the natural world, facts concerning the goodness or wrongness of certain acts, for example. These lie beyond the purview of the natural sciences. Naturalism thus appears to conflict with the possibility of moral truth because the latter seems to require the existence of a kind of fact that is ruled out by the naturalistic view of the world. Both Kitcher (2011, 3–4) and Pettit (2018, 17–18) present the overcoming of this naturalistic challenge as one of their main goals, and both rely on claims about the social function of morality in their pursuit of this goal (albeit in rather different ways, as we will see below). In the remainder of this chapter, I will focus on this project of providing a strong vindication of morality that meets the naturalistic challenge, and the role that claims about morality’s function can play in it.

140  Andreas Müller That is partly because this strikes me as the most ambitious and hence the most interesting attempt at deriving metaethical conclusions from claims about the function of morality. Moreover, the debunking project has been discussed – and criticized – extensively over the last couple of years (see, e.g., Kahane 2011; FitzPatrick 2015; Vavova 2015; Sauer 2018), whereas attempts at function-based vindication have attracted less attention. Lastly, if those attempts turn out to be successful, they will also amount to a rebuttal of the debunker’s skeptical argument. For since function claims are in principle epistemically accessible, grounding moral truths in such claims brings them within the reach of our epistemic capacities, too. So what does the function of morality have to do with the truth of moral judgments? We can distinguish three general approaches to answering this question. According to the first approach, the facts that true moral judgments accurately represent can be reduced to facts concerning morality’s function. This approach thus proposes a naturalistic reduction of moral truths, with function facts serving as the reductive basis. The second approach also adduces claims about the function of morality in support of a naturalistic reduction of moral truths. But on this approach, function facts do not provide the reductive basis. Instead, the role of function claims is more indirect: identifying the function of moral concepts helps us understand that they are used to refer to other naturalistically unsuspicious properties or facts (i.e., ones that do not concern functions). The third approach, like the first approach, takes function facts to provide the ground for the truth of moral judgments. But it does so within an antirealist framework. On this approach, moral judgments are not true because they accurately represent anything, so function facts make them true in some other way than grounding facts that are accurately represented by them. In the following sections, we will discuss each of these approaches in turn.

Reductive Naturalism and Evolutionary Ethics The first approach to establishing a connection between moral truth and morality’s function seeks to reduce the former to (truths about) the latter. A proponent of this approach might argue, for example, that a certain type of action, such as fidelity, is morally good because it serves a certain function, such as the (biological) function to increase reproductive fitness or the (social) function of facilitating cooperation within the group. More specifically, the idea is that facts of the former kind – an action’s being good or bad, right or wrong, a character trait being virtuous, etc. – can be reduced to facts of the latter kind – facts about the functionality of that type of behavior. Thus, the truth of judgments that purport to accurately represent facts of the former kind actually only requires the existence of facts of the latter kind. Insofar as these latter

The Social Function of Morality  141 facts about functionality are deemed unproblematic from a naturalistic point of view, such a view hence offers a straightforward answer to the naturalistic challenge. Nevertheless, at least in the current metaethical literature, this first approach is by far the least common one. It is often associated with Herbert Spencer (1879) and, more generally, with the project of biologizing ethics, that is, bringing moral questions within the scope of evolutionary biology (Wilson 1978; Ruse and Wilson 1986; see also Krebs 2005). G.E. Moore (1903, 100–101) characterized such a view, which he found entailed by – though not explicitly endorsed in – Spencer’s writings, as claiming that the moral goodness of a kind of conduct consists in its being highly evolved, that is, its satisfying the function of increasing reproductive fitness especially well. By investigating the evolution of human behavior and thus discovering “the direction in which we are developing,” biologists could thus determine “the direction in which we ought to develop” (Moore 1903, 97). The project of developing such an “evolutionary ethics” also has more recent advocates. In an article entitled “Defense of Evolutionary Ethics,” for example, Robert Richards argues that “[s]ince each man has evolved to advance the community good, each ought to act altruistically” (1986, 288). Another example is William Casebeer, who proposes a “fully naturalized conception of morality” according to which “morality [is] a matter of proper function,” where the relevant function is “at least partially fixed by our evolutionary history,” and “[m]oral claims should be reduced to functional claims,” which “should be treated as they are in biology and the life sciences” (2003, 3–4). This kind of view faces a number of widely acknowledged problems, and it is probably more interesting as an illustrative contrast foil, rather than a serious contender in contemporary metaethics. One main problem has to do with the extensional worries to which such views give rise. By effectively reducing moral questions to questions about functions, these views are committed to substantive answer to the former questions, that is, accepting them entails accepting certain first-order ethical claims about the morality of, say, meat-eating, killing, rape, etc. And because those views typically focus on biological, specifically evolutionary, functions, some of those commitments deviate significantly from our considered moral judgments on those matters. For example, this kind of reductive naturalism will have a hard time accommodating the categorical wrongness of nonconsensual sex, or the moral scruples many people have regarding the practice of killing animals for food. A second problem with the view under consideration is that it fails to capture the characteristic authority that comes with moral requirements and the judgments that express them. Unlike, for example, the requirements of etiquette, the bindingness of moral requirement does not depend on the agent’s accepting or wanting to conform to them. Or at least, they purport to have that inescapable authority, so that, from the perspective of

142  Andreas Müller those they address, moral requirements come with a distinctive practical oomph. But if a kind of behavior’s being morally required is boiled down to its being highly evolved, or enhancing the subject’s change of reproductive success, this distinctive dimension of moral judgments seems to get lost, since facts of this latter kind do not appear to have a comparable practical authority. The two problems that I have just outlined can be seen as two aspects of a more general difficulty that advocates of any reductive naturalism about morality face. In order to meet the naturalistic challenge, they need to argue that the naturalistically unsuspicious domain of truths that they propose as a reductive base are indeed what we are looking for when we are looking for moral truths. That means that these truths need to line up sufficiently well with those cases in which we are reasonably sure that we have identified a moral truth. And it means that they must offer some basis for the distinctive features of moral truths, such as their practical authority.

Pettit’s Genealogical Argument for Naturalist Realism Philip Pettit (2018) also argues for a naturalist realism about morality, that is, for the view that our moral concepts refer to properties that are unproblematic from a naturalistic point of view. Function claims are at the core of his argument, too, but they play a very different role in his account than they do in the account we discussed in the previous section. Pettit does not claim that moral claims can be reduced to function claims, or that moral concepts refer to functional properties. In fact, according to his account, function claims play no role in grounding the truth of any particular moral judgment. Nevertheless, his argument for naturalist realism crucially relies on a claim about the function of moral concepts. In a nutshell, Pettit argues that a proper understanding of the function of our moral concepts (he focusses on the concepts of desirability and responsibility) supports naturalist realism because it shows us that those concepts can fulfil their function without referring to anything other than naturalistically unproblematic properties. However, Pettit does not tell us which naturalistic property our moral concepts refer to (2018, 25). That is because he does not formulate a reductive analysis of moral concepts that states necessary and sufficient conditions for their correct application. The function-based argument is offered as an alternative to this standard philosophical methodology (2018, 20–25). Pettit suggests that if we can give a functional explanation of the emergence of a concept in a context that does not involve anything naturalistically unacceptable, then we can infer that this concept refers to a naturalistic property: since the appearance of those concepts … is naturalistically explicable, the properties they ascribe … must be naturalistic, too; if the

The Social Function of Morality  143 concepts ascribed non-naturalistic properties, after all, then those properties would presumably have played a role in explaining how the concepts came into use. (Pettit 2018, 22) Pettit does not aim to show this for our moral concepts directly – he does not offer a historical account of how those concepts came into use by our ancestors. Instead, Pettit wants to show how the emergence of concepts that are very similar to ours could be explained in a hypothetical naturalistic world he calls “Erewhon,” an anagram of “nowhere” and a reference to Samuel Butler’s 1872 work of that title.5 The insights gained about those concepts are then transferred to our actual moral concepts: insofar as the terms or concepts that emerge in the story [about Erewhon] respond to the same sort of prompts, and serve the same sorts of purposes, as our actual ethical terms, the properties they predicate are good candidates for the properties we ourselves predicate with our terms. (Pettit 2018, 22) Because our actual moral concepts have the same functional role as the concepts in Erewhon (they “respond to the same sort of prompts, and serve the same sorts of purposes”), it is plausible that they also refer to the same – naturalistically unproblematic – properties. Pettit’s argument for naturalist realism about moral concepts is thus indirect in two respects. It is indirect because the thesis that our actual moral concepts refer to naturalistic properties is derived from an analogous thesis about certain hypothetical concepts. And it is indirect because Pettit does not actually identify the naturalistic properties those concepts refer to but supports the latter thesis about the Erewhon concepts by offering a naturalistic explanation of their emergence. Because of this indirectness, Pettit’s argument does not commit him to any first-order moral claims, e.g. about the desirability of specific outcomes.6 That makes his version of naturalist realism much less vulnerable to extensional objections than the reductionist one considered in the previous section. Moreover, Pettit pays attention to the fact that moral concepts are prescriptive concepts, as well as to the special authority of the judgments in which they are used (2018, 149–153), and his story about the emergence of the analogous concepts in Erewhon is meant to explain how Erewhon’s residents come to use concepts with those very features. Pettit’s account thus promises to fare much better when it comes to the second problem faced by functionalist reductive naturalism, too. But is his account successful in meeting the naturalistic challenge by establishing that moral concepts refer to naturalistic properties? That depends on two conditions. First, the concepts whose emergence

144  Andreas Müller is reconstructed in Erewhon must be sufficiently similar to our actual moral concepts to warrant the transferal of insights into their semantic features to the case of the latter. Second, the explanation that is offered for the emergence of the concepts in Erewhon must only rely on naturalistic resources. In particular, the hypothetical conditions that give rise to their emergence must not presuppose anything that is problematic from a naturalistic point of view. Both conditions will attract skepticism by those who are not already convinced that some version of naturalist realism must be correct. The first condition presupposes that our actual moral concepts resemble the Erewhon concepts in that their cognitive job is to “respond to” to the presence of certain properties, so that the judgments in which they are used serve to indicate that certain states of affairs obtain. But such a representationalist conception of the nature of moral judgments is rejected both by expressivists (Gibbard 1990; Blackburn 1998) and by constructivists (Bagnoli 2002; Korsgaard 2003; James 2007; Müller forthcoming) in metaethics, as well as by some fellow vindicators, including Kitcher (2011, Chapter 5). Pettit briefly addresses expressivism as an alternative to his view, but sets it aside because it is “downgrading ethical discourse” in a way that he wants to avoid (2018, 19). But expressivists will of course reject this allegation, which Pettit does not elaborate further. Constructivism is not considered by Pettit at all, even though it certainly does not merit a similar allegation, since it neither disputes that our moral judgments can be true in a substantive (i.e., nondeflationist) sense nor that some of them are indeed true. Regarding the second condition, there is reason to doubt that Pettit’s characterization of the initial conditions in Erewhon – that is, the conditions that obtain before the emergence of prescriptive concepts and that are supposed to provide the resources for explaining their emergence – is thoroughly naturalistic. Pettit assumes the agents in that scenario to have the ability to engage in practical reasoning and deliberation (2018, 48–49). For example, they are able to reason from the desire to stay dry during their walk to the train station and the belief that this requires taking an umbrella to the intention to take an umbrella. But if the initial conditions in Erewhon permit the residents to engage in such reasoning, then Pettit has to assume that they can have that ability without having any prescriptive or other normative concepts. After all, it is the emergence of such concepts that the story is intended to offer a plausible explanation of. But there are very good reasons to think that practical reasoning already requires the ability to make certain normative judgments. It is a natural thought that reasoning involves taking something to be a reason. According to a plausible account of reasoning that builds on this idea, every episode of practical reasoning involves a judgment according to which the consideration from which that episode starts (that staying dry requires taking an umbrella) provides a reason for its

The Social Function of Morality  145 conclusion (the intention to take an umbrella). This distinguishes reasoning from other, merely causal processes of attitude formation, such as association, which are not guided by a normative judgment. Though admittedly not uncontroversial, such an account of practical reasoning offers many advantages over competing accounts.7 By simply assuming it to be incorrect, Pettit’s story about the emergence of prescriptive concepts incurs significant theoretical costs right from the beginning. The reservations outlined in the previous paragraphs do not show that Pettit’s argument suffers from a fatal flaw, nor does it suggest that his view is internally inconsistent. But it provides some ground for questioning how compelling Pettit’s response to the naturalistic challenge is to anyone who does not already accept the metaethical view that he seeks to establish. This brings us to the third approach to employing claims about the function of morality in vindicating the possibility of moral truth in a naturalistic world, which has been promoted by Philip Kitcher. Since this is perhaps the most unusual but also, in my option, the most promising approach, I will discuss it in somewhat greater length than the previous two options.

Kitcher on Progress and Truth in Ethics Like the proponents of the views discussed in the previous two sections, Kitcher believes that moral statements can be true or false. He also believes that insights into the function of morality can help us explain what the truth of a moral statement consists in. Unlike Pettit and others, however, he does not think that moral judgments like “Lying is wrong” are true in virtue of correctly representing an action to have a certain property. Kitcher advocates a conception of the truth of moral statements that differs fundamentally from the idea that truth consists in the accurate representation of particular facts. Moral statements, he suggests, correspond to ethical rules which issue certain commands. They are the “descriptive counterparts” of such rules (Kitcher 2011, 186). For example, the statement “Lying is wrong” corresponds to a rule that commands us not to lie. Kitcher’s proposal is that ethical statements, understood as descriptive counterparts of ethical rules, count as true just in case those rules would be adopted in ethical codes as the result of progressive transitions and would be retained through an indefinite sequence of further progressive transitions. (Kitcher 2011, 246) This needs some unpacking. Ethical rules are elements of ethical codes. Such codes can thus be changed by adding a new rule or eliminating an old one.8 Crucially, such changes to the ethical code can constitute progressive transitions. Not all changes are progressive – there are also

146  Andreas Müller regressive ones. Now suppose that introducing some rule into the ethical code constitutes progressive transition. Suppose further that no future change to the code which involves eliminating that rule will ever qualify as progressive. That is, if all future changes to the code were progressive, the rule would never be eliminated. The descriptive counterpart of such a rule, and only the descriptive counterpart of such a rule, would count as true according to Kitcher’s proposal. Clearly, the notion of a progressive transition is doing important work here. Kitcher argues that ethical truth can be explained in terms of ethical progress. He takes the latter to be the more fundamental concept. To avoid circularity, he thus cannot explicate such progress as accumulating truths, or as somehow getting closer to the truth. What, then, makes a change in the ethical code progressive? Here, the function of morality comes into play. Kitcher tells us that “[e]thical progress consists in functional refinement” (2011, 221). His conception of ethical progress is modelled on a conception of technological progress; indeed, Kitcher takes ethical rules and our ability to comply with them to be a “social technology” (2011, 221). He spells out this analogy as follows. Technological artifacts serve a certain function. The function of a toaster, for example, is to toast bread. And when it comes to toasters, functional refinement constitutes progress: if toasters are changed so that they perform their function more reliably, more quickly, or with less effort for the user, then this change can be meaningfully characterized as a progressive one. Ethical codes have a function, too: to remedy altruism failures and thereby to promote social harmony (2011, 225).9 And just like in the case of technological artifacts, a change to the code constitutes progress if the new code performs that function more thoroughly, more reliably, and with less effort (2011, 221). Combining this conception of ethical progress with Kitcher’s account of ethical truth yields that a moral statement counts as true just in case adding the corresponding rule to the ethical code improves the code’s functionality, and as long as all future changes to the code were to also improve its functionality, the rule would not be eliminated again. We can call such a rule robustly functional. Then, a moral statement counts as true just in case it is the descriptive counterpart of a robustly functional ethical rule. On Kitcher’s account, the truth of moral statements is thus ultimately grounded in the truth of certain function claims. Does a functionality-based account of ethical truth along the lines of Kitcher’s proposal offer a basis for a promising metaethical view? Insofar as function claims are unproblematic from a naturalistic point of view, the view promises to meet the naturalistic challenge. Moreover, because it takes moral truths to be grounded in a social rather than a biological function, it is less vulnerable to extensional objections than the views discussed above in the section “Reductive Naturalism and

The Social Function of Morality  147 Evolutionary Ethics”.10 But the view also raises a number of questions, not least because Kitcher lays out some of the crucial parts of his view rather swiftly. In the following sections, I will address the two questions that strike me as the most pressing ones. In doing so, I will try to elaborate some aspects of Kitcher’s view in a way that goes some way towards allaying the worries underlying those questions, but I will also highlight some unresolved issues.

Function and Truth in Ethics11 Why should we accept Kitcher’s proposal that a moral statement’s truth is somehow grounded in and dependent upon the functionality of an ethical code which includes the rule that corresponds to the statement? After all, such an account of truth – which denies that truth consists in the accurate representation of certain facts – does not seem plausible for statements from other domains. This might make his proposal seem unfounded and ad hoc. Moreover, Kitcher presents only a rough online of his account, and he does little to establish its adequacy as an account of truth for moral judgments, nor does he compare it to other accounts in the literature.12 Nevertheless, Kitcher’s remarks on truth can be fruitfully understood as tying into an important tradition in the theory of truth that goes back at least to Charles S. Peirce and that has found its most comprehensive elaboration and defense in the work of Crispin Wright. Interestingly, Wright also presents his account as a way of accommodating ethical truth without the commitment to robust ethical facts that comes with a representationalist conception of truth (see especially Wright 1995).13 His view is best understood as a response to and improvement of Peirce’s pragmatist theory of truth. Peirce defined truth as “[t]he opinion which is fated to be ultimately agreed to by all who investigate” (1878, 300). This view faces serious problems. Firstly, it imposes an epistemic constraint on truth: at some point, there has to be sufficient justification available to inquirers for accepting every truth. But it seems that there might well be truths that are forever beyond the reach of our epistemic capacities, such as truths concerning events in the distant past. Hence, the epistemic constraint is implausible for some domains of judgment. A second problem concerns the very intelligibility of the account. In particular, the idea of ideal conditions – such as those at the supposed end of inquiry – which ensure that the judgments made under those conditions include all and only true judgments is of questionable coherence.14 The neo-pragmatist account of truth developed by Wright in Truth and Objectivity (1992) and elsewhere is designed to avoid these problems. On his account, a judgment’s truth does not depend on whether it would be justified under certain ideal conditions, but rather on whether

148  Andreas Müller justification will at some point be available and whether some of that justification will remain undefeated upon arbitrarily close and extensive scrutiny: Rather than ask whether a statement would be justified at the limit of ideal empirical investigation, or under ideal empirical circumstances, whatever they are, we can ask whether an ordinary carefully controlled investigation, in advance of attaining any mythical limit, justifies the statement, and whether, once justified, that statement continues to be so no matter how much further information is accumulated. (Wright 1992, 47) Wright’s proposal is continuous with earlier versions of pragmatism about truth in that it understands a statement’s being true in terms of having a certain justificatory status, but he idealizes that status by requiring it to be especially robust, rather than by requiring it to be present in certain counterfactual circumstances. Wright calls this particularly robust form of justification “superassertibility” and characterizes it more explicitly as “the property of being justified by some (in principle accessible) state of information and then remaining justified no matter how that state of information might be enlarged upon or improved” (1992, 47). Kitcher’s proposal is thus remarkably similar to Wright’s, at least when it comes to their basic structure: just like Wright understands truth as an especially robust form of justification, Kitcher characterizes it in terms of an especially robust form of functionality. Like Peirce’s definition, Wright’s account still entails an epistemic constraint on truth. This has important consequences. As we saw above, the constraint is implausible for some domains, such as judgments about the distant past. Wright responds to this by endorsing alethic pluralism: his superassertibility account is meant to apply to the truth of judgments in some domains, but not to others. That is, truth in some domains is realized by superassertibility, whereas in other domains, it might be realized by, e.g., representational accuracy. In particular, the account is applicable only in domains for which the epistemic constraint is plausible on independent grounds. Wright’s example of choice is discourse about the comic: “there seems no sense to be attached to the idea that the comedy of a situation might elude the appreciation even of the most fortunately situated judge” (1992, 58). Unlike Wright, Kitcher does not characterize truth in epistemic terms. But his account does entail a constraint that is structurally similar to Wright’s epistemic constraint, which we might call a functionality constraint: this constraint rules out the truth of a statement whenever the corresponding rule does not increase a code’s performance with respect to the relevant function, just like Wright’s epistemic constraint rules out

The Social Function of Morality  149 the truth of a statement whenever that statement is beyond our epistemic reach. Such a constraint is certainly not plausible for all domains of judgment. Therefore, like Wright, Kitcher is committed to alethic pluralism.15 He thus has to argue on independent grounds that the moral domain is subject to such a functionality constraint, e.g. by showing that there is, as Wright puts it, “no sense attached to the idea” of an action being morally wrong even though adding a rule that prohibits such actions to an ethical code does not improve the code’s functionality. Is it plausible that moral truth is subject to such a functionality constraint? To a significant degree, that depends on which function we have in mind here. According to Kitcher, it is remedying altruism failures. Adopting that proposal, the functionality constraint entails that a statement cannot be true if the corresponding rule does not contribute to remedying altruism failures (or if it even increases such failures). I don’t think that it is obvious that this constraint holds, that it makes no sense to deny it. Whether you find it plausible will depend on your substantive moral views. Thus, whether Kitcher’s account is applicable to moral statements depends on certain moral commitments. We will get back to this dependency and its consequences for the account’s ability to meet the naturalistic challenge below. For now, we can conclude that while the conception of moral truth that Kitcher proposes deviates from the well-known representationalist account, according to which judgments are true in virtue of correctly representing an independent realm of facts, it is not an ad hoc proposal but can be seen as building on a wellestablished position in the theory of truth.

Which Function? Let us grant Kitcher for the moment that the truth of ethical statements can be grounded in the functionality of ethical codes in the way he suggests. This leads us to a second important question: which function is the relevant one for determining the progressiveness of changes to the ethical code and hence the truth of moral statements? Kitcher, of course, gives an answer to that question: the relevant function is remedying altruism failures (as well as certain secondary functions that derive from this main function, see Kitcher 2011, 237–241). But for the purposes of assessing whether Kitcher’s view offers a compelling solution to the naturalistic challenge, it is crucial to have a closer look at how Kitcher argues for the claim that this is the relevant function for progress and truth. After all, remedying altruism failures is not the only function that our ethical codes serve, and it is certainly not the sole force behind changes to those codes. Ethical progress is rare, as Kitcher himself acknowledges (2011, 173–174). Many changes do not improve or even diminish the code’s functionality in remedying altruism failures. But these changes can be explained, too, and at least in some cases,

150  Andreas Müller that explanation will also refer to a function that is better served by the changed code. Examples abound. Ethical codes have served the purpose of establishing or protecting the power that members of one group have over members of another (e.g., by requiring a wife’s submission to the will of her husband). They have also served the purpose of stabilizing and protecting certain ideologies (e.g., by prohibiting the questioning or ridiculing of its core assumptions). There are thus various functions that have been served by changes to the ethical code. Clearly, though, not all of these changes would qualify as progressive. But then, which of these functions is the relevant one for ethical progress and truth? According to Kitcher, it’s the remedying of altruism failures. But what sets this function apart from the other ones and qualifies it as the ground of ethical truth? Kitcher often seems to suggest that remedying altruism failures is the relevant function for ethical truth because it is the original function of ethics (see, e.g., 2011, 221, 225). Characterizing the remedying of altruism failures as the original function of ethics refers to his account of how a genuine ethical practice first emerged among our ancestors (2011, Chapter 2). According to Kitcher, a crucial step in that process was the development of a “capacity for normative guidance” (2011, 69). This capacity is, primarily, a “capacity for following orders,” which involves the “ability to apprehend and obey commands” (2011, 74). It affects a subject’s preferences and can thus reinforce altruistic preferences or override conflicting ones. The capacity for normative guidance enables us to avoid or remedy potential altruism failures “by means of a commitment to following a rule: you obey the command to give weight to the wishes of the other” (2011, 74). Ethical practice thus involves two main components: the capacity for normative guidance and a commitment to a certain set of rules. Together, these two components function to reduce the occurrence of altruism failures, and that is why they emerged. Importantly, the capacity for normative guidance itself is not sufficient for performing this function. It can be combined with any number of rules, including egoistic or repressive ones. Thus, when Kitcher writes that the “products of normative guidance (in its simplest and original form) are desires that issue in behavioral altruism” (2011, 76), what he means must not just be any exercise of the capacity for normative guidance, but an exercise that is guided by a specific set of rules to which the subject is committed: roughly, rules that require altruistic or prohibit harmful behavior. So, according to Kitcher, the relevant function for ethical truth is the original function of ethics. But is it relevant for truth because it is the original function? Various commentators have argued that this would involve an illicit inference from “is” to “ought”: the fact that the emergence of a particular capacity or type of behavior can be explained by the fact that they perform a certain function does not suffice to establish that current exercises of that capacity or instances of that behavior

The Social Function of Morality  151 which satisfy this same function therefore have a special normative, let alone moral status (Meyer 2011; FitzPatrick 2012). After all, various other rule systems, including language, have likely developed initially due to their role in satisfying a certain function, too. Nevertheless, we do not continue to classify changes in those norm systems as progressive when they increase functionality with respect to that initial function, nor do we assign any normative authority to them which resembles that of morality. What Kitcher shows at the most is that the capacity for normative guidance initially emerged, together with a certain set of rules, in order to remedy altruism failures. But that does not entail that these are moral rules, unless we already presuppose that altruism is the or a core aim of morality.16 If you don’t presuppose or even reject that (we will discuss such a challenger presently), then you might well think that the capacity for normative guidance is employed to serve a genuinely moral purpose only once it is combined with a different set of rules. Surprisingly, when he introduces his approach to identifying criteria for moral progress, Kitcher does not actually suggest that remedying altruism failures is the relevant function because it is the function that explains why ethics has emerged in the first place. Instead, the strategy he announces is to (i) identify a number of progressive transitions in the history of ethics (which he does in chapter 4 of the book) and then (ii) determine what they have in common (2011, 213). Kitcher describes this as an “empiricist” strategy for identifying the grounds of ethical progress. But that mischaracterizes his actual argument, which requires moral judgments at both steps. Take a look at the second step first. One thing to note right away is that isolating features that various cases of progressive transitions have in common is not sufficient for identifying the features in virtue of which they qualify as progressive, for these might be features that those transitions share with nonprogressive ones. In that case, bearing those features does not ensure, and therefore cannot ground, the progressiveness of a transition. Kitcher seems to be aware of this, since a paragraph later, he characterizes the second step as “isolating common features of the favored transitions, features revealing why we might be concerned to make ethical changes of this particular kind” (2011, 214 [my emphasis]). But requiring that a feature “might” render a transition an object of our concern is either a nearly vacuous condition (if the “might” simply denotes psychological possibility). Or the condition is to be read as requiring that it would be appropriate (rather than just psychologically possible) to be concerned to make the transition. But that is not an empirical but a normative question. The first step – identifying actual cases of progressive changes in the ethical code – also relies on moral judgments. Kitcher discusses a number of cases at great length, including the emancipation of women and the abolition of chattel slavery (2011, 138–165). But he does not offer any

152  Andreas Müller reasons for selecting those cases to illustrate the existence of progressive transitions. He simply introduces them as “compelling examples of transitions that look progressive” (2011, 139 [my emphasis]). Apparently, Kitcher takes it to be obvious that these are paradigmatic examples of ethical progress. I agree, as most of his readers will. But what we agree with is a moral judgment, albeit an uncontroversial one. We don’t actually see or observe them to be progressive. Insofar as their progressiveness is intuitively clear, the intuitions in question are moral intuitions. Imagine someone with a radically different ethical outlook, perhaps someone who believes that women have a moral duty to submit themselves to the will of men because that corresponds to what he considers to be the “natural order.” That person will identify very different cases as paradigmatic instances of progress or regress and then, following Kitcher’s proposed strategy, end up endorsing a different function as the ground of ethical progress and truth. Does Kitcher have the resources to argue that this person is mistaken about the relevant function? As we saw above, it does not help to point out that it is Kitcher’s preferred function, rather than the function endorsed by his opponent, that explains why ethics emerged in the first place. His opponent can grant that this was the original function of ethics, but insist that what makes for ethical progress is, perhaps among other things, the preservation or restauration of some natural hierarchy among people.17 Note that the character I have described is not a moral skeptic. He does not question the existence of moral truths or the possibility of moral knowledge. Neither does he question that truths which are grounded in functionality in the way Kitcher describes have the authority that is distinctive of moral truths.18 He simply insists that the function that is relevant for ethical progress and truth is one of the other functions that is sometimes furthered by changes in the ethical code. Can Kitcher offer anything to argue that this character is mistaken that is not an implicit or explicit appeal to a moral judgment? Kitcher discusses the challenge posed by Hume’s “sensible knave” (Hume 1751, 81), whom he describes as someone who does not want the satisfaction of his own egoistic desires to be interfered with and who will only conform to the ethical principles of those around him if that furthers his own ends (Kitcher 2011, 274–275). In response to the challenge posed by the sensible knave, Kitcher tries to establish the privileged status of the original function by “[offering] a diagnosis of what he is doing. He wants to take advantage of the products of social evolution without acknowledging the functions that have made those products possible” (2011, 276). But that will not do to refute the challenge posed by our character. For what is “acknowledging” supposed to mean here? If it means to acknowledge remedying altruism failures as the original function and hence as an important part of the explanation of the emergence of ethics, then he will have no problem acknowledging that.

The Social Function of Morality  153 But, as we saw above, he will insist that moral truth is grounded in a different function, one that is satisfied by rules that have perhaps only emerged at a later stage. If, on the other hand, it means acknowledging the relevance or authority of the original function, then this simply begs the question: why should he acknowledge that, if the mere fact that it is the original function does not render it relevant? Kitcher offers an additional response: “Being human, we tell him, consists in participating in this project through which altruism failures are remedied and further ethical functions generated and fulfilled” (2011, 276). That might well be true. But it does not explain the moral relevance of that function, unless we add the – hardly uncontroversial – assumption that our moral norms are grounded in anthropological facts of this kind, an assumption that is certainly not borne out by any empirical investigation. Moreover, note that being human also consists in participating in the project of using language for communication. Yet, we don’t consider the rules of language to have a normative status akin to ethical rules. Thus, Kitcher’s additional response is not sufficient to establish the special moral relevance of his preferred function, either. This shows that Kitcher must rely on substantive moral judgments in making a compelling case for the view he advocates. His (and our) disagreement with the above challenger is, and remains, a moral one. Appeals to moral convictions are thus essential both to his case for the applicability of a functionality-based conception of truth to the domain of moral judgment and to his case for taking the remedying of  altruism failures to be the relevant function for ethical progress and truth. Is that a problem for Kitcher’s overall project? In particular, does it disqualify his view from being a satisfactory response to the naturalistic challenge? In answering this question, it is helpful to distinguish between two forms of moral naturalism: an ambitious and a modest one. Ambitious moral naturalists seek to vindicate and explain the nature of moral truths without relying on any nonnaturalistic claims or intuitions in their argument. In particular, they are not willing to presuppose the truth of any of our moral convictions, unless these convictions can be verified by naturalistically acceptable methods, such as those employed by the natural sciences. If successful, this project would ultimately allow us to derive particular moral claims from a set of nonmoral, naturalistic truths. A satisfactory version of ambitious naturalism would also allow us to adjudicate the disagreement between Kitcher and his challenger on purely naturalistic, scientific grounds. We have seen, however, that Kitcher fails to provide a view that qualifies as ambitious in this sense, as many have before him. Modest moral naturalism, on the other hand, seeks to ultimately ground moral truths and facts in nonmoral, naturalistically unproblematic facts, such as psychological facts or facts about functions. This view

154  Andreas Müller is motivated by a commitment to a certain metaphysical picture. According to this picture, at the most fundamental level, the world only contains naturalistic properties or facts. Importantly, this does not rule out the existence of nonnatural facts, as long as they are grounded in naturalistic facts on a more fundamental level. To show that certain facts are compatible with this version of naturalism, we have to establish that they are ultimately grounded in naturalistic facts. But that does not entail that, in arguing for such a grounding claim, we must only rely on naturalistic claims. Rather, we can, at least in principle, appeal to the plausibility of certain judgments even though they are concerned with facts of the domain that we are trying to ground. A view that relies on certain moral convictions in arguing for the claim that moral facts can be grounded in certain other facts would thus qualify as naturalism in this modest sense, as long as these latter, grounding facts are naturalistic, and as long as those convictions, too, can be understood as owing their truth to the naturalistically grounded moral facts. Kitcher’s view might well be a version of such modest naturalism in metaethics. Whether or not that is enough to vindicate moral truth and reconcile it with a naturalistic worldview depends on whether philosophical naturalists are – and, of course, on whether they should be – content with this more modest form of their project.

Conclusion This chapter has surveyed various attempts at deriving metaethical conclusions from claims about the function of morality. Those attempts are usually made in the context of two broader projects: either they seek to debunk a conception of morality as a domain in which truth and knowledge can be sought, or they seek to vindicate such a conception. Our discussion focused on the latter, which has received significantly less attention in the recent literature than the debunking project. In particular, we discussed three approaches to providing a strong vindication of moral truth that also seeks to meet the naturalistic challenge by appealing to claims about morality’s function. The first of them proposes to reduce truths about the morality of certain types of behavior to truths about their evolutionary function. This approach is often associated with the works of Herbert Spencer and Edward Wilson, but it also has more recent advocates. And while it serves as a useful contrast foil for the other approaches, it is widely discredited due to extensional problems and problems to capture the characteristic authority of moral requirements. The second approach, championed most recently by Philip Pettit (2018), also argues that our moral concepts refer to naturalistically unproblematic properties, but it does not do so by offering a reductive account. Instead, it seeks to establish this conclusion indirectly, by

The Social Function of Morality  155 showing that we can give a functional explanation of how concepts that are very similar to our moral concepts emerge in a hypothetical world that does not include any nonnaturalistic properties that those concepts might be used to refer to. This approach promises to avoid the problems of the first one, but because it makes controversial and unsupported assumptions about the nature of moral judgments and the prerequisites of practical reasoning, it does not offer a compelling case for its preferred metaethical view to anyone who does already accept it. Like the first one, the third approach to providing a strong vindication of moral truth takes it to be grounded in facts about morality’s function. But it denies that moral judgments are true in virtue of accurately representing those, or indeed any other facts. Instead, this approach, which is advocated most prominently by Philip Kitcher (2011), offers an antirealist conception of the truth of moral judgments in terms of the progressiveness of adopting a corresponding ethical rule. Such progress is in turn explicated in terms of the robust functionality of the adopted rule. I argued that this proposal is best seen as a variation of the neo-pragmatist theory of truth developed by Crispin Wright. We also saw that contrary to what Kitcher himself suggests, his argument for identifying the remedying of altruism failures as the relevant function for ethical progress and truth relies on an implicit appeal to substantive moral judgments. Hence, his view qualifies only as a modest form of moral naturalism.

Acknowledgments I am grateful to Rebekka Hufendiek for very helpful discussions and comments on previous versions of this text. I would also like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

Notes



156  Andreas Müller

12 This might be partly due to Kitcher thinking that providing a suitable conception of progress is more important than the notion of truth – at one point, he writes that his overall view retains a notion of ethical truth merely “for explanatory purposes” (2011, 210). But in order to fully meet the naturalistic challenge, it is crucial to accommodate our commitment to the truthaptness of moral statements.





References Bagnoli, Carla. 2002. “Moral Constructivism: A Phenomenological Argument.” Topoi 21, no. 1: 125–138. Blackburn, Simon. 1998. Ruling Passions: A Theory of Practical Reasoning. Oxford: Oxford University Press. Casebeer, William D. 2003. Natural Ethical Facts Evolution, Connectionism, and Moral Cognition. Cambridge, MA: MIT Press. Cummins, Robert. 1975. “Functional Analysis.” Journal of Philosophy 72 (November): 741–764. FitzPatrick, William J. 2012. Review of The Ethical Project, by Philip Kitcher. Ethics 123, no. 1: 167–174. ———. 2015. “Debunking Evolutionary Debunking of Ethical Realism.” Philosophical Studies 172, no. 4: 883–904. Gibbard, Allan. 1990. Wise Choices, Apt Feelings: A Theory of Normative Judgment. Cambridge, MA: Harvard University Press. Hufendiek, Rebekka. “The Evolution of Morality and Morality as an Explanandum.” Unpublished manuscript.

The Social Function of Morality  157 Hume, David. (1751) 1983. An Enquiry Concerning the Principles of Morals. Indianapolis, IN: Hackett. James, Aaron. 2007. “Constructivism about Practical Reasons.” Philosophy and Phenomenological Research 74, no. 2: 302–325. Joyce, Richard. 2006. The Evolution of Morality. Life and Mind. Cambridge, MA: MIT Press. Kahane, Guy. 2011. “Evolutionary Debunking Arguments.” Noûs 45, no. 1: 103–125. Kincaid, Harold. 2007. “Functional Explanation and Evolutionary Social Science.” In Philosophy of Anthropology and Sociology, edited by Stephen P. Turner and Mark W. Risjord, 213–247. Handbook of the Philosophy of Science Series. Amsterdam: North-Holland. Kitcher, Philip. 1993. “Function and Design.” Midwest Studies in Philosophy 18, no. 1: 379–397. ———. 2011. The Ethical Project. Cambridge, MA: Harvard University Press. ———. 2016. “Reply to Smith.” In The Philosophy of Philip Kitcher, edited by Mark Couch and Jessica Pfeifer, 267–272. Oxford: Oxford University Press. Korsgaard, Christine M. 2003. “Realism and Constructivism in TwentiethCentury Moral Philosophy.” Journal of Philosophical Research APA Centennial Supplement: 99–122. Krebs, Dennis. 2005. “The Evolution of Morality.” In The Handbook of Evolutionary Psychology, edited by David M. Buss, 747–771. Hoboken, NJ: Wiley. Manne, Kate. 2018. “Reconsider the Lobster.” Review of 12 Rules for Life, by Jordan Peterson. Times Literary Supplement. Accessed October 08, 2019. www. the-tls.co.uk/articles/public/jordan-peterson-12-rules-kate-manne-review/. Mason, Kelby. 2010. “Debunking Arguments and the Genealogy of Religion and Morality.” Philosophy Compass 5, no. 9: 770–778. McLaughlin, Peter. 2001. What Functions Explain: Functional Explanation and Self-Reproducing Systems. Cambridge: Cambridge University Press. Meyer, Kirsten. 2011. “Naturalismus und moralischer Fortschritt.” In Naturalismus in der Ethik, edited by Thomas Schmidt and Tatjana Tarkian, 69–86. Paderborn: mentis. Moore, G. E. (1903) 1993. Principia Ethica. Edited by Thomas Baldwin. Cambridge: Cambridge University Press. Müller, Andreas. 2019. “Reasoning and Normative Beliefs: Not Too Sophisticated.” Philosophical Explorations 22, no. 1: 1–14. ———. forthcoming. Constructing Practical Reasons. Oxford: Oxford University Press. Peirce, Charles S. 1878. “How to Make Our Ideas Clear.” Popular Science Monthly 12: 286–302. Pettit, Philip. 2018. The Birth of Ethics: Reconstructing the Role and Nature of Morality. Oxford: Oxford University Press. Plantinga, Alvin. 1982. “How to Be an Anti-Realist.” Proceedings and Addresses of the American Philosophical Association 56, no. 1: 47–70. Richards, Robert J. 1986. “A Defense of Evolutionary Ethics.” Biology and Philosophy 1, no. 3: 265–293. Ruse, Michael, and Edward O. Wilson. 1986. “Moral Philosophy as Applied Science.” Philosophy 61, no. 236: 173–192.

158  Andreas Müller Sauer, Hanno. 2018. Debunking Arguments in Ethics. Cambridge: Cambridge University Press. Searle, John. 2010. Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press. Singer, Peter. 2005. “Ethics and Intuitions.” The Journal of Ethics 9, no. 3: 331–352. Smith, Michael. 2016. “Function and Truth in Ethics.” In The Philosophy of Philip Kitcher, edited by Mark Couch and Jessica Pfeifer, 253–267. Oxford: Oxford University Press. Spencer, Herbert. 1879. The Data of Ethics. London: Williams & Norgate. Sterelny, Kim. 2012. “Morality’s Dark Past.” Analyse & Kritik 34, no. 1: 95–115. Street, Sharon. 2006. “A Darwinian Dilemma for Realist Theories of Value.” Philosophical Studies 127, no. 1: 109–166. Vavova, Katia. 2015. “Evolutionary Debunking of Moral Realism.” Philosophy Compass 10, no. 2: 104–116. Wilson, Edward O. 1978. On Human Nature. Cambridge, MA: Harvard University Press. Wright, Crispin. 1992. Truth and Objectivity. Cambridge, MA: Harvard University Press. ———. 1995. “Truth in Ethics.” Ratio 8, no. 3: 209–226. ———. 2000. “Truth as Sort of Epistemic: Putnam’s Peregrinations.” Journal of Philosophy 97, no. 6: 335–364. Wright, Larry. 1973. “Functions.” Philosophical Review 82, no. 2: 139–168.

8

The Function of Gender as a Historical Kind Mari Mikkola

Introduction Feminism is often said to be the movement to end the sexist oppression of women (hooks 2000, 26). We can understand the term ‘woman’ in this claim as a sex term: it picks out human females, and being a human female depends on various biological and anatomical features (like chromosomes or genitalia). However, in response to biologically deterministic accounts that took human anatomy to determine all behavioral, psychological, and sociopolitical features, feminists in the 1960s/1970s began using ‘woman’ differently: not as a sex term, but as a gender term. Being a woman or a man was no longer considered to depend on the kind of anatomy one has but on particular social and cultural factors, like one’s social position or the amount of masculinity/femininity one exhibits. In so doing, feminists distinguished sex (being female/male) from gender (being a woman/a man). This distinction undergirds a major theoretical presumption: that feminist work should orient itself around those who fall under the gender concept woman. Since genders and gender roles were taken to depend on social factors (broadly conceived), and it is the social realm that feminism aims to alter, the gender concept woman became feminism’s defining concept both theoretically and politically. It became commonplace to treat woman as the concept around which feminist politics is organized and the term ‘woman’ as that which picks out the category that makes up feminism’s subject matter. Over the past few decades, the focus on the concept woman has nevertheless become increasingly unsettled. Key challenges are historically and conceptually rooted in (what I term) ‘gender skeptic’ views. First, woman is ultimately meaningless and necessarily open-ended, which renders any attempt to define it futile and even politically counterproductive (Butler 1999). Second, the social kind of women does not exist in any meaningful sense: either its existence is illusory (Butler 1999), or the kind is so hopelessly fragmented that any talk of ‘women’ is meaningless (Spelman 1990). These influential views have played a pivotal role in generating the following clusters of puzzles.

160  Mari Mikkola •

•

Semantic puzzle: Given that ordinary language users tend not to distinguish sex and gender (treating ‘woman’ largely as a sex term, or a mixture of social and biological features), what precisely are feminists talking about when they talk about ‘women’? What, if any, are the necessary and sufficient conditions that the concept woman encodes? Ontological puzzle: How should we understand the category of women that is meant to undergird feminist political solidarity, if there are no necessary and sufficient conceptual conditions underlying our gender talk? Do women make up a genuine kind, or simply a gerrymandered and random collection of individuals?

Feminist philosophers have provided numerous answers to these questions over the past few decades, but precious little agreement exists (Martin 1994; Stoljar 1995, 2011; Frye 1996; Tanesini 1996; Young 1997; Haslanger 2000, 2003a, 2003b, 2005; Heyes 2000; Stone 2004; Zack 2005; Alcoff 2006; Mikkola 2007; Witt 2011a, 2011b; Saul 2012). Moreover, resolving these puzzles is not supposedly just an exercise in philosophy of language and metaphysics. Rather, they are said to generate serious political concerns. Many have found the inability to pick out women’s social kind (feminism’s purported subject matter), and the kind’s subsequent fragmentation problematic for political reasons (for an overview, see Mikkola 2016). The possibility that – strictly speaking – gender kinds might be ontological illusions appears to pull the rug from under feminist politics. As Iris Marion Young put it, unless there is “some sense in which ‘woman’ is the name of a social collective [that feminism represents], there is nothing specific to feminist politics” (1997, 13). Over the past two decades, feminist philosophers have tackled the semantic and ontological puzzles precisely to avoid politically paralyzing consequences. Responding to these worries has been considered a prerequisite for effective feminist politics, and a rich literature has emerged aiming to conceptualize gender (Stoljar 1995, 2011; Frye 1996, 2011; Young 1997; Haslanger 2000; Heyes 2000; Zack 2005; Alcoff 2006; Ásta 2011). In the attempt to resolve this (as Linda Alcoff puts it) “feminist identity crisis,” Theodore Bach (2012) argues for an understanding of gender that comes apart from other contemporary accounts. First, although many feminists implicitly maintain that a system of gender classification (somehow) functions to shape individual social agents and broader social interactions, hardly any explicitly functionalist elucidations of gender exist.1 A functional explanation of (say) a social practice explains why that practice exists by appealing to the purpose or needs served by the practice (Kincaid 2006). So a functional explanation of the existence of gender kinds would involve identifying the needs that are being served by their existence. For Bach, gender

The Function of Gender as Historical Kind  161 kinds function to ensure the continued existence of a broader gender system. Second, contra prevalent views that take women to make up a social kind, Bach’s functionalist solution to the ontological puzzle holds that women make up a natural kind with a historical essence. In short: The essential property of women, in virtue of which an individual is a member of the kind ‘women,’ is participation in a lineage of women. To exemplify this relational property, an individual must be a reproduction of ancestral women, in which case she must have undergone the ontogenetic processes through which a historical gender system replicates women. (Bach 2012, 271) Bach takes his view to be an alternative to social constructivism, although he is a historical constructivist about gender. In other words, Bach thinks that some social kinds are to be incorporated into the framework of natural kinds understood to have historical essences. Bach’s historical constructivism, however, substantially diverges from more typical social constructionist and feminist analyses of gender in that it holds gender should be incorporated into the same ontological framework as species and electrons. This view of gender aims to resolve two clusters of problems for feminists. • •

Representation problem: “if there is no real group of ‘women,’ then it is incoherent to make moral claims and advance political policies on behalf of women” (Bach 2012, 234). Commonality problems: 1 2 3

4

Inseparability: Gender does not exist and develop independently of other (social) features such as race, class, and religion. Universality: There is no feature that all women share across cultures and history. Immutability: “By defining women according to property P it follows that (i) the elimination of P entails the elimination of women, (ii) if an individual possesses P at time 1 and loses P at time 2, then that individual is no longer a woman.” Normativity: Delimiting women’s social kind with the help of some essential property privileges those who possess it, and marginalizes those who do not. (Bach 2012, 235)

On Bach’s view, then, gender kinds serve explanatory and justificatory functions relative to the broader gender system. That is, Bach provides

162  Mari Mikkola a functional explanation of gender kinds, where they exist in order to and because they serve the requirements of the gender system as a whole. Gender kinds are replicated in order to maintain the overall gender system, which is why gender should be thought of as a natural kind with a historical essence. Moreover, this account of gender has a theory-internal role in feminism: historical essentialism can help resolve the above worries in order to prevent the disintegration of feminist politics. Importantly, the account Bach puts forward is allegedly better than available social constructionist alternatives like that of Sally Haslanger’s: historical essentialism is alleged to be explanatorily superior, which makes it politically more powerful. This paper then asks: Is Bach’s functional explanation theoretically compelling in that it provides an analysis of gender that is explanatorily and politically superior to Haslanger’s prominent social constructionist analysis? My answer is, in short, no. First, the story Bach tells about gender combines different components from philosophy of biology and social sciences in a way that renders the account puzzling and opaque. Second, Bach fails by his own lights to offer compelling responses to the problems identified above – that is, he doesn’t resolve the feminist identity crisis. I will begin by outlining Bach’s view (“Gender as a Historical Kind” section). Next, I will consider how the different aspects of Bach’s etiological picture do not hang together as well as one would hope (“How Theoretically Compelling Is Bach’s Functional Analysis?” section). Finally, I will argue that it does not offer a politically compelling metaphysics of gender (“The Political Tenability of Historical Essentialism” section).

Gender as a Historical Kind Alcoff characterizes the typical feminist focus on gender as follows: the concept of woman is “the central concept of feminist theory” in being the necessary point of departure for any feminist theory and feminist politics, predicated as these are on the transformation of women’s lived experience in contemporary culture and the reevaluation of social theory and practice from a woman’s point of view. (2006, 133) A theoretically satisfactory conception of woman should be thick enough to support specific normative demands, which include being able to elucidate how and why patriarchy damages women, to justify critical feminist claims about extant states of affairs, and to ground a vision of a feminist future. An account of what it is to be a woman ought to enable theorists to conceive of feminism’s subject matter by providing a way to

The Function of Gender as Historical Kind  163 conceptualize their social kind or category – perhaps even enabling us to ground feminist political solidarity. Bach terms the prevalent feminist strategy for doing so “social objectivism”, of which Sally Haslanger’s analysis of gender is a prime example. In a series of articles, Haslanger argues for a way to define woman that is meant to be politically useful, serve as a tool in feminist fights against sexism, and that shows woman to be a social (not a biological) notion. More specifically, Haslanger argues that gender is a matter of occupying either a subordinate or a privileged social position. Her analysis is explicitly ameliorative: it aims to elucidate gender concepts feminists should be using in order to identify and explain persistent social inequalities between males and females. With this in mind, Haslanger holds: S is a woman/ man iff [by definition] S is systematically subordinated/ privileged along some dimension (economic, political, legal, social, etc.), and S is ‘marked’ as a target for this treatment by observed or imagined bodily features presumed to be evidence of a female’s/ male’s biological role in reproduction. (Haslanger 2003a, 6–7) This is constitutive of being a woman or a man: what makes calling S a woman apt, is that S is oppressed on sex-marked grounds. Haslanger’s ameliorative analysis is prima facie counterintuitive in that females who are not sex-marked for oppression, do not count as women. However, this is not a reason to reject her analysis as it is meant to be revisionary and not to capture our intuitive gender notions. Haslanger’s proposal is also eliminativist: gender justice would eradicate gender since it would abolish those sexist social structures responsible for sex-marked oppression and privilege. If sexist oppression were to cease, women and men would no longer exist (although there would still be males and females). In arguing for this view, Bach takes Haslanger to make two dialectical moves. First, her social objectivist account defines women “according to a suitably abstract relational property” (Bach 2012, 236) in order to avoid the above commonality problems – that is, the abstract relational property of being systematically subordinated (along some dimension). Second, it employs “an ontologically thin notion of ‘objectivity’” (Bach 2012, 236), thus providing an answer to the representation problem. Haslanger’s solution (Bach holds) is specifically to argue that women make up an objective type because women are objectively similar to one another, given their social position occupancy. As he puts it, Haslanger’s method of determining whether some collection of objects has unity is “to investigate if, and to what degree, they are in some sense similar. For a type to be objective, then, the axis of similarity must be objective as opposed to merely conceptual” (2012, 238). Social objectivism supposedly avoids the representation problem by arguing for a nonessentialist

164  Mari Mikkola conception of gender that nevertheless tracks “a real structure of reality as opposed to a conventional aspect of our categorizing” (Bach 2012, 238) – a systematic structure of social subordination and privilege. However, Bach holds, Haslanger’s account is not objective enough. We can – in fact, we should on political grounds – provide a stronger metaphysical characterization of gender kinds as natural kinds that takes an explanatory approach to natural kinds, where natural kinds are “mind-independent ontological structures that support successful inductive practice. In other words, a natural kind is a stable set of correlated properties that can be fruitfully studied” (Bach 2012, 240). One way to conceive of kinds in this way has been offered by Ron Mallon (2007), drawing on Richard Boyd’s homeostatic property cluster (HPC) theory, which explicates a way to understand generalizationsupporting categories. On this view, kinds are to be understood as property clusters, where the properties are imperfectly shared by kind instances. Nevertheless, the clustered properties are reliably coinstantiated by some causally homeostatic – causally stabilizing – mechanism. This property cluster conceivably makes good kind relevant inductions in the absence of a core essence necessary and sufficient for kind membership. Although taking a natural kind approach to gender, Bach rejects Boyd’s and Mallon’s HPC view because it places insufficient importance on historical properties and fixes kind membership in terms of some approximation of similarities of manifest properties that cluster together. This view, Bach claims, cannot satisfyingly address the commonality problems: “HPC’s commitment to shared properties would require individual women and men to exemplify some subset of the gender role property cluster” (264). Although on this view no single feature is necessary for membership in a gender kind, “some unspecified percentage of the delimited gender role properties is necessary” (264). Instead, Bach favors historical essentialism as an alternative explanatory approach. On this view, “historical facts about genetic replication and past selective pressures explain why the members of a species share a phenotypic profile” (242). There is no lawlike causal connection between a kind’s manifest surface features and the kind’s historical essence. For instance, take the kind Homo sapiens: there is no nomological connection between the “essential historical relations that unify members of the kind” (243) – their historical essence – and any particular phenotype. Replication or reproduction is the essential relation that fixes kind membership, provides unity across members given phenotypic diversity, and explains kind-wide similarities (Bach 2012, 243). Drawing on Ruth Millikan’s work, Bach understands direct replication to take place when a token is a direct copy of another; indirect replication or reproduction takes place when a historical model or blueprint is used to reproduce tokens, where the model’s function is to do just that. This connects to general etiological theories of function, where tokens

The Function of Gender as Historical Kind  165 possess teleological functions because they were replicated given past selective pressures: a particular trait T is selected for replication, over and against historical alternatives, because of a past favorable effect that T caused for ancestors that possessed T. Once selected, T has the function of causing this effect. For example, human hearts are reproduced from ancestral hearts because ancestral hearts, more so than historical alternatives, had the favorable effect of pumping human blood. “Pumping blood” is thus the teleological function of hearts, and we can say of a token heart that was reproduced from ancestral hearts that it is “supposed to” pump blood. (Bach 2012, 244) Human hearts have a teleofunction because our ancestor’s hearts happened to play a causal role in the relevant supersystem – the human organism. Human hearts were selected for and replicated because of their contribution to stabilizing the overall, selected-for system (the human organism type) and because the extant human heart “design” pumped blood more successfully than alternative “designs.” The “deep” historical properties to do with replication explain surface manifest properties. Hence, historical kinds, in general, have seven features: 1 Their essential property is participation in a lineage. 2 This property defines kind, not individual, essence. 3 To exemplify this essence, an individual must be a replication or reproduction of ancestors from a lineage. 4 Reproduction and replication are ontogenetic processes caused by mechanisms that produce historical kind members. 5 Historical kind members tend to be similar to one another in various ways because of their history; similarities that are due to something other than shared history are not necessary for kind membership. 6 Individuals reproduced from a selected-for historical lineage possess teleological functions. 7 This function is defined by “the contribution made by ancestral kind members to the stability and effect of a more general historical system.” (245–246, italics mine) With this general picture of historical kinds in mind, Bach argues that gendered kinds are to be understood as such kinds. B eing a woman (or a man) is not a matter of having some biological or psychological properties, or of occupying a certain social position. These are mere indicators of a deeper shared history or origin. Hence, womanhood is about having “the right sort of origin and replicative history in

166  Mari Mikkola relation to a more fundamental historical kind – a replicating gender system” (246). To illustrate, Bach offers the following analogy. The historical kind “1990 Nissan Sentra” is a broader replicating system. All token Sentras have been reproduced from the same historical design, which is why individual token Sentras have certain features (good gas-mileage, front-wheel drive). The various subcomponents of Sentras were selected for and reproduced given the role that they played in successfully replicating the Sentra system, a system that met the demands and interests of consumers more than other competing car-model systems did. Now, the subpart “Sentra transmission” is also a historical kind because individual Sentra transmission tokens are reproductions of a lineage of Sentra transmissions. Sentra transmissions possess a teleological function of shifting gears in Sentras. Given the history of reproduction under past selection pressures, the kind 1990-Sentra transmissions plays a causal role in the overall Sentra system: since the teleological function of shifting gears in Sentras was performed particularly well by the 1990-Sentra transmission kind, this historical kind fulfilled a system-relative function that contributed to the stabilization and reproduction of the “1990 Nissan Sentra” system (246). To summarize Bach’s picture: both the “1990 Nissan Sentra” and “1990-Sentra transmission” are historical kinds because they are reproductions of historical designs, where the former kind denotes a broader system that is produced and kept in place by Sentra subparts. These subparts (like the “1990-Sentra transmission”) have a teleological function: Sentra transmissions have the function of shifting gears in Sentras and in successfully performing this teleofunction, Sentra transmissions contribute to the existence and persistence of the overall Sentra system. Gender kinds supposedly function alike. The overall gender system is a historical kind akin to the historical kind of “1990 Nissan Sentra,” and the individual components of our gender system are like Sentra transmissions: they were selected for and reproduced given the role that they played in successfully replicating and stabilizing the overall gender system. Although there might not be anything ontologically necessary or ethically desirable about such a system, our gender system is homeostatic – “its various components have achieved a stability that is resistant, though not impervious, to change” (247). For Bach, the homeostatic gender system has (at least) seven components: 1 Binary sex categories: categories of male and female. 2 Conceptual gender dualism: “The tendency for individuals to think and categorize in terms of the masculine and feminine binary.” 3 Gender identity: identification with a particular gender, its roles, and taking certain gender norms to be applicable to oneself.

The Function of Gender as Historical Kind  167 4 Binary gender socialization practices: e.g., sex segregation. 5 Social and legal institutions: e.g., gender differences in law. 6 Binary gendered artifacts: e.g., masculine and feminine toys and clothing. 7 Binary roles: e.g., gendered roles, sexual division of labor. (247–248) The idea is that these individual components have been selected for and reproduced because they play an important role in the successful replication of the gender system as a whole; they have made our gender system homeostatic. On the picture put forward, then: “Reproduced members of a historical gender kind possess a teleological gender function if the gender lineage of which they are a reproduction resulted from selective processes” (249). It does not matter whether these processes resulted from pressures that were biological or social – both are consistent with the historical account. We can (for instance) tell a story about cultural processes doing the work. Cultural mechanisms that specifically favor the selection of gender include institutions like the media, schools, and the medical profession that effectively transmit gendered norms, expectations, and other components of the gender system. Nevertheless, we cannot elucidate the teleological gender function via conceptual analysis and must investigate “the contributions made by the historical lineages of men and women toward the stabilization of a gender system” (254). This requires social scientific research, and some such research already exists relative to the gender roles of men and women (component seven above). Plausible candidate features for the teleological functions of gender roles then include: (a) social hierarchy: women having less power and status than men; (b) division of labor: women performing more household chores, sex segregation of/in household and paid labor, earning differentials; (c) personality traits: men having more agentic traits, being aggressive; women having communal traits, being “tender-minded”; and (d) body management: women wearing makeup, electing for cosmetic  surgery, the increased frequency of women smiling, and so on (254–255). Those individuals in our ancestral past who acted in conformity with these features acted to stabilize the gender system. If the above characteristics were culturally reproduced because of their stabilizing effects, we can conclude that they were culturally selected for. On the picture put forward, then, the above four binary sets of properties “constitute the teleological functions of men and women” (255). So, no one is born a man or a woman; instead, certain ontogenetic processes reproduce members of gendered historical kinds out of sexed individuals. The most important of these processes is differential socialization (component four of the gender system): parents attributing

168  Mari Mikkola gender stereotypic behaviors to their children, talking about them using gender stereotypic language, sex segregation in schools, and among tasks assigned to boys and girls. In sum: [B]oth conscious and unconscious design mechanisms mold infants as reproductions of historical men and women from the moment they are born. Through these processes, an individual comes to participate in a lineage of men or women and thereby becomes a member of the historical kind men or women. (260) One’s sex is not necessary for participation in a particular gender kind: sex merely marks individuals as targets for particular ontogenetic processes. Furthermore, to be a woman does not require that one exemplify the properties outlined above that are constitutive of women’s historical gender role (“to be subordinated, tender-minded, present her body according to the norms of the fashion-beauty complex, and perform more housework” [Bach 2012, 261]). For womanhood, it is necessary to have undergone an ontogenetic process through which one comes to take part in the lineage of women. In short: female gender socialization is the mechanism whereby one comes to partake in women’s gender lineage, but one need not exemplify any typically female properties or undertake stereotypical female roles. One is a woman because one has “the right history” (261). Bach’s own example is of a swamp-woman created by a lightning-strike: she may be relatively powerless, perform housework, be tender-minded, and pay attention to feminine fashion norms, thereby exemplifying the properties that define women’s historical gender role. But she would not, therefore, be a member of women’s kind. This is because the swamp-woman has the wrong history, and has not undergone the ontogenetic process of gender socialization that would give her ‘the right’ history. Let me summarize the view on offer: at the most basic and overarching level, we find the replicating gender system. Next level up, we find seven components of this underlying system, which were selected for, given their success in making the gender system homeostatic. The components were replicated and reproduced precisely because of this success, and so have (causal) teleofunctions relative to the gender system. Out of these components, gender socialization represents the most important ontogenetic process that reproduces gender out of sex. So, on the third level, we find the historical kinds of women and men; they are created and recreated by the components of the gender system via the ontogenetic process of gender socialization. These historical kinds also have a teleofunction of keeping the overall gender system in place. Finally, on the manifest level, we find typical gender traits: social position occupancy, social roles, behavioral patterns, body management. And so,

The Function of Gender as Historical Kind  169 We can assume that ancestral males and ancestral females exhibited a range of social and behavioral traits. Among these competing gender role cultural variants, sexed individuals who exemplified properties [emblematic of gender traits] in conformity to the dualisms that structure a gender system acted to stabilize that system more than individuals who did not exemplify these properties. If these gender roles were culturally reproduced on account of this stabilizing effect, we can conclude that they were culturally selected for. On the etiological account of functions sketched earlier, the binary sets of properties … [emblematic of gender traits] constitute the teleological functions of men and women. (Bach 2012, 255) The surface similarities in (say) roles and behavior are not definitive of gender. Instead, they are manifestations of a deeper, more fundamental similarity: having the right history and shared ontogeny. What does the important ontological work, and provides responses to the representation and commonality problems, is the third level. What makes x a woman is x’s participation in a lineage of women. To participate in this lineage, x must be a reproduction of ancestral women. To be such a reproduction, x must have undergone the right ontogenetic process. Undergoing this process (socialization) is one mechanism with which the base-level gender system replicates women. As a result, women have certain teleological functions captured by the set of typical gender traits – they have the function of contributing to the evolution and reproduction of historical gender kinds.

How Theoretically Compelling Is Bach’s Functional Analysis? The account Bach gives is complex and novel within contemporary metaphysics of gender. But a closer look suggests that this complexity generates tensions internal to Bach’s analysis. Consider first the claim that the seven components of the gender system have rendered the system homeostatic insofar as “its various components have achieved a stability that is resistant, though not impervious, to change” (Bach 2012, 247). Homeostasis is about the ability of some organism or environment to regulate itself to retain stability in the face of changes. For instance, the human body provides numerous examples. Paradigmatically, the internal ability to regulate bodily temperature to keep it at a constant is an example of homeostasis. The same is true of keeping the body’s glucose levels at a constant by either the pancreas releasing insulin if the levels are too high or the liver converting glycogen in the blood to glucose if blood glucose levels drop too low. The basic idea is that if one part of the system malfunctions, other parts of the system come to the aid and

170  Mari Mikkola compensate for this malfunction – otherwise the entire system is at risk. This being the case, homeostatic systems cannot change radically in the short run since the system’s persistence is caused and explained by its parts compensating for changes in one part of the system to stabilize the system as a whole. If the gender system is homeostatic though, it is hard to see how the system can radically change. That is, any radical changes to (say) practices of binary gender socialization might be compensated for perhaps by strengthening and congealing the genderedness of social and legal institutions or through binary gender roles. Nonetheless, Bach seemingly holds that the gender system can undergo radical change: The proposed ontology interprets the destination point of gender egalitarianism as a future gender system that culturally reproduces men and women so that they are supposed to have (in the historical/ etiological sense) equal status, power, and labor opportunities. The proposed ontology helps locate the departure point by describing the natures and boundaries of the targeted social groups. (259) However, given that the starting point is a homeostatic system, it is unclear to me how the desired changes could come about without overturning the system or taking an extremely long time to happen. Understanding oppressive systems like the gender system in this way then suggests that without a thoroughgoing revolution, the system will simply reproduce itself – and in the absence of such a revolution, the status quo will be replicated. This is a politically depressing result, contra Bach’s view that conceptualizing gender as he does supposedly affords us political levers of change. In fact, Bach takes one of the advantages of his view in contrast to Haslanger’s precisely to be that historical essentialism can account for changes in the gender system better. For Haslanger, if sexmarked patterns of social subordination cease, there will no longer be gendered kinds. By contrast, Bach takes his metaphysics to retain gender while allowing that the gender system can be transformed in ways that make it characterized by gender justice and while remaining historically continuous with our current unjust gender system. If the system is homeostatic, though, it is hard to see how this is possible. That said, thinking about changes that the gender system has already undergone in a relatively short period might suggest a different (and less depressing) conclusion: maybe the gender system isn’t homeostatic after all, which would undermine Bach’s analysis from the start. Bach’s functional analysis relies on the gender system being homeostatic due to which the individual parts of the system have teleofunctions that reproduce the system. But if the gender system isn’t homeostatic in the sense Bach has in mind, his functional analysis of gender kinds is less compelling, and the selection pressures he has identified start to look exaggerated.

The Function of Gender as Historical Kind  171 This connects to my second worry about just how explanatorily valuable historical essentialism is. What motivates Bach’s etiological view is precisely that thinking about gender kinds as natural kinds with explanatory essences can better ground inductive practices. Genders are historical kinds because they are reproductions of ancestral women/ men via the mechanism of binary gender socialization. Complying with certain gendered norms and behaviors functions to reproduce gendered kinds. These kinds (among other components) then function to make the traditional gender system homeostatic. One might, however, wonder whether there is anything particularly surprising or explanatorily illuminating about the claim that dividing people along gendered lines helps to keep in place a broader binary gender system. Consider a parallel explanation of slavery: a system of slavery is kept in place and stabilized by dividing people into “natural” slaves and “natural” masters. This claim certainly has surface intuitive plausibility, but it does not offer a great explanation of why slavery exists and what keeps the system in place. At the very least, a good explanation should provide systematized answers to questions that are not self-evident and that illuminate something that would otherwise have been left unilluminated. Explaining the existence and persistence of a binary gender system by pointing out a division that places individuals in binary gendered kinds does no systematic explaining whatsoever. Bach does hold that social hierarchy, division of labor, personality traits, and body management are plausible candidate features for the teleological functions of gender roles insofar as those who acted in accordance with these roles have congealed the gender system. But again, think about the parallel to slavery: to say that those who performed the roles of “natural” slaves and “natural” masters particularly well acted to congeal a system of slavery is not a strong functional explanation of the system of slavery. A functional explanation of slavery should explain why it existed by appealing to the purpose or needs served by slavery; however, pointing out that the social practice of slavery was maintained by social agents acting in accordance with some relevant social expectations does not provide an answer to the why-question. However, one might hold, my above points are uncharitable. Over and again, Bach stresses that his model is explanatory superior in that it supports inductive practices better than alternatives. So perhaps this is what we should focus on rather than focusing on an explanation of why the gender system exists and persists. Bach’s basic idea is that “reproductive properties tend to make kind members alike—and this is why historical kinds support induction and qualify as explanatory natural kinds” (Bach 2012, 264). By contrast, Haslanger’s social objectivism allegedly has at least four problems: Haslanger’s definition of subordination is too general to be helpful, her account is unintuitive, it makes future-oriented political action impossible in having the result that gender justice would

172  Mari Mikkola eradicate gender, and Haslanger’s elucidation of gender qua objective types is not explanatory (265–271). Specifically, relative to the final point Bach writes: claiming that a group of items is an objective type is to state a unity rather than explain a unity. In contrast, an ontology of historical kinds provides explanations for objective similarities. The present proposal agrees with Haslanger’s social objectivism that women generally share the objective similarity of social subordination. But rather than making this property definitive of women, it explains this property as one of several properties that reliably co-occur as a result of a more fundamental, historical property. (271) Even so, I remain unconvinced that the explanation Bach provides is particularly compelling. It simply asserts that women and men are reproductions of ancestral women and men; hence, they typically manifest certain surface properties given that the mechanism of gender socialization is at work. Moreover, Haslanger’s social objectivism does not preclude social explanations of the kind Bach is after – Haslanger’s project is rather doing something subtly different. Haslanger is offering a constitutive elucidation: what it is to be a woman is to be sex-marked for social subordination; this objective similarity binds women’s type together. Bach is offering a causal elucidation: how is it that women came to have manifest similarities? Answer: there is an ontogenic process that generates these similarities. To build causality into her view, Haslanger could just as well appeal to gender socialization as one mechanism that positions women in subordinate ways. It is just that Haslanger is aiming to provide a different kind of explanation: a non-causal one.2 Finally, it is not obvious to me that Haslanger’s constitutive account fails to support inductive practices. If we take womanhood to be about occupying a socially subordinate position, does this support the prediction that (say) the next n-number of women you encounter are likely to experience some form of gendered subordination? Given structural and statistical realities (of, say, sexualized violence, gender segregation in various social spheres, and the gendered pay gap), this prediction is highly likely to be confirmed. Of course this is vague and imprecise – but then, so is Bach’s proclamation that his account better supports inductive practices and explanatory endeavors. What his metaphysics tells us is that those who have been gender socialized in particular ways tend to share behavioral, psychological, and social properties, where the replicating ontogenetic process explains this sharing. Bach may provide a more specific causal story than Haslanger, but since Haslanger’s explanatory burden is different from Bach’s in being constitutive, I am not yet convinced that Haslanger’s general metaphysics is explanatorily inferior.

The Function of Gender as Historical Kind  173

The Political Tenability of Historical Essentialism Historical essentialism is meant to conceptualize women’s kind in a politically helpful manner, and thus overcome pressing feminist problems. Above I noted that if the gender system is strongly homeostatic, there is little point to feminist politics: the system will adapt to changes to reproduce itself. This prima facie undermines the idea that Bach’s conceptualization of gender is politically usefulness. There are other grounds to find his analysis politically worrisome, though. In particular, Bach’s model faces serious concerns relative to his solutions to the representation and normativity problems. Consider political representation first. One of the most puzzling moves in Bach’s ontology is generated by his claim that members of historical gender kinds have teleological functions. This being the case, gender – as we ordinarily know it – can be analyzed as a historical kind and a teleofunctional kind. Token replicating systems are individuated historically: tokens of the US gender systems of the past and the present are members of the same historical kind because the extant system descended from the former historical system. Contemporary US women and those of (say) the 1950s are “members of the same historical kind because they descend from the same ancestral population of American women” (Bach 2012, 261). Gender as a historical kind does not, however, travel. For instance, American and Japanese women are “not members of numerically the same historical kind” (Bach 2012, 262). This is because there isn’t significant enough of a historical unity between the American and Japanese gender systems. Hence, American and Japanese women are descendants of different ancestral populations of women. Still, there are similarities between the two gender systems: they are similar types of replicating gender systems (262). As a result, the gender roles we find in the two systems are allegedly roughly analogous and therefore: On account of their shared type of history, then, American and Japanese women are members of numerically the same teleofunctional gender kind. This means that both American and Japanese women can fail to satisfy their teleofunctional gender norm and yet both are still members in the cross-cultural teleofunctional gender kind (and, of course, they are still also members in their respective historical kinds). (262) Despite not being members of numerically the same historical gender kind, insofar as American and Japanese women have analogous historical gender roles, they are members of the same teleofunctional gender kind “given the logical relationship between historical kind membership and possession of a teleological function” (266).

174  Mari Mikkola This move, however, is deeply puzzling. Bach’s metaphysics is ultimately committed to there being historical gender kinds and teleofunctional gender kinds, but the relationship between these types of gender kinds is highly opaque. Moreover, it is unclear what work is being done by the two types of kinds. In response to the representation problem, it seemed that Bach’s solution was to argue that women make up a historical kind. But if women from ancestrally isolated populations end up being from numerically distinct historical kinds, the representation problem remains unresolved. Feminists cannot make claims and advance policies in the name of women – only in the name of (say) American or Japanese women. If the unity needed for feminist political representation is provided by the teleofunctional gender kind membership, there seems to be no meaningful work that historical gender kinds do. Then again, if teleofunctional gender kind membership is not meant to ground feminist political representation, what work is it meant to do? Perhaps the idea is that teleofunctional gender kinds home in on what women have in common (patriarchal distribution of power, similarities in division of labor, and gender stereotypes), whereas the historical kinds of American and Japanese women account for what distinguishes them (e.g. specific gender norms that facilitate teleofunctional kind reproduction).3 Be that as it may, Bach seems to think that gender understood as a historical kind plays a more fundamental role in responding to political concerns raised by the feminist identity crisis. Having a historical essence is precisely what is meant to bind women together – but if women from different geographical areas have qualitatively distinct historical essences qua women, they are not bound together by a shared historical essence after all. Distinguishing these different sorts of kinds looks even more puzzling, given what Bach writes about the type of political representation that feminism is supposedly after. A politically adequate theory of gender must be sufficiently inclusive, and to avoid positing some exclusionary universal feature members of a gender kind must exemplify. For this to be met, feminists are allegedly after natural kind political representation, as opposed to ad hoc political representation. The former “advocates on behalf of categories for which members tend to share certain social and political properties (such as oppression and opportunities) for the same reason,” whereas the latter “advocates for individuals who accidentally share similarities with natural political kinds” (267). Haslanger’s view (according to Bach) is politically too weak because it cannot make good natural kind political representation. This is what the historically essentialist view is said to offer, which elucidates a stronger basis for binding women’s kind together and supposedly provides a better way to make the kind inclusive. One need not exhibit any particular features to be a woman as long as one has the right origin and history. Only individuals who have not undergone the ontogenetic process of

The Function of Gender as Historical Kind  175 gender socialization will be excluded from women’s kind (266). The historical account is hence allegedly more inclusive and marginalizes fewer individuals: only those who have not been socialized as women fail to count as women, but since the socializing ontogenetic processes are ubiquitous, Bach takes his account to be genuinely inclusive. This does exclude (for instance) swamp women from women’s kind. Although such individuals still deserve moral consideration, Bach holds, they nevertheless should not gain feminist natural kind political representation. The basic idea is this: although oppression is not definitive of womanhood, it is no accident that women tend to be oppressed. This surface feature is a manifestation of a deeper shared essence of women having undergone the same ontogenetic process of gender socialization. Surface trait similarities are nonaccidental and non-ad hoc, which supposedly provides a stronger basis for feminist political organizing. So, femalessocialized-as-women will gain natural kind political representation. The swamp woman’s oppression, however, is not a manifestation of a deeper similarity with other women – it is a feature that she accidentally shares with females-socialized-as-women. Thus, swamp women only qualify for ad hoc political representation. However, squaring this idea about different bases for political organizing ill fits the idea that femalessocialized-as-women from different populations make up a teleofunctional gender kind, but not a historical gender kind. Historical essentialism is put forward as the solution to feminism’s identity crisis and is meant to provide a basis of political organizing that is oriented around natural kinds. But if women from isolated populations make up numerically distinct historical kinds, natural kind political representation that Bach is after won’t be possible. There is a further way in which historical essentialism fails to be genuinely inclusive. Some individuals clearly fail to satisfy Bach’s criteria for womanhood: trans* women.4 On Bach’s account, trans* women would not count as women if they have not undergone the right ontogenetic process of gender socialization. Moreover, these processes are said to take place “through events that occurred primarily before the age of ten. Thus, gendered individuals can change in myriad ways while not affecting the relational property that is constitutive of their membership in a gender kind” (268–269). In fact, this suggests that trans* women might be gendered men contrary to self-identification, which raises questions about how genuinely inclusive the historically essentialist account is – it looks instead highly marginalizing, and problematically so.5 Bach does briefly consider transitioning. He asks: when does an individual transition into a new gender? His reply, however, isn’t all that satisfying: this is a difficult question without a clear answer (269), though we do learn from the phenomenon of transitioning that historical gender kinds have vague boundaries – which looks like the right result. Still, his treatment of trans* women is deeply worrying, particularly given the

176  Mari Mikkola history of trans* women’s marginalization within feminist movements. On his picture of historical gender kinds, trans* women end up being either men or like swamp women. If trans* women who have undergone male gender socialization end up counting as men, they won’t be even prima facie candidates for feminist political representation. If trans* women are like swamp women, it will be wholly accidental that they share features with females-socialized-as-women. Again, trans* women will be explicitly excluded from feminist political representation understood in Bach’s natural kind political representation sense – they merely deserve ad hoc political representation.6 One might wonder whether ad hoc political representation is any worse, though. Bach does note that “human flourishing and justice require both forms of representation,” but he adds: “the need for ad hoc political representation should not serve as a corrective to natural kind political representation” (267). So, there is a sense in which natural kind political representation is somehow more expedient. Now, Bach can, of course, retort claiming that this is politically less than ideal, but if the metaphysical facts pull us in one direction and ethical considerations in another, so much worse for the ethical considerations! Metaphysical facts trump political desiderata, and it is metaphysics that should fix the relevant political categories – our politics cannot fix metaphysics. Implicitly such a view seems to figure in Bach’s metaphilosophical commitments: he is seeking to elucidate the metaphysical facts “out there,” which may or may not be ethically desirable. Just consider his way of elucidating the feminist identity crisis that he is aiming to resolve: Nominalists about gender address the Commonality Problems by avoiding them: concerns about universality and normativity do not arise if there are no membership conditions for the kinds men and women. This tactic creates the Representation Problem: denying the existence of women as a group undercuts women’s advocacy. (270) Nominalists supposedly “deny that there are real ontological kinds that correspond to the categories ‘women’ and ‘men.’ On this view, the only unity that exists for members of a gender category is a conceptual or linguistic unity” (234). Elsewhere he puts the point more explicitly (Bach 2016): what distinguishes Haslanger-type social objectivism from historical essentialism is that the former is aiming to classify social structures whereas the latter is aiming to identify them (Bach 2016, 181). In other words, projects classifying gender kinds are driven by political and ethical considerations, rather than the aim of identifying some real extant gender structures. In being value-driven, social objectivism is liable to be mistaken, and, hence, it does not offer a strong enough metaphysics of gender to do the politically relevant work.

The Function of Gender as Historical Kind  177 To alleviate political problems that feminism faces, Bach seemingly holds, we must affirm the existence of women’s real ontological kind – and the only satisfying way to do so is by showing that women make up a natural kind with a historical essence. Only in this way (the argument goes) can feminists advocate on behalf of women. Bach’s meta-ontological commitments differ from more typically feminist ones, though, and the responses to the normativity and universality problems demonstrate that Bach has a particular metaphysical background picture in mind. However, I contend, this skews his formulation of the politically pertinent issues, in which case the retort that although trans* women’s exclusion is ethically undesirable it is metaphysically grounded fails. To see this more clearly, consider some of Bach’s background ontological presuppositions. He distinguishes two ways in which we can conceive of kinds: some kind x is either part of reality’s structure, and therefore objective and non-conventional; or x is conventional and consequently nonreal and ontologically dependent of our conceptual schemes/ classificatory practices. Given this background ontological picture, the feminist quest for a real kind of women appears to be a quest for a kind that is objective (explanatory in the sense of supporting inductions), nonconventional (in the sense of not taking into account, say, political goals), and carved in nature’s joints. Otherwise, feminists cannot make good the reality of women’s kind: if the kind is unified merely conventionally in hinging on our choice of conceptual schemes and classificatory practices, it won’t be real. However, Bach’s picture does not exhaust the available ontological options. Moreover, when feminists are after a real kind of women, they are after something other than Bach’s real kinds. Feminists are more usually after a kind that is objective, conventional in the sense of depending on our conceptual schemes, and real. Or to put the point differently: they reject the stark distinction between value-laden classification and value-neutral identification that Bach (2016) makes. First, we can identify a kind that is objective relative to certain strongly normative practices; second, this is compatible with the kind being conventional in the sense that its existence and persistence depends on our conceptual choices and classificatory practices; and third, the kind is real insofar as it has causal and explanatory powers. Consider traffic rules in the United Kingdom and the United States by way of example. These rules are conventional: not only do they depend entirely on contingent human practices, but the rules are a consequence of more or less arbitrary choices about conceptual schemes and social practices. There is no mind-independent “fact of the matter” about which side is the correct one to drive on – in this sense, traffic rules are not to be found in nature’s joints. Nevertheless, the rules are objective: relative to certain social frameworks, norms, and practices, traffic rules are not “up to us” or up for grabs. They have (what Searle (1995) calls) deontic powers, although we confer them the relevant authority. Finally, these objective and conventional traffic rules are real in

178  Mari Mikkola the sense of having causal and explanatory powers. Were I to bunk the norms and drive on the right-hand side in the United Kingdom, I would very probably cause an accident. Were I to follow UK traffic rules while driving in the United Kingdom, the likelihood of my driving causing an accident would be diminished. If I have a severe accident, my bunking the objective and conventional traffic rules can be used to explain what caused the accident. So, traffic rules support inductions and causal explanations. The upshot is this: we have more than Bach’s two ontological options available to us, and women’s kind seems ontologically more akin to traffic rules. The response that trans* women deserve ad hoc political representation only because our ontology tells us so is thus not only ethically problematic but also metaphysically unwarranted. If women’s social kind is conventional in the sense of depending on the choice of conceptual scheme, objective, and real in the manner outlined, we can spell it out in a way that includes trans* women. Hence, I see little reason to accept Bach’s metaphysics of gender. He could, of course, reply further that my “real” kinds are not really real, but supervene on properly real ones. So, traffic rules supervene on psychological states of conscious agents, and these states are carved in nature’s joints. This would undermine my alternative ontological position. But if it turned out that traffic rules are not real because they supervene on psychological states, then – I submit again – feminist philosophers simply are not after a real kind of women in Bach’s sense. This (once more) would demonstrate that his way of framing the crucial political issue differs from more usual ways that feminists frame them. Hence Bach’s above characterization of nominalism that supposedly threatens the viability of feminist politics isn’t quite right. To see this another way: Mallon helpfully distinguishes skeptical antiessentialists, who deny that category membership hinges on some individually necessary and jointly sufficient conditions, from constructionist antiessentialists, who deny that category membership is delimited by intrinsic or natural features (Mallon 2007, 148). With this in mind, he writes: Skeptical anti-essentialists typically attempt to problematize, deconstruct or eliminate human category concepts and generalizations, while constructionist anti-essentialists typically offer non-essentialist reinterpretations of human category concepts in order to employ them to do theoretical and practical work. (Mallon 2007, 149) At the start of this chapter, I mentioned gender-skeptical positions that have, to a significant extent, generated the feminist identity crisis; these positions are akin to Mallon’s skeptical antiessentialists. Gender nominalists are committed to something else: to antiessentialist reinterpretations

The Function of Gender as Historical Kind  179 of human categories, as Mallon puts it (for more, see Mikkola 2016). In framing the central feminist problems, Bach has in mind gender skeptical or skeptical antiessentialist positions. But in so doing, his presentation of the problems becomes questionable. Nominalists are also aiming to elucidate a real kind of women: a kind that is not a mere fiction, that is objective relative to certain strongly binding practices and norms, and one that has causal and explanatory efficacy. Thus, Bach’s formulation of the representation problem already presupposes his background meta-metaphysical commitments, rather than those feminist philosophers more commonly hold: namely that there are gender kinds to be discovered, even though these kinds are historically constructed through past selection pressures. If we reject this way of framing the pertinent problems (as feminist philosophers typically do), the motivation to provide a metaphysically stronger elucidation of gender kinds in the service of natural kind political representation is undermined.

Conclusion In this chapter, I have argued against Bach’s functional analysis of gender kinds. His account has internal tensions that render it opaque, and it fails to satisfy the requirement to resolve the feminist identity crisis. However, one might ask, is there another functional account of gender that might do the job better? I wish to finish the paper briefly addressing this point. The feminist identity crisis has generated much attention since it supposedly undermines feminist politics. Bluntly put, we must be able to account for the “glue” that binds members of gender kinds together in order to formulate sensible emancipatory policies. If women merely make up a gerrymandered kind, there is no basis for making political demands on behalf of women as a group. Now, Bach’s historical essentialism is one way to account for this “glue,” and I have argued above for rejecting his account. But I am not convinced that we require a social-functionalist account of what binds women together – in fact, I do not think that we require any metaphysical account to elucidate this. I argue for this conclusion in length elsewhere (Mikkola 2016). But let me briefly elucidate my view. Simply put, a metaphysical elucidation of the “glue” isn’t politically necessitated, and the efficacy of feminist policies does not stand or fall with the plausibility of “our” gender metaphysics. Think back to Bach’s claim that social constructivist terms aim to classify the world, whereas historical essentialist terms aim to identify it. As I noted before, feminist philosophers typically reject this distinction. Specifically, with respect to gender this seems rather strikingly obvious. For instance, Naomi Zack holds that the feminist identity crisis isn’t about the denotation of woman; it is about its connotation (2005, 27). Marilyn Frye’s remarks about oppression are also illustrative. She holds

180  Mari Mikkola that a system of gender oppression works to “reduce, immobilize and mold people who belong to a certain group, and effect their subordination to another group” (1983, 33). However, Frye continues: Such a system could not exist were not the groups, categories of persons, well defined. Logically, it presupposes that there are two distinct categories. Practically, they must be not only distinct but relatively easily identifiable; the barriers and forces [of oppression] could not … be applied if there were often much doubt as to which individuals were to be contained and reduced, which were to be dominated. (Frye 1983, 33) So, the very existence of gendered injustices suggests that feminists and ordinary speakers alike do engage in practices of identifying women, rather than stipulatively classifying women. Insofar as this is the case, we need not metaphysically elucidate the kind-binding “glue” in order to respond to practices of oppression. Instead, we need to analyze those practices – doing so and responding to them adequately is what grounds the efficacy of feminist politics. This is not to say that functional explanations of gender are utterly useless and uninteresting. There may be legitimate and important projects that would benefit from – even demand – explanations of why a gendered practice exists in virtue of the purpose or needs served by the practice. But these sorts of explanations are not necessitated by the feminist identity crisis: in short, responding to the representation and commonality problems simply isn’t necessary to ground and justify feminist political action.

Acknowledgments I would like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.” Furthermore, I am particularly grateful to Daniel James and Rebekka Hufendiek for their extremely helpful and constructive comments on earlier drafts of this chapter.

Notes

The Function of Gender as Historical Kind  181 consider themselves to be transgender. The term “transsexual” is typically used to refer to individuals who use medical means and technologies to alter their bodies, so that their bodily presentation conforms to their gendered sense of self (Bettcher 2009). By contrast, “transgender” refers to people who ‘do not conform to prevailing expectations about gender’ by presenting and living genders that were not assigned to them at birth or by presenting and living genders in ways that may not be readily intelligible in terms of more traditional conceptions of gender. (Bettcher 2009) trans* is contrasted with “cis” that denotes (roughly) women-born-female and men-born-male, who have typical gender identities and presentations.



References Alcoff, Linda. 2006. Visible Identities. Oxford: Oxford University Press. Ásta. 2011. “The Metaphysics of Sex and Gender.” In Feminist Metaphysics: Explorations in the Ontology of Sex, Gender and the Self, edited by Charlotte Witt, 47–65. Dordrecht: Springer. Bach, Theodore. 2012. “Gender Is a Natural Kind with a Historical Essence.” Ethics 122, no. 2: 231–272. ———. 2016. “Social Categories Are Natural Kinds, not Objective Types (and Why it Matters Politically).” Journal of Social Ontology 2, no. 2: 177–201. Bettcher, Talia Mae. 2009. “Feminist Perspectives on Trans Issues.” The Stanford Encyclopedia of Philosophy (Winter 2009 Edition), edited by Edward N. Zalta. http://plato.stanford.edu/archives/win2009/entries/feminism-trans/. Butler, Judith. 1999. Gender Trouble. London: Routledge. Frye, Marilyn. 1983. The Politics of Reality: Essays in Feminist Theory. Trumansburg, NY: Crossing Press. ———. 1996. “The Necessity of Differences: Constructing a Positive Category of Women.” Signs 21, no. 4: 991–1010. ———. 2011. “Metaphors of Being a ϕ.” In Feminist Metaphysics: Explorations in the Ontology of Sex, Gender and the Self, edited by Charlotte Witt, 85–95. Dordrecht: Springer.

182  Mari Mikkola Haslanger, Sally. 2000. “Gender and Race: (What) Are They? (What) Do We Want Them to Be?” Noûs 34, no.1: 31–55. ———. 2003a. “Future Genders? Future Races?” Philosophic Exchange 34: 4–27. ———. 2003b. “Social Construction: The ‘Debunking’ Project.” In Socializing Metaphysics: The Nature of Social Reality, edited by Frederick Schmitt, 301–325. Lanham, MD: Rowman & Littlefield Publishers Inc. ———. 2005. “What Are We Talking About? The Semantics and Politics of Social Kinds.” Hypatia 20, no. 4: 10–26. Heyes, Cressida. 2000. Line Drawings. Ithaca, NY; London: Cornell University Press. hooks, bell. 2000. Feminist Theory: From Margins to Center. 2nd ed. London: Pluto Press. Kincaid, Harold. 2006. “Functional Explanation.” In Handbook for the Philosophy of Science: Philosophy of Anthropology and Sociology, edited by M. Risjord and S. Turner, 205–239. Amsterdam: Elsevier. Mallon, Ron. 2007. “Human Categories Beyond Non‐essentialism.” Journal of Political Philosophy 15, no. 2: 146–168. Martin, Jane. 1994. “Methodological Essentialism, False Difference, and Other Dangerous Traps.” Signs 19, no. 3: 630–655. Mikkola, Mari. 2007. “Gender Sceptics and Feminist Politics.” Res Publica 13, no. 4: 361–380. ———. 2016. The Wrong of Injustice: Dehumanization and Its Role in Feminist Philosophy. Oxford: Oxford University Press. Saul, Jennifer. 2012. “Politically Significant Terms and Philosophy of Language.” In Out from the Shadows: Analytical Feminist Contributions to Traditional Philosophy, edited by Sharon Crasnow and Anita Superson, 195–216. Oxford: Oxford University Press. Searle, John. 1995. The Construction of Social Reality. London: Penguin Books. Spelman, Elizabeth. 1990. Inessential Woman. Boston, MA: Beacon Press. Stoljar, Natalie. 1995. “Essence, Identity and the Concept of Woman.” Philosophical Topics 23, no. 2: 261–293. ———. 2011. “Different Women: Gender and the Realism-Nominalism Debate.” In Feminist Metaphysics: Explorations in the Ontology of Sex, Gender and the Self, edited by Charlotte Witt, 27–46. Dordrecht: Springer. Stone, Alison. 2004. “Essentialism and Anti-Essentialism in Feminist Philosophy.” Journal of Moral Philosophy 1, no. 2: 135–153. Tanesini, Alessandra. 1996. “Whose Language?” In Women, Knowledge, and Reality, edited by Ann Garry and Marilyn Pearsall, 353–365. London: Routledge. Witt, Charlotte. 2011a. The Metaphysics of Gender. Oxford: Oxford University Press. ———. 2011b. “What Is Gender Essentialism?” In Feminist Metaphysics: Explorations in the Ontology of Sex, Gender and the Self, edited by Charlotte Witt, 11–25. Dordrecht: Springer. Young, I. M. 1997. “Gender as Seriality: Thinking about Women as a Social Collective.” In Intersecting Voices, edited by I.M. Young, 12–37. Princeton, NJ: Princeton University Press. Zack, Naomi. 2005. Inclusive Feminism. Lanham, MD: Rowman & Littlefield Publishers, Inc.

9

Function Without Intention? A Practice-Theoretical Solution to Challenges of the Social Domain Amrei Bahr

Introduction Artifact functions are commonly taken to be intended, be it by the creators of the artifacts or their users or recipients. While many philosophers take this to be true of paradigmatic functions of technical and artistic artifacts (i.e., artworks and their instances) in particular, they have paid considerably less attention to the role of intentions with respect to functions of social artifacts1 such as laws, institutions, corporations, and the like. 2 In fact, in philosophical approaches to artifact functions, functions of social artifacts are often explicitly excluded.3 Although an account that grants intentions a central role appears to be very plausible concerning technical and artistic artifacts, given our experiences of the genesis and use of these items, this is less so in the case of social artifacts. Prima facie, not only the persistence but also the functionality of social artifacts seems to depend on complex processes, at least some of which dispense with intentions.4 This idea is prominently reflected in sociological research, first and foremost in Robert K. Merton’s famous distinction between manifest and latent functions. Whereas manifest functions trace back to intentions, this is not the case with latent functions – they are, as Merton puts it, “unintended and unrecognized consequences” (Merton 1968, 117). Moreover, social factors call into question that intention-centered theories are universally applicable to artifact functions regardless of their nature when it comes to social functions of artifacts which themselves are not social entities. These artifacts also cast doubts on an overall necessity of intentions for artifact functionality. Candidates for unintended social functions of nonsocial artifacts most prominently include what Achinstein dubs service functions (Achinstein 1977), that is, activities of these artifacts that – albeit not intended – are still beneficial for a social group. In this chapter, I will explore the impact of social factors on artifact functionality in the two senses outlined above. Thus, I will inquire whether the challenges that social factors pose to intentionalist accounts of artifact functions should ultimately motivate us to abandon such

184  Amrei Bahr accounts (at least when it comes to social artifact functions).5 The chapter falls into three parts. In the first part, I briefly review intentionalist accounts of artifact functions to make clear what characterizes these accounts and to lay bare the benefits of taking the intentionalist stance. In the second part, I show in how far social factors pose challenges to intentionalist accounts. I present three cases of statements appealing to social functions which seem to suggest that intentions should be considered superfluous when it comes to social artifact functions. In the first two cases, social functions of nonsocial artifacts – one of them technical, one artistic – appear to be unintended. The third case centers on a putatively unintended social function of a social artifact. In the third part, I present an intentionalist solution to the challenges posed by these cases, drawing on insights from practice theory. As will become apparent, although intentions play a special role for artifact functionality in the social domain, we nonetheless should not do without them. On the contrary, it is insightful to acknowledge how intentions that are embedded in practices play a prominent role in the genesis of social artifact functions. To put the investigation of social artifact functions in relation to both the debate on social functions and the other contributions to this volume, a concluding remark regarding the particular concept of social function that is subject of this investigation is called for. According to a widespread use of the term ‘social function’, something is a social function only if it is functional in relation to a social whole. The candidates for social functions discussed in this chapter are not social functions in this sense, but are nevertheless referred to as social functions. The aim of this chapter is to clarify to what extent the artifact-related use of the term ‘social function’ in these function statements calls intentional accounts of artifact functionality into question.

Intentionalist Accounts of Artifact Functions In philosophical debates about artifact functions, intentionalist accounts are prevalent. Proponents of such accounts include Bahr (2019), Dipert (1995), McLaughlin (2001), Neander (1991a, 1991b), and Searle (1995). According to intentionalist accounts, artifact functions are essentially6 determined by intentions of human agents, i.e. the designers and/or the users of the artifacts that have the functions in question. By contrast, effects of artifacts that cannot be traced back to intentions do not constitute functions – they are mere side effects. An example that is frequently mentioned to illustrate this difference is the example of aspirin. An obvious function of aspirin is to cure headaches – that is what the drug was designed for in the first place. But it turned out that the agent in aspirin that is responsible for this intended effect – acetylsalicylic acid – can have an additional effect that is desirable as well: acetylsalicylic acid also has the disposition to reduce

Challenges of the Social Domain  185 the risk of suffering from heart attacks and strokes. After the discovery that acetylsalicylic acid has that disposition, something happened which is quite usual in dealing with medicine: aspirin, which was initially produced as a painkiller, was prescribed to people with a high risk of suffering from heart attacks and strokes. This established practice in medicine is called off-label use. Clearly, a particular aspirin pill that is neither produced to lower the risk of heart attacks and strokes nor put to this particular (off-label) use can nonetheless have that particular effect on a patient. However, according to intentional accounts, that effect would not constitute a function of aspirin because there was no intention directed at this effect – neither by the designer of the drug nor by its user. Hence, we can formulate the principle of intention (PI) that is at the heart of intentionalist accounts of artifact functionality: PI The ascription of the function of Φ-ing to an artifact a is justified only if there are good reasons for the assumption that a is intended by its designer or user to perform the function of Φ-ing.7 Intentionalist accounts of artifact functionality have a number of undeniable benefits.8 First and foremost, they fit together very well with our experiences in designing and using artifacts and therefore have an immediately intuitive appeal to them. On the one hand, given these experiences, it is eminently plausible that artifacts owe certain functions to the intentions of their designers. Take the construction of a slingshot – an artifact that many of us have built in our childhood, using a forked branch and elastic band. Ascribing the function to fling small projectiles such as stones using the slingshot appears to be reasonable not only because the item is capable9 of serving that function, but precisely because it has intentionally been made to serve that function. Moreover, as for designing artifacts, the intentionalist stance corresponds very well with the prevalent view on how artifacts come into existence and obtain (some of) their features. According to this view, artifacts owe both their existence and some of their properties to the intentions of their creators (see, e.g., Hilpinen 1993, 157, 159; Baker 2004, 53; Reicher 2013). It therefore stands to reason that certain functional characteristics of an artifact are also due to the intentions of its designer. On the other hand, it is equally plausible that artifacts acquire functions via intentional usage. Take the chair that I regularly use to reach a high shelf in my kitchen. It seems plausible to say that this chair has the function of serving as a ladder precisely because I use it as such. In addition, granting intentions a prominent role in function ascription has the benefit that phenomena of malfunctioning can be accounted for (Houkes and Vermaas 2010, 51). For instance, even if artifacts (temporarily) lack the capacity to perform a function, we can explain

186  Amrei Bahr by reference to intentions why it is reasonable to ascribe these functions to them anyway.10 By contrast, many theories of artifact functionality which do not entail that intentions are necessary for artifact functions cannot account for malfunctioning. For instance, ascribing the function of showing broadcasts to a broken-down television set cannot be done with causal-role function theories, since the television set does not have the respective disposition, a disposition these theories require to ascribe a function (Houkes and Vermaas 2010, 59). By contrast, an intention-centered theory of artifact functionality can account for the malfunctioning television set by drawing on the designer’s intention to create an item with the function of showing broadcasts.

Challenges of the Social Domain to Intentionalist Accounts of Artifact Functions The social domain poses two challenges to intentionalist accounts of artifact functions as presented above. First, these accounts are confronted with social functions of nonsocial artifacts that appear to be unintended. Second, they are called into question by functions of social artifacts that appear to be unintended. I will address these two challenges subsequently. Since there is a supporting theoretical framework available for each of the two challenges, I will integrate the respective theories into my outline of them. For my presentation of the first challenge, I will draw on Achinstein’s analysis on function statements; for my presentation of the second challenge, I will make recourse to Elster’s account of functional explanations in the social sciences. As will become clear below, the choice of these two accounts is motivated by the fact that prima facie, they appear to be readily applicable to the function statements in question. Candidates for Unintended Social Functions of Non-Social Artifacts The first challenge of the social domain that I want to take into account here is posed by social functions of nonsocial artifacts. Prima facie, there seem to be functions among them that do not owe their presence to an intention after all. I will discuss two examples for which this holds. The first is an example of a putative unintended social function of a technical artifact. In the second example, the putative unintended social function is one of an artwork. To set the stage for these two examples from a theoretical point of view, Peter Achinstein’s work on function statements can provide helpful considerations. Achinstein offers a theory of functions that acknowledges the role that both designer’s and user’s intentions play for a significant number of functions, but leaves room for functions that dispense

Challenges of the Social Domain  187 with intentions. To illuminate what Achinstein has in mind, let us take a look at his throne example (Achinstein 1977, 349). According to this example, which Achinstein uses to introduce his preliminary distinction of three function types, the functionality of a particular magnificent chair can be described as follows: first and foremost, the chair is designed to be a seat for the king – seating the king is what Achinstein calls the design function of the chair. Besides, the chair is used by the king’s guards to block a doorway in the palace – Achinstein calls this a use function of the chair. In Achinstein’s example, however, the guard’s plans to block the doorway by means of the chair do not pan out after all: since the chair is eminently beautiful, it attracts crowds that come to the palace to catch a glimpse of the chair. Hence, instead of blocking the doorway, the chair prompts people to walk through it to have a look at the chair, although there was no intention whatsoever in this regard. Achinstein calls this unintended function of the chair to draw crowds to the palace a service function. Thus, whereas the first two kinds of functions require intentions on the part of either designers or users, this does not hold for the third kind. This theory boils down to the following analysis of function statements: x actually has the function y iff. 1 2

y is in fact done by means of x, and a either x was designed (produced, etc.) to be or to serve as a means of doing y, or b x is used as a means of doing y, or c y’s being done confers a good.11

If (1) and (2a) are met, the function in question is a design function. If cases (1) and (2b) are met, the respective function is a use function. And if (1) and (2c) are met, the function at play is a service function. As we can see, conferring a good somehow serves as a substitute for intentions in the case of service functions in Achinstein’s account, which thereby opens up the possibility of unintended functions. That these functions are unintended can either mean a lack of awareness or a lack of acknowledgment with regard to their benefit. According to Achinstein, not only unawareness but also awareness and rejection of the existence of the respective benefit are possible on the part of the beneficiary. Doing y can confer a good upon a person S, even if doing y is not something which S knows about or regards as beneficial, even if she knows about it. An example is basic training in the army, which can teach recruits to obey orders, even if they do not realize this or do not regard it as beneficial (Achinstein 1977, 355). Now, is it plausible to drop PI in favor of acknowledging nonintended functions such as Achinstein’s service functions? Should we dismiss the intentionalist stance to be able to integrate unintended social functions

188  Amrei Bahr of nonsocial artifacts? The following two examples seem to suggest exactly that move. Nomadic Intimacy: An Unintended Social Function of Smartphones? The first candidate for a social function that appears to be unintended stems from the domain of digital communication. Due to the digitalization of our lifeworld, human communication has undergone significant changes. Thanks to portable digital devices, keeping in touch with friends and family is possible even if these people are not present. Thus, these devices can serve functions for maintaining relations. For instance, smartphones can have the social function of providing their users with so-called nomadic intimacy: The cell phone can be seen as a device that [empowers] moving individuals to connect to any distant partners at any point in time, regardless of location and speed. Thus, one of their major social functions is to provide a “nomadic intimacy” (Fortunati 2000) by making it possible for people on the move to remain embedded in their personal social networks. (Geser 2004 [citation in original]) Using the framework that Achinstein provides, we can now analyze the function ascription that Geser presents. Notably, despite being designed as a means of communication, it seems doubtful that the specific function of providing nomadic intimacy can be considered a design function of smartphones. It is more likely that companies who produce mobile phones have less specific communicative functions in mind in the design process. And it appears to be equally doubtful that providing nomadic intimacy regularly is a use function of smartphones, namely for a similar reason. Even though these devices are without question being used to reach communicative aims, these aims will most probably be less elaborate in many cases and often result from less sociological reflection than the aim of bringing about nomadic intimacy would require. Still, if we take being embedded in a social network while being on the move as beneficial, Achinstein’s account provides us with a reason why Geser’s function statement is plausible after all. Providing nomadic intimacy would justifiably be called a social function of smartphones because it is beneficial for smartphone users, even if they are unaware of this benefit. Moreover, smartphone users might even be aware of the respective effect and reject its benefit; that would not affect providing nomadic intimacy’s status as a service function.12 Hence, if we take the way of speaking that marks out nomadic intimacy as a social function seriously, it appears that providing nomadic

Challenges of the Social Domain  189 intimacy can be a social function of a technical artifact – namely a smartphone – that does not trace back to any intention. Accordingly, this example clearly poses a challenge to PI. If we reach the conclusion that the effect of providing nomadic intimacy can justifiably be called a function of the smartphone despite the fact that we have good reasons to believe that it can be unintended, this should ultimately prompt us to drop PI and thus reject intentionalist accounts of artifact functionality. Substantiating Existing Social Hierarchies as an Unintended Social Function of Modern Art? Examples for putatively unintended social functions of nonsocial artifacts are not limited to the domain of technology. Let us take a look at an example from a yet different domain, namely the domain of art. Here, too, we can find social functions that appear to be unintended. Wolfgang Ullrich studies cases in which modern art acquires the social function of being a status symbol (Ullrich 2000). Whereas art has frequently been used throughout the ages to represent social status and wealth, using contemporary art as a status symbol is, according to Ullrich, a  novel phenomenon (Ullrich 2000, 12). One of the examples he gives is insightful for our topic: in 1997, Rolf Breuer, executive board spokesman of the Deutsche Bank, posed in front of a painting by German artist Günther Förg.13 Ullrich argues that in cases like this, modern art acquires an important function, namely the function of substantiating social hierarchies. The executive board spokesman who routinely smiles while standing in front of a contemporary artwork demonstrates his superiority over his employees who struggle with the artwork. Thereby, the feeling of not being equal in appreciating the artwork extends to the relation the employees have to their superior: he appears to justifiably have his position, given his intellectual competence which he flaunts by posing in front of the work of art (Ullrich 2000, 39). Nonetheless, according to Ullrich, it is improbable that this function of substantiating social hierarchies is intended by the one who benefits from it. On his account, it is not likely that people in a position of power deliberately use works of art in this way – instead, he takes it that it is a fascination for these artworks that prompts powerful people to purchase them and surround themselves with them (Ullrich 2000, 40).14 Now, let us consider the case of Breuer posing in front of Günther Förg’s painting. If Ullrich is rightly assuming that Breuer does not intend to assign the social function of substantiating social hierarchies to the painting, then the function in question is not a use function. Moreover, it is doubtful that Förg intended for his painting to have this particular function. Therefore, it would not be plausible to consider it a design function. However, it is plausible to regard substantiating social hierarchies as beneficial, at least for Breuer, who can thereby consolidate

190  Amrei Bahr his social status in the respective status hierarchies. Hence, Achinstein’s account supports our talk of a function in this case: on this account, the function in question is a service function. Again, this example raises the question as to whether PI is an inadequate requirement for function ascriptions. If we can indeed take the painting’s social function of substantiating social hierarchies to be unintended, this would cast doubt on the plausibility of PI. Once again, we would be confronted with an unintended social function of a nonsocial artifact, which would compromise intentionalist accounts of artifact functionality. A Candidate for an Unintended Function of a Social Artifact The social domain does not only challenge intentionalist accounts of artifact functionality with respect to putatively unintended social functions of nonsocial artifacts. There are also certain functions of social artifacts that appear to be unintended. In what follows, I will present an example of such a function. Here again, there is a theoretical framework at hand that we can adopt as a backdrop for the example, namely Elster’s account of functional explanations15 in the social sciences16: On my definition, then, an institution or a behavioral pattern X is explained by its function Y for group Z if and only if: 1 2 3 4

Y is an effect of X; Y is beneficial for Z; Y is unintended by the actors producing X; Y (or at least the causal relationship between X and Y) is unrecognized by the actors in Z; 5 Y maintains X by a causal feedback loop passing through Z. (Elster 1979, 28)

Now, there are certain obvious parallels between Elster’s account of functional explanations and Achinstein’s account of artifact functionality. Both authors include the feature of being beneficial in their accounts. Also, both address functions that are unintended and even unrecognized – albeit with a slight but important difference: as we have seen, Achinstein includes not only unintended but also intended functions, and with respect to the former, he also allows for cases in which a function and its benefit are recognized while the benefit as such is rejected. By contrast, the functions that Elster is interested in are both completely unintended and beyond recognition.17 This requirement of being entirely unrecognized and unintended all at once is criticized by Peter McLaughlin.18 According to him, this is too strong a requirement for the functions that social entities can have. Complete absence of intentions by all

Challenges of the Social Domain  191 members of group Z cannot be necessary for X to have function Y. McLaughlin gives two examples to illustrate that intentions of single group members are compatible with the presence of certain social functions, namely the religion as opium example and the Hopi rain dance example. As for the first example, suppose that Marx was right in assuming that religion has the social function of being opium for the people. In this case, “religion would nonetheless not then cease to have its function just because one cynical priest (or even the pope) were to read Marx and then redouble his efforts” (McLaughlin 2001,  92). Similarly, in the second example, fostering social cohesion would remain a function of the Hopi rain dance “even if one of the native dancers has an Ivy League degree in cultural anthropology and, by participating in the dance, is deliberately seeking to preserve his native culture” (McLaughlin 2001, 92). I think McLaughlin’s critique is persuasive. Since Elster’s conditions (3) and (4) exclude cases like the ones just mentioned, the class of functions would not only be implausibly restricted; moreover, cases of the ascription of social functions which are especially interesting would fall from view. Ultimately, broadening the conception of function to include cases in which at least some members of a group intend effects or are at least aware of them seems necessary for this purpose. As for the issue of awareness, we should, therefore, consider cases in which awareness of an effect is present, without ruling them out as potential cases in which a function can be ascribed right from the start. However, regarding the issue of intentions, holding on to Elster’s condition (3) is prerequisite, given our epistemic goal. What is of interest to us here is whether the idea that there are entirely unintended social functions is persuasive: since intentions of single group members would be sufficient reason to hold on to PI, cases in which such intentions are at play are not of interest here. Instead, we should focus on a case in which there is a complete lack of intentions with reference to the effect that is supposedly a function. Is it plausible to assume that a social artifact – for instance an institution – X can indeed have a function Y even though no member of a group Z whose lifeworld X is a part of intends X to have that particular function? Let us take a look at an example that might indicate that we should indeed proceed on the assumption that this can be the case. Supporting The Interests of Individual Citizens as an Unintended Function of the U.S. Senate An example of a social artifact with regard to which we could arguably claim that it has an unintended function is the institution of the U.S. Senate. To illustrate how institutions can routinely acquire “significant unintended effects” (Pierson 2000, 484), Paul Pierson refers to this institution, which, in his view, has acquired such an effect. The Senate

192  Amrei Bahr was initially designed to support the interests of the states by having senators appointed by them to serve as delegates. Over time, senators seeking greater autonomy were able to gradually free themselves from state oversight. Since the popular election of senators is required, the Senate now instead serves to support the interests of individual citizens (Pierson 2000, 484). Now, if we adopted a narrow variant of functionalism that requires functions to be intended, supporting the interests of individual citizens would, of course, be a mere effect, but not qualify as a function of the Senate. However, this is not Pierson’s understanding of functionalism. Creator-intended functions mark only one side of the coin, whereas emergence through enhancement mechanisms marks the other: as Pierson puts it, a functionalist logic requires one of two supporting lines of argument: either functional institutions must be products of rational design, or they emerge over time through mechanisms of institutional enhancement. (Pierson 2000, 477) This broader understanding of functionalism would allow for ascribing the social function of supporting the interests of individual citizens to the Senate despite the fact that this effect is unintended, be it by the institution’s designers or the ones who benefit from it. All that is required is that it is likely that certain enhancement mechanisms which do not involve intentions19 have brought about this effect. A possible enhancing mechanism in the example of the Senate could be a lack of efforts to establish a competing institution that could serve as a means to the same end, namely the support of individual citizens’ interests, and would thereby have the potential to end the Senate’s persistence because it takes over one of its effects. Since the Senate serves this particular effect, there is no pressure to form such an institution. But that does not necessarily amount to intending the Senate to have this effect, but rather to not intending another institution to have this effect, because the respective need is already met. Considering the political setting as a whole, we can arguably describe this mechanism as enhancing since establishing a new institution without need would be inefficient. Besides, Pierson’s understanding of functionalism is obviously not the only theoretical resource to vindicate ascribing the function of supporting the interests of individual citizens to the Senate. With Elster’s account of functional explanations, this function ascription can likewise be backed up. We can say that this is an effect of the Senate that is beneficial for the citizens, even though they do not intend it. Moreover, it appears reasonable to assume that the Senate persists and maintains this effect because of its respective benefit. If we take McLaughlin’s critique

Challenges of the Social Domain  193 of Elster’s account serious, then the requirement (4) would be too strong anyway. Even if single citizens were aware of the respective effect, that would not speak against calling the effect a social function. And as argued above, this critique is comprehensible. Hence, even if single citizens are aware of the Senate’s effect in supporting their interests or the causal connection between this effect and the Senate’s persistence, a slightly modified version of Elster’s account would still allow for calling this effect a function that features in the explanation of the persistence of the Senate. And concerning the persistence of the Senate, there might very well be a causal feedback loop that links the persistence (or maintenance, to stick to Elster’s formulation) of the Senate to its effect of supporting individual citizen’s interests. So, the example of the Senate’s allegedly unintended function raises the question as to whether we should drop PI to acknowledge – in line with Elster’s account of functional explanation in the social sciences and Pierson’s understanding of functionalism in political science – that there are functions of social artifacts that are unintended.

A Practice-Theoretical Solution to the Challenges of the Social Domain: From (Initially) Unintended Effects to Functions via Practices Should the three examples I presented above prompt us to drop PI and thereby reject intentionalist accounts of artifact functionality, despite their benefits? I take it that this is not the case. In what follows, I will present a practice-theoretical solution to the challenges of the social domain just outlined that not only justifies holding on to the intentionalist stance, but also adds another explanatory twist to it. To do that, I will first briefly outline the practice-theoretical element of my solution by drawing on an family of theories that operate under the name “practice theory”. Afterwards, I will review the three examples in light of this solution and lay bare the benefits it has in dealing with each of the examples. Even though theories of practice have been on the rise in social theory for the last 20 years, these theories form a multifaceted cluster (Reckwitz 2003) rather than adding up to a consistent practice theory that shares a congruent theoretical core of research propositions, subject matter and concepts (Elias et al. 2014, 4). Still, the respective theories share certain core tenets that justify referring to all of them as theories of practice. Here, I will confine myself to the commonality that matters for the present context, namely the focus on practices. An innovative potential of practice theories lies in their new take on the theory of action. The category of practices substitutes the hitherto decisive category of actions; thereby, intentions are faded into the background (Elias et al. 2014, 4). As Haslanger points out, “[s]ocial practices are not fully intentional,

194  Amrei Bahr as such, but neither are they mere regularities in behavior such as blinking or squinting in bright light” (Haslanger 2018, 237). Still, intentions do not disappear from the scene entirely. Social practices can be taken to “depend on collective attitudes (more precisely shared we-attitudes), which all express collective intentionality” (Tuomela 2004). But how are practices connected to artifact functions? I want to suggest the following connection: practices turn initially unintended effects into social functions, whereas these functions can then serve to individuate the respective social practices. But how does this happen? The idea is that initially unintended effects can be detected and embraced by agents who then establish practices in which these effects are incorporated. Recognizing these effects and perpetuating them through the establishment of respective practices is the transformative moment, as it were, in which they are turned into functions. I take it that this is illuminating both with regard to the distinction of unintended effects and functions and the practices as such. Distinguishing functions from initially unintended effects in this way has the overall theoretical benefit that we can distinguish actions of artifact usage that are determined by social practices from actions that are not determined by such practices. This approach fits together well with a practice-theoretical view on the nature of practices as such. Socially established practices can be taken to have purposes and are (among other things) determined by these purposes. They are what they are because of the purposes that are pursued by exercising them (Jaeggi 2013, 100). Hence, when an initially unintended effect is turned into a function by becoming the purpose of a newly established practice, this is also insightful with regard to the individuation of that particular practice. The practice is characterized by the function it is geared towards realizing. At the same time, the function is a function in the first place because it is embedded in the practice. To be clear, the idea is not that we should reject theoretical frameworks that back up function statements in cases in which there is no intention present as a matter of principle. Many of these frameworks can serve (other) epistemological interests. Still, I take it that retrieving the intentionalist stance by reference to practice-theoretical considerations as I have done in the solution sketched above has an explanatory value. To see how the solution can be applied to our three examples and in how far this has explanatory benefits, I will address each example individually, starting with the example of the smartphone. If we apply the practice-theoretical solution to this example, the following can be stated: there might indeed be actions that cause nomadic intimacy as an unintended side effect of smartphone usage. As indicated above, this is rather probable, given that smartphone users will often lack the respective sociological knowledge to even form intentions towards this particular effect. But if we adopt the solution presented above, these cases can be distinguished in an instructive manner from cases in which actions

Challenges of the Social Domain  195 constitute an established communicative practice that aims at nomadic intimacy. If we limit ascribing the function of providing nomadic intimacy to cases of intentionally realizing this practice, we win on two counts. On the one hand, we can preserve the intentionalist stance with all its benefits. On the other hand, we can individuate a communicative practice by marking out the effect that it is directed at as a function – after all, it is this function that makes the practice what it is. The solution presented above can also be insightful with regard to our second example of a nonsocial artifact with a putatively unintended social function, namely the example of the painting that has the effect of substantiating social hierarchies. Here again, we can distinguish between cases in which this effect comes about without any intentions and others in which it is embedded in a social practice and thereby transformed into a function. This is also instructive with regard to developments in dealing with art. By differentiating cases in which substantiating social hierarchies is an unintended side effect of a painting of modern art from cases in which this is a function embedded in a practice, it is possible to trace the historical development of a practice aiming at this particular way of dealing with items of modern art. As this practice is individuated by the respective function, referring to this function is also a means to precisely single the practice out. And what about our example of a supposedly unintended social function of a social artifact, namely the example of ascribing the function of serving the interests of citizens to the U.S. Senate? Drawing on the practice-theoretical solution outlined, we can also derive a helpful distinction that allows us to trace back a sociopolitical development. Differentiating between the Senate at first serving the interests of citizens as an unintended effect and the subsequent intended recognition and perpetuation of this effect – in the sense of the enhancement mechanism of learning anticipated by Pierson (Pierson 2000, 486f.) – allows for individuating a new sociopolitical practice pertaining to the Senate. At the same time, it allows for spotting the point of time in which the effect is transformed into a function.

Conclusion As we have seen, the social domain poses serious challenges to intentionalist accounts of artifact functionality. We also ascribe social artifact functions in cases in which no intention is present, i.e. neither on the part of the artifact’s creator nor on the part of its user(s). Notably, this equally holds for social functions both of nonsocial and of social artifacts. In light of this use of the term ‘function’, I have called intentionalist accounts of artifact functionality into question by way of three examples for function ascriptions that involve seemingly unintended effects. However, I have argued that despite the fact that these function ascriptions

196  Amrei Bahr can theoretically be backed up, we should maintain the intentionalist stance, albeit with a practice-theoretical element. If we acknowledge that unintended effects can be integrated into social practices and consequently mark them out as functions, this has explanatory benefits not only with regard to functions, but also to the practices that these functions are embedded in. Besides, this approach reveals that it is plausible to assume that there is a certain class of social functions that crucially depend on social practices. Extending the approach from the analysis of function statements to functional explanations could further illuminate the causal interplay between social practices and artifacts that perform certain functions because they are embedded in social practices.

Acknowledgment I would like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

Notes







Challenges of the Social Domain  197







12 Moreover, if we take nomadic intimacy to be relevant for social cohesion, this example would even allow to be analyzed as a social function in the commonly discussed sense as indicated above, since fostering social cohesion could reasonably be regarded as functional for a society as a social whole.





198  Amrei Bahr



References Achinstein, Peter. 1977. “Function Statements.” Philosophy of Science 44, no. 3: 341–367. Bahr, Amrei. 2019. “What the Mona Lisa and a Screwdriver Have in Common. A Unifying Account of Artifact Functionality.” Grazer Philosophische Studien 93, no. 1: 81–104. Baker, Lynne Rudder. 2004. “The Ontology of Artifacts.” Philosophical Explorations 7, no. 2: 99–111. Dipert, Randall R. 1995. “Some Issues in the Theory of Artifacts: Defining Artifact and Related Notions.” The Monist 78, no. 2: 119–135. Ehrenberg, Kenneth M. 2016. The Functions of Law. Oxford: Oxford University Press. Elias, Friederike, Albrecht Franz, Henning Murmann, and Ulrich Wilhelm Weiser. 2014. “Hinführung zum Thema und Zusammenfassung der Beiträge.” In Praxeologie. Beiträge zur interdisziplinären Reichweite praxistheoretischer Ansätze in den Geistes- und Sozialwissenschaften, edited by Friederike Elias, Albrecht Franz, Henning Murmann, and Ulrich Wilhelm Weiser, 3–12. Berlin: De Gruyter. Elster, John. 1979. Ulysses and the Sirens: Studies in Rationality and Irrationality. Cambridge: Cambridge University Press. Fortunati, Leopoldina. 2000. “The Mobile Phone: Towards New Categories and Social Relations.” In Sosiale konsekvenser av mobiletelefoni: proceedings fra et seminar om samfunn, barn og mobile telefoni, edited by Rich Ling and K. Thrane, 9–14. Kjeller, Telenor FoU. Geser, Hans. 2004. “Towards a Sociological Theory of the Mobile Phone.” Sociology in Switzerland: Sociology of the Mobile Phone. Online Publications, Zuerich (Release 3.0). Accessed December 1, 2019. http://socio.ch/mobile/t_ geser1.htm.

Challenges of the Social Domain  199 Haslanger, Sally. 2018. “What Is a Social Practice?” Royal Institute of Philosophy Supplement 82: 231–247. Hilpinen, Risto. 1993. “Authors and Artifacts.” Proceedings of the Aristotelian Society 93, no. 1: 155–178. Houkes, Wybo and Pieter Vermaas. 2010. Techical Functions. On the Use and Design of Artefacts. Dordrecht: Springer. Jaeggi, Rahel. 2013. Kritik von Lebensformen. Berlin: Suhrkamp. Jansen, Ludger. 2013. “Artefact Kinds Need Not Be Kinds of Artefacts.” In Johanssonian Investigations. Essays in Honour of Ingvar Johansson on His Seventieth Birthday, edited by Christer Svennerlind, Jan Almäng, and Rögnvaldu Ingthorsson, 317–337. Frankfurt: Ontos. Kroes, Pieter and Ulrich Krohs. 2009. “Philosophical Perspectives on Organismic and Artifactual Functions.” In Functions in Biological and Artificial Worlds. Comparative Philosophical Perspectives, edited by Pieter Kroes and Ulrich Krohs, 3–12. Cambridge, MA; London: MIT Press. McLaughlin, Peter. 2001. What Functions Explain. Functional Explanation and Self-Reproducting Systems. Cambridge: Cambridge University Press. Merton, Robert K. 1968. Social Theory and Social Structure. New York: Simon and Schuster. Neander, Karen. 1991a. “Functions as selected effects: The conceptual analyst’s defense”. Philosophy of Science 58, no. 2: 168–184. Neander, Karen. 1991b. “The teleological notion of ‘function’”. Australasian Journal of Philosophy 69, no. 4: 454–468. Pierson, Paul. 2000. “The Limits of Design: Explaining Institutional Origins and Change.” Governance 13, no. 4: 475–499. Reckwitz, Andreas. 2003. “Grundelemente einer Theorie sozialer Praktiken. Eine sozialtheoretische Perspektive.” Zeitschrift für Soziologie 32, no. 4: 282–301. Reicher, Maria E. 2013. “Wie aus Gedanken Dinge werden. Eine Philosophie der Artefakte.” Deutsche Zeitschrift für Philosophie 61, no. 2: 219–232. Searle, John. 1995. The Construction of Social Reality. New Haven: Free Press. Tuomela, Raimo. 2004. “The Philosophy of Social Practices: A Collective Acceptance View.” Cambridge: Cambridge University Press. Ullrich, Wolfgang. 2000. Mit dem Rücken zur Kunst. Die neuen Statussymbole der Macht. Berlin: Wagenbach. Wright, Larry. 1973. “Functions.” The Philosophical Review 82, no. 2: 139–168.

10 Revealing Social Functions through Pragmatic Genealogies Matthieu Queloz

Introduction There is an underappreciated tradition of genealogical explanation that is centrally concerned with social functions. I shall refer to it as the tradition of pragmatic genealogy. It runs from David Hume (T, 3.2.2) and the early Friedrich Nietzsche (TL) through E. J. Craig (1990, 1993) to Bernard Williams (2002) and Miranda Fricker (2007).1 These pragmatic genealogists start out with a description of an avowedly fictional “state of nature” and end up ascribing social functions to particular building blocks of our practices – such as the fact that we use a certain concept or live by a certain virtue – which we did not necessarily expect to have such a function at all. That the seemingly archaic device of a fictional state-of-nature story should be a helpful way to get at the functions of our actual practices must seem a mystifying proposal, however; I shall therefore endeavor to demystify it in what follows. My aim in this chapter is twofold. First, by delineating the framework of pragmatic genealogy and contrasting it with superficially similar methods, I argue that pragmatic genealogies are best interpreted as dynamic models whose point is to reveal the function – and noncoincidentally often the social function – of certain practices. Second, by buttressing this framework with something it notably lacks, namely an account of the type of functionality it operates with, I argue that both the type of functional commitment and the depth of factual obligation incurred by a pragmatic genealogy depend on what we use the method for: the dynamic models of pragmatic genealogy can be used merely as heuristic devices helping us spot functional patterns, or more ambitiously as arguments grounding our ascriptions of functionality to actual practices, or even more ambitiously as bases for functional explanations of the resilience or the persistence of practices. By bringing these distinctions into view, we gain the ability to distinguish strengths and weaknesses of the method’s application from strengths and weaknesses of the method itself.

Revealing Social Functions  201

The Nature and Point of Pragmatic Genealogies What are the characteristics of pragmatic genealogies? It quickly becomes apparent that they have little in common with the genealogies approximating regular historiography that one finds predominantly outside philosophy. 2 Unlike regularly historiographical genealogies, pragmatic genealogies do not in the first instance aim to describe the complex historical roots of a practice and its contingent transformations over the course of history. At the purely formal level which remains largely neutral between different substantive interpretations, the following are typical (though not necessary) features of a pragmatic genealogy: i

It begins not in a particular time and place, but in a state of nature, which differs from the Hobbesian version in that it already contains a small community of social and language-using creatures.



But how are these somewhat quaint-sounding state-of-nature stories best understood? And what are they supposed to tell us? One way of making sense of them is to read them as conjectural depictions of early hominid life in our so-called “environment of evolutionary adaption” (EEA). The state of nature would then be a stand-in for a historical situation about which we have very little data. For just this reason, we would then lack the means to decide whether things actually developed as the genealogical story presents them, but we might at least conjecture how things might possibly have developed. The state of nature would then be a narrative device by which to give “how possibly” explanations. An example of a genealogical approach that explicitly embraces this interpretation is Philip Kitcher’s The Ethical Project (2011).

202  Matthieu Queloz Where the pragmatic genealogists I listed at the beginning are concerned, however, things must be more complicated. This is because they explicitly deny that their state of nature is in the business of depicting, however conjecturally, the real history of our conceptual practices. Hume writes that the state of nature is “a mere philosophical fiction, which never had, and never cou’d have any reality” (T, 3.2.2.14). Williams insists that “the state of nature is not the Pleistocene” (2002, 27), and Craig likewise emphasizes that the “question ‘when?’ just doesn’t apply to it” (2007, 193). Philip Pettit (2018) drives the point home by calling his state of nature “Erewhon,” a Butlerian anagram of “nowhere.” How, then, are mere philosophical fictions depicting nowhere in particular supposed to tell us anything about reality? The answer comes into view once one sees genealogical fictions as akin to idealizing models in the sciences. Drawing on the interpretation of Craig’s genealogy developed by Martin Kusch and Robin McKenna (Kusch 2009, 2011, 2013; Kusch and McKenna 2018a, 2018b), I suggest that pragmatic genealogies are best interpreted as dynamic models – as idealizations with a time axis. These dynamic models start out with a strongly idealized situation, highlight certain practical needs, and show how these needs would drive a community to develop a prototypical form of a certain conceptual practice. The model is then de-idealized towards our actual situation by successively factoring in further needs: needs entailed by the initial needs the model started out with, but also, as in Williams’s case, needs factored into the model based on what we know about the actual history, sociology, and psychology of human beings. But what are these dynamic models for? The point of these genealogical models is to help us understand what our conceptual practices do for us, i.e. whether and how they tie in with our practical needs. To this end, pragmatic genealogies typically work to reveal social functions – to render visible unsuspected ways in which our conceptual practices meet social needs. Thus Hume offers a pragmatic genealogy of the virtue of “justice,” i.e. the virtue of respect for property, which exhibits it as a solution to conflict over external goods; Nietzsche presents truthfulness as a solution to the problem of deceit and dissimulation within the community; Craig describes the concept of knowledge as a tool by which to flag good informants and pool information; Williams shows how truthfulness facilitates the gaining and sharing of information; and Fricker highlights the importance of the virtue of testimonial justice in correcting for prejudice. All of these pragmatic genealogists seek to uncover an underappreciated social function performed by a conceptual practice – a respect in which it proves beneficial, not (or not just) to the individual who engages in it, but to the social community as a whole. The genealogical narrative is useful to this end because it perspicuously shows how, from certain practical needs we uncontroversially

Revealing Social Functions  203 have, a need can be derived which we did not necessarily know we had, namely the practical need for the target practice – the practice whose social functionality is to be revealed. In contrast to invisible hand explanations, pragmatic genealogies do not seek to account for the appearance of design by explaining how something obviously functional came to be so; rather, they serve to show how something that does not even look designed in fact turns out to be functional in nonobvious ways.3 Using a set of needs which one’s addressees identify as having, or which they at least recognize to community to have, we can tell a genealogical narrative showing how, given this set of needs, the target practice is in fact called for by the fact that this set of needs entails a series of further needs issuing in the need for the target practice. This genealogical derivation of needs from needs will then have the form: need A, hence need B, hence need C … hence need X, where need X is the need for the target practice. Granted that we actually have need A, the genealogical story will then give us reason to think that we also have need X.4 It is along just these lines that Williams (2002), for example, seeks to reveal the social function of a community’s valuing the truth for its own sake. To value the truth for its own sake, on Williams’s view, is in the first instance to value the various states and attitudes expressive of truthfulness, such as accuracy and sincerity, for their own sake (2002, 6–7). 5 To uncover the point of doing so, he begins with a state-of-nature situation highlighting the need of each individual to gather information about the immediate environment. This is a need that human beings have even on a highly generic conception of the person, which is to say that in ascribing such a need to our agents in the state of nature, we are not expressing a sociohistorically local conception of the person (we are not, for instance, expressing a distinctively liberal conception of the person by assuming that each individual has a strongly demanding need for autonomy). On the basis of this plausibly generic need-ascription, Williams then points out that the mere fact that individual inquirers occupy different spots at different times already entails that any given inquirer would, under certain circumstances, come to possess what Williams dubs a “purely positional advantage” (2002, 42) over other inquirers. And this means that there are strong pressures on these inquirers not to rely merely on their own senses in acquiring information, but to engage in the practice of sharing or pooling information. But this in turn generates the social need to cultivate in all participating inquirers the dispositions that make good contributors to the information pool: the dispositions which Williams brings under the capitalized headings of “Accuracy” and “Sincerity” to mark the fact that these are, at this point in the story, merely prototypical forms of what we now understand by “accuracy” and “sincerity.” They are as yet merely the dispositions involved in getting one’s beliefs right and openly passing them on to others.

204  Matthieu Queloz Yet as Williams’s dynamic model makes clear, development cannot stop there, as the need to cultivate dispositions that make good contributors to the pool itself leads us to consider further needs. The reason is that the dispositions in question cannot deliver what is demanded of them if they do not develop further. The practical value of the individual inquirer’s exhibition of these dispositions consists largely in their instrumental value to the community of inquirers: for the individual inquirer, Sincerity is rarely directly of much use, and Accuracy pays only to the extent that its benefits for the individual outweigh the costs to that individual – but there are many pieces of information that would be of great value to the community even if they are of no particular interest, or involve forbiddingly high risks and dangers, to the individual who could acquire them. Consequently, mere dispositions of Accuracy and Sincerity are overly vulnerable to the temptation to free ride, i.e. to profit from the Accuracy and Sincerity of others while failing to exhibit them reliably oneself. And since this holds true for any individual in the community, Accuracy and Sincerity in this form are not stable solutions to the problem of information pooling. A salient way to overcome this problem, Williams argues, is for Accuracy and Sincerity to come to be regarded as dispositions worth having and worth exhibiting for their own sake – they need to come to be regarded (and be commonly known to be regarded) as intrinsically valuable dispositions or virtues, as Williams puts it (2002, 89–90). And this in turn generates the need for individuals to be capable of making sense of these dispositions as intrinsically valuable, which, for Williams, means that their “value must make sense to them from the inside”  – they must be able to relate Accuracy and Sincerity “to other things that they value, and to their ethical emotions” (2002, 91–92). If we are to grasp how these prototypical forms of Accuracy and Sincerity have been fleshed out “now and around here,” however, and how they have been “changed, transformed, differently embodied, extended and so on” (Williams 2007, 132) in response to further needs that are more clearly historically and socially situated, the dynamic model must then be de-idealized in that direction. Williams consequently factors in, first, developments in ancient Greece that led to the extension of Accuracy to the distant past (2002, ch. 7); second, developments in eighteenth-century Europe that led to Sincerity’s elaboration into the value of authenticity (2002, ch. 8); and lastly, the needs of modern-day liberal democracies to cultivate Accuracy and Sincerity about politics and political history (2002, chs. 9–10, esp. 231–232, 265–266). Williams’s genealogy is not coincidentally about the social function of Accuracy and Sincerity. Such a pragmatic genealogical model serves to sharpen our eye for functional patterns within our actual conceptual practices. If it is to have a point, the functionality it reveals must be one that we are not already fully aware of, and social functions are often

Revealing Social Functions  205 particularly hard to discern. The functionality of practices for the individual are often easier to discern because looking out for individual benefits is something we do anyway, whereas the social point of view is one we only take up in special circumstances or upon reflection. Moreover, Williams’s genealogy helps explain why the social function of Accuracy and Sincerity is hard to discern: it brings out that the fact that we do not primarily think of Accuracy and Sincerity in functional terms at all – neither in terms of individual nor of social functionality – is an essential part of what renders them functional. Their very functionality demands that their functionality be effaced in favor of intrinsic motivations – which is why, elsewhere, I have called these particular functional dynamics the dynamics of self-effacing functionality (Queloz 2018b, §3). But a pragmatic genealogy derives its interest not just from the fact that it reveals functionality we are not yet aware of; it also derives it from its being something of a mystery how and why the target of the genealogy would have emerged in the first place. Why did individuals ever come to be Accurate and Sincere where it does not pay for them? Yes, this has to do with their valuing Accuracy and Sincerity for their own sake, but to leave it at that is simply to restate the puzzle: why did they ever come to value them intrinsically? Is this more than a fetish, more than a relic from the enchanted world in which God is truth and truth is divine, as Nietzsche (GM, III, §24) put it? In answer to these questions, we can usefully construct a model that renders perspicuous how and why Accuracy and Sincerity would arise naturally, without mysterious saltations, in response to individual and social needs – especially if that model can itself explain why these needs would come to include the need for Accuracy and Sincerity to be valued for their own sake, even without metaphysical reasons for doing so. Williams’s model to that extent naturalizes the intrinsic valuing of Accuracy and Sincerity, and uses this functionally justified intrinsic valuing to explain how a set of practices serving a social function could have arisen despite the fact that they were only of limited use to the individuals engaging in them. The genealogy bridges the gap between individual and social functionality on the back of an insight into the social function of intrinsic values. If pragmatic genealogies are dynamic models revealing social functions, however, it remains unclear which notion of functionality this particular brand of functionalism is supposed to operate with. The pragmatic genealogists themselves provide little guidance in this regard. Their writings leave it underdetermined whether their talk of “functions” should be cashed out in terms of a causal role account of functions à la Cummins (1975), where functions are ascribed to elements of a system – such as a heating system – on the basis of what they contribute to the realization of some systemic capacity we are interested in (such as the capacity to keep room temperature constant); or in terms of an etiological account of functions à la Wright (1973), Millikan (1989), and Neander

206  Matthieu Queloz (2017), where functions are ascribed to items based on what these items were selected to do; or an agentive functions account à la Searle (2010), where functions are imposed on objects by the purposes of agents, and even biological functions are thought of as causes that serve a purpose relative to agents’ values (58–60). On the one hand, it can be seen as a strength of these genealogical approaches that they remain ecumenical regarding the notion of function they involve. This invites one to try and plug in whichever notion one prefers or is interested in. On the other hand, it can also be seen as a weakness, especially in the light of the barrage of objections to functionalist approaches in the social sciences.6 These objections express and foster a discomfort with functionalist approaches that is bound to prove an obstacle also to pragmatic genealogies once their functionalist spirit is recognized. In the remainder of this essay, I will therefore try to dispel such discomfort by laying bare the functionalist innards of pragmatic genealogies.

Functions in and Functions of Pragmatic Genealogy The key to understanding how pragmatic genealogies work is to distinguish what they minimally and primarily do from what they can then be used to do on that basis. In the first instance, pragmatic genealogies serve to reveal instrumental relations between certain needs and certain conceptual practices within a fictional model. They issue in conclusions of the form: “The function of the prototype of X is to satisfy a need to Y.” For the pragmatic genealogists discussed above, this formula yields the following claims: • • • • •

The function of the prototype of the virtue of justice is to satisfy a need to avoid conflicts over external goods (Hume). The function of the prototype of the virtue of truthfulness is to satisfy a need to avoid deception within the community (Nietzsche). The function of the prototype of the concept of knowledge is to satisfy a need to flag good informants (Craig). The function of the prototype of the virtues of Accuracy and Sincerity is to satisfy a need to gain and share information effectively (Williams). The function of the prototype of the virtue of testimonial justice is to satisfy a need to correct for prejudice (Fricker).

Each of these claims involves an ascription of functionality: it highlights an instrumental relation between a prototypical concept or virtue on the one hand and a need as represented in the model on the other. This does not yet in itself carry any claims about our actual situation, past or present. But the point of working with such a model is to render visible similar instrumental need-concept or need-virtue relations in the actual history of our conceptual practices or within our current conceptual

Revealing Social Functions  207 practices. The dynamic model that is the pragmatic genealogy constitutes a useful guide to the discernment of such functional patterns because it represents them in their clearest, most generic form, free of the clutter and complexities of reality, thus sharpening our eye for similar functional patterns in reality – much as a priming look at a prototypical morel will assist the morel seeker in spotting morels of varying shapes and colors hidden underneath the tangle of twigs. And to the extent that we indeed find such instrumental relations in the past or the present, we can use that as a basis for functional explanations or functional assessments: insofar as the need-concept or need-virtue relation obtained in the past, this helps explain why we came to have the concept or virtue; insofar as the need-concept or need-virtue relation obtains now and around here, this gives us reasons for or against continuing to cultivate the concept or virtue. Moreover, the dynamic models provide prima facie evidence for these functionality ascriptions in much the same way that design analyses in evolutionary biology do: in both cases, a model is used to show that a given trait would solve a problem, and this is used as evidence for thinking that the traits we actually find actually solve similar problems (see Kincaid (1996, 118–119) for a discussion of design analyses in biology). An important difference is that in biology, design analyses typically turn on optimality arguments – arguments to the effect that a given trait, such as a certain foraging strategy or a reproductive strategy, can be mathematically shown to form an optimal solution to a problem. Pragmatic genealogies, by contrast, turn on what might be called indispensability arguments. These can be very roughly characterized as taking the following form: This characterization still begs numerous questions, including notably questions about how much like us these creatures and their environment are, and what the criteria for similarity between RN and RN* and P and P* should be – I say more on these questions in Queloz (Forthcoming-b). Moreover, it should be noted that (P3) does not figure prominently in all pragmatic genealogies, although it is an important and much celebrated feature of Hume’s genealogy that he traces out a path by which the virtue of justice might arise without much foresight or explicit coordination.7 But the point I want to press here is that this line of argument does not

208  Matthieu Queloz involve the idea that some behavioral trait can be mathematically shown to be optimal. The line of argument turns on the idea that a behavioral trait – more specifically, a concept or virtue – of some broadly outlined form, characterized only by its ability to discharge the function at issue, constitutes an indispensable solution to a problem that any creature with certain needs faces. The argument does not purport to show that P is optimal; it purports to show that P is conditionally necessary. On this account, the core of a pragmatic genealogy is an ascription of functionality relative to needs. Drawing on a certain understanding of what human beings need, certain concepts or virtues are presented as having, among their various actual and potential effects, such effects as tend either directly or indirectly to meet those needs. As Kincaid (1990, 1996, 2006, 2007) and others have argued, ascriptions of this sort are innocuous and capable of surviving scrutiny by those who are skeptical of functionalism in the social realm, because such ascriptions of functionality are straightforwardly identifiable with a set of causal claims: the practice of using or living by a concept or a virtue has certain effects; these effects contribute to the satisfaction of certain needs; and, given the transitivity of causation, the practice of using or living by the concept or virtue therefore helps satisfy these needs. With any ascription of functionality, there is a question about the extent to which the functionality in question is observer-relative; to what extent is this the case for the functionality ascriptions of pragmatic genealogy? Are they more like Searle’s agentive functionality ascriptions, which are dependent on the purposes agents happen to have? Or are they more like etiological ascriptions of “proper functions,” in Millikan’s terminology, which achieve independence from subjective purposes by basing ascriptions of functionality on objective selection histories? There are reasons to think that pragmatic genealogies are interestingly situated between these two poles. On the one hand, they involve an observer-relative dimension insofar as they take their basic normative orientation, which all functional talk requires in one way or another if it is to allow some sort of discrimination between the functional and the dysfunctional,8 from needs, and need ascriptions involve a substantial degree of interpretation: needs are something that an entity lacks as long as it is described merely in the vocabulary of physics; it is only once it is described in more normative terms that needs come into view. At the minimum, these terms must permit a distinction between survival and death, and more richly normative terms might allow for further distinctions, first between bare survival and flourishing, and then between flourishing by, in Williams’s phrase, “the ethological standard of the bright eye and the gleaming coat” (2011, 52) and more demanding (but also more socio-historically local) standards of flourishing, which may include such things as the need for autonomy or political selfdetermination. On the other hand, needs are also more objective than

Revealing Social Functions  209 ends and purposes. Unlike ends or purposes, needs are something one cannot decide to have – they are not subject to one’s will. Nor are needs subject to the will of the observer, just as their presence or absence does not depend on the observer’s ends or purposes. Furthermore, needs are not luminous, which is to say that one can have them without knowing that one has them. This is also part of what gives pragmatic genealogies their point, because they can reveal that we have certain needs which we may not have been aware we had. Hence, whether or not we have certain needs – such as a need for the virtues of Accuracy and Sincerity, or a need for the concept of knowledge – is, in all these respects, an objective matter and something that can come as a discovery. Functionality ascription, then, is the primary business of pragmatic genealogy; but once a pragmatic genealogy has suggested and buttressed a functionality ascription, this ascription can be used as an explanatory basis for functional explanations of varying ambitions. One thing that this explanatory basis might be used for is to explain the stability or, in Philip Pettit’s (1996, 2000) terminology, the resilience of the building blocks of our way of life: Which are the more or less passing ephemera and which the phenomena that are deeply embedded in the society? Which are more or less incidental or contingent features and which are features apt to last? There is an interesting research programme suggested by such questions. It would take any society or culture or institution and, reviewing the data on various traits displayed by the entity in question, would seek to separate out the dross from the gold. It would try to identify and put aside the features that may be expected to come and go. And it would seek to catalogue the more or less necessary features that the society or culture or institution displays. It would give us a usefully predictive stance on the society, providing us with grounds for thinking that such and such features are likely to stay, such and such other features likely to disappear. (Pettit 1996, 299–300) The research program described here might well be pursued using pragmatic genealogy as one’s method. Note, however, that this research program need not be backwards-looking at all: claims of resilience need not involve any factual claims about history being a certain way. To say that a concept or virtue is resilient because it stands in an instrumental relation to needs is not necessarily to say that it came to exist for that reason, or even that this instrumental relation played any role in its historical development. The conceptual practice could just have popped into existence, or it might have been instituted by a mad king on a whim, and yet it might truly be said to be resilient for functional reasons, i.e. in ways that could be explained by reference to the instrumental relations

210  Matthieu Queloz to needs that the conceptual practice, whatever its genesis, now stands in. Pettit (2000, 48) gives the example of golf clubs, asking us to imagine that they have come into existence purely because people enjoy the motions involved in playing golf. “Consistently with the absence of any such historical selection,” he notes, “what might well be the case is that golf clubs have certain effects, certain functional effects, such that were they to come under any of a variety of pressures, then the fact of having those effects would ensure that they survived the pressure,” and if so, we can conclude that “though not the beneficiaries of actual selection, golf clubs do enjoy the favor of a virtual process of selection” (2000, 48). Yet to say that the effects of golf clubs would ensure that they survived any of a variety of pressures is no doubt too strong, since it would imply that golf clubs are not just resilient, but, as one might put it, hyperresilient. The more modest claim advanced by a resilience explanation should rather be that if, as a community, we were to move away from golf clubs, this would rob us of some of their functional effects, and this loss would make itself felt (at least among those for whom golf clubs are presumed to have functional effects; the fact that this is likely to be a rather exclusive group, and that the functions in question are perhaps not functions anyone outside that group would want to see discharged in the first place, indicate further reasons why the ambitions of resilience explanations are better reined in). The conclusion we then reach is that there would be some pressure, however limited in strength and scope, to maintain golf clubs – though whether there would be enough pressure for them ultimately to survive in the face of countervailing pressures remains a further question. A second type of explanation that a functionality ascription might be used for is the explanation of the actual historical persistence of a concept or virtue. This still stops short of claiming that it came into existence because it serves a need (a claim that would raise questions about how as yet unrealized effects can bring something to exist). But it does involve committing oneself to the claim that the fact that the concept or virtue endured or was retained once it had come into existence had something to do with its relation to our needs. In sum, a pragmatic genealogy is in the first instance a narrative device by which to reveal and ground ascriptions of functionality which can then be used to explain the resilience or even the persistence of concepts or virtues. We thus get the following schema, with (1)–(3) together potentially acting as the explanatory basis for an explanation either of the form of (4) or of the form of (5): Functionality Ascription (Explanatory Basis) 1

The practice of living by concept/virtue A causes (sometimes via inferential consequences) the consequence B.

Revealing Social Functions  211 2 3

B helps satisfy need C. The practice of living by A helps satisfy C.

Functional Explanation 4 5

A is resilient because it helps satisfy C. A persists because it helps satisfy C.

The key idea that this schema is meant to bring out is that pragmatic genealogy minimally and primarily serves to substantiate (1)–(3), and it can, though it need not, also serve to substantiate (4) or (5), be it with regard to our present situation, with regard to some historical situation, or both. Two aspects of the schema bear further clarification, however. First, it is true that on a possible reading of (5), (5) claims no more than (4): in Kincaid’s (2007, 223) rendering, for example, one way for something to persist because it has certain consequences is for it to be resilient now, whatever its history, in virtue of the fact that there would be pressures to keep it if we moved away from it. By contrast, I distinguish (4) and (5) precisely to register the fact that we do undertake commitments about the history of the concept or virtue as soon as (5) is read as implying not just counterfactual claims, i.e. claims to the effect that certain forces would be actualized if a conceptual practice were to diverge from a certain functionally specified configuration, but factual claims to the effect that such forces were actualized and are part and parcel of the causal-historical story explaining why we now find the concept or virtue. Second, the qualification “sometimes via inferential consequences” in (1) registers the fact that the relevant practical consequences of living by a concept can be more or less immediate consequences of particular acts of concept application. Sometimes, the relevant practical consequence of living by a concept is simply the fact that this renders the concept-user suitably sensitive to the presence of certain items in the world; but sometimes, the relevant consequences lie further downstream, and the path towards them might lead through inferential consequences, i.e. through the inferences the concept-user is put in a position to draw by coming to live by the concept. Michael Dummett (1973, 454) gives the example of a student learning the concept of validity, and thus coming to be able to distinguish between valid and invalid arguments. But what gives the concept of validity its point is not the capacity to make this distinction in itself, but rather the practical difference made by the inference one can then draw from an argument’s being valid, namely that one has reason to accept the conclusion given that one accepts the premises. Consequently, a student who reliably applied the concept but never drew this inference – who treated the distinction between valid and invalid arguments as being like the distinction between Petrarchan and

212  Matthieu Queloz Shakespearean sonnets – could perhaps be said to have acquired the concept of validity, but when used in this way, the concept would fail to tie in with anyone’s needs. We can illustrate the schema with Craig’s genealogy. Craig’s dynamic model highlights the way in which social and language-using creatures would be driven to develop something like our concept of knowledge by two sets of practical pressures: the first set of practical pressures grows out of the fact that each individual needs information about his or her immediate environment, and is to that extent in the position of the inquirer: someone who wants to find out whether p. But for social and language-using beings like us, there are strong incentives to rely not just on one’s own senses in acquiring information, but to tap into others’ stores of information. This means that there are pressures on each inquirer to become able to identify what, given the particular needs and capacities of that inquirer, are good informants as to whether p. This the inquirer becomes able to do by developing the concept of what Kusch (2009, 65) aptly calls proto-knowledge – a concept that serves to flag good informants. Proto-knowledge is still markedly different from our concept of knowledge, however, in that it remains strongly indexed to the situation of the individual concept-user. It tracks whomever is a good informant for me, given my needs and capacities, here and now. But if we factor in the second set of practical pressures, we come to see why we in fact operate the concept of knowledge rather than the concept of proto-knowledge. This second set of practical pressures arises from the fact that inquirers have a strong interest in recommending informants to each other; and the more they do – the more they socially cooperate not just in exchanging information, but in exchanging information about who is a good source of information on a given question – the more they have reason to operate a concept that it less subjectivized than proto-knowledge. It is this second set of pressures which leads to the concept of someone who is a good informant whether p for anyone, whatever their needs and capacities, anywhere and at any time: someone, in other words, who knows whether p. The concept of knowledge we end up with in Craig’s model is thus revealed to perform a social function that is central to a kind of epistemic division of labor, namely the social practice of information pooling. Plugging Craig’s genealogy of the concept of knowledge into the schema then yields the following: Functionality Ascription (Explanatory Basis) 1 2

The practice of living by the concept of knowledge causes the flagging of good informants. The flagging of good informants helps satisfy the need to pool information.

Revealing Social Functions  213 3

The practice of living by the concept of knowledge helps satisfy the need to pool information.

Functional Explanation 4 5

Were the practice of living by the concept of knowledge to come under pressure, there would be some pressure to drive it back into use because it helps satisfy the need to pool information. The fact that the practice of living by the concept of knowledge helped satisfy the need to pool information in the past caused the concept to be retained.

To which of these claims does Craig’s genealogy commit him? As he himself points out, this depends on the purposes which the genealogical story is taken to serve. “The depth of factual obligation incurred by a state-of-nature theory depends on its aims,” he writes; it “will be greatest when its intentions are explanatory, to account for the existence of the target phenomenon” (Craig 2007, 193). This underscores the important methodological point that pragmatic genealogies in themselves – i.e. the bare genealogical narratives considered in isolation from the context and spirit in which they are advanced – do not yet determine how much is being claimed and what evidence they are beholden to. These parameters only receive determinate values once the pragmatic genealogy is put to use in a particular context with a view to performing a particular task, and these will be different values in different contexts. In Craig’s own case, the aim is in the first instance to cure us of the temptation to define our present concept of knowledge in terms of necessary and sufficient conditions by getting us to look at the concept from a pragmatic point of view. Once we let our understanding of the concept grow genealogically out of our understanding of the needs of inquirers, we will understand why we should not expect the application conditions of the concept of knowledge to be necessary conditions at all, but rather to reflect the typical conditions under which the concept has a point for inquirers given their needs. As we saw, this need in principle amounts to no more than to offering a model of a need-concept relation as a heuristic device by which to reveal whether such relations also obtain in past or present societies. But Craig aims to do more than that. In his later reflections on Knowledge and the State of Nature, he emphasizes that he “was trying to explain how certain real results have arisen, and only real pressures can produce real results” (Craig 2007, 190, emphasis mine). In accordance with this aim, he notes: I do and must suppose that there were societies whose members, collectively and individually, had the needs I ascribe to them and were able, whether as the outcome of some conscious process or of other equally real tendencies, to find their way to the solution I describe. …

214  Matthieu Queloz My line was, and had to be, that the needs were real and the persons concerned would have come, in one way or another, to satisfy them.  … I had to maintain that the circumstances that favour the formation of the concept of knowledge still exist, or did until very recently, since otherwise I would have had no convincing answer to the obvious question why it should have remained in use …. (Craig 2007, 191) On this Craigean reading of Craig, the pragmatic genealogy does not just serve as a device by which to sharpen our eye for (1)–(3); nor does it rest content to claim that (1)–(3) are now, for whatever reason, the case, and that therefore the concept of knowledge is now resilient in the sense of (4); rather, the genealogy is used to claim (5): to explain why the concept of knowledge persists. Of course one might invoke the Death-of-the-Author principle and insist that authorial intentions are not always the most reliable guide to a book, but if the account of pragmatic genealogy offered here is along the right lines, we might accept Craig’s self-interpretation, find it insufficiently corroborated by evidence and short on detail and mechanisms, and yet think no worse of his pragmatic genealogy, because the merits of Craig’s genealogy are distinct and separable from the use he made or took himself to make of it. We must distinguish the method – offering a dynamic model of the functional relations between needs and conceptual practices – from the use to which it is put. If, like the Craig of 2007, we harbor the ambition to use the genealogy outlined in Knowledge and the State of Nature to explain why the concept of knowledge persists, we will have incurred a different type of functional commitment – namely an etiological one – and correspondingly deeper factual obligations than if we used it simply to reveal the relation of the concept we now have to some of our present needs. Using the genealogy effectively to explain the persistence of the concept of knowledge might well require rather more supporting material than Craig had room for in his dense and admirably concise book. But the important point is that this does not invalidate the genealogy. The genealogy itself is merely a multipurpose model, a tool that earns its keep in many trades. Our conclusion, then, is that both the type of functional commitment and the depth of factual obligation incurred by a pragmatic genealogy depend on what we use it for. We can use the dynamic models of pragmatic genealogy merely as heuristic devices by which to sharpen our understanding of how certain needs bring certain problems and call for certain solutions while retaining an open mind as to whether a given society exhibits either those needs or anything like the modelled answer to them; the model would then serve to sharpen our eye for the needs and the functional patterns they tend to engender. But we can also use pragmatic genealogy to reveal (and to bolster our case for)

Revealing Social Functions  215 factual claims about the present or the past. On this basis, we might make backward-looking use of pragmatic genealogy, deploying it to account for the persistence and ubiquity of certain arrangements on the basis of what the genealogy reveals to be their social function. But we might equally make forward-looking use of pragmatic genealogy, either confidently to predict that certain arrangements are not going to go away, because the genealogy reveals them to be resilient; or, on the contrary, to highlight the reasons we have to cherish and defend these arrangements against countervailing pressures because the genealogy reveals unsuspected ways in which the satisfaction of certain needs depends on them. Pragmatic genealogy can thus not only be used to show what something would do for us if we were in a fictional state of nature; but also what it has done, what it now does, and what it can continue to do.

Notes



216  Matthieu Queloz

References Ariew, Andre, Robert Cummins, and Mark Perlman. 2002. Functions: New Essays in the Philosophy of Psychology and Biology. Oxford: Oxford University Press. Ayala, Francisco J. 1970. “Teleological Explanations in Evolutionary Biology.” Philosophy of Science 37, no. 1: 1–15. Baier, Annette. 1988. “Hume’s Account of Social Artifice: Its Origins and Originality.” Ethics 98, no. 4: 757–78. Barnes, Barry. 1995. The Elements of Social Theory. Cambridge: Cambridge University Press. ———. 2000. Understanding Agency: Social Theory and Responsible Action. London: SAGE. Binmore, Ken. 2005. Natural Justice. New York: Oxford University Press. Blackburn, Simon. 2008. How to Read Hume. London: Granta. Brandon, Robert N. 1981. “Biological Teleology: Questions and Explanations.” Studies in History and Philosophy of Science 12, no. 2: 91–105. Buller, David J. 1999. Function, Selection, and Design. New York: SUNY Press. Charron, William C. 1980. “Convention, Games of Strategy, and Hume’s Philosophy of Law and Government.” American Philosophical Quarterly 17, no. 4: 327–334. Cohen, G. A. 1978. Karl Marx’s Theory of History: A Defence. Princeton, NJ: Princeton University Press. Cohon, Rachel. 2008. Hume’s Morality: Feeling and Fabrication. New York: Oxford University Press. Craig, Edward. 1990. Knowledge and the State of Nature: An Essay in Conceptual Synthesis. Oxford: Clarendon Press. ———. 1993. Was wir wissen können: Pragmatische Untersuchungen zum Wissensbegriff. Wittgenstein-Vorlesungen der Universität Bayreuth, edited by Wilhelm Vossenkuhl. Frankfurt am Main: Suhrkamp. ———. 2007. “Genealogies and the State of Nature.” In Bernard Williams, edited by Alan Thomas, 181–200. Cambridge: Cambridge University Press. Cummins, Robert C. 1975. “Functional Analysis.” Journal of Philosophy 72: 741–764. Dummett, Michael. 1973. Frege: Philosophy of Language. New York: Harper and Row. Fricker, Miranda. 2007. Epistemic Injustice: Power and the Ethics of Knowing. Oxford; New York: Oxford University Press. Hampton, Jean. 1987. Hobbes and the Social Contract Tradition. Cambridge: Cambridge University Press. T: Hume, David. 2000. A Treatise of Human Nature, edited by David Fate Norton and Mary J. Norton. Oxford: Oxford University Press. Kavka, Gregory S. 1986. Hobbesian Moral and Political Theory. Princeton, NJ: Princeton University Press.

Revealing Social Functions  217 Kincaid, Harold. 1990. “Assessing Functional Explanations in the Social Sciences.” PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association 1990: 341–354. ———. 1996. Philosophical Foundations of the Social Sciences: Analyzing Controversies in Social Research. Cambridge: Cambridge University Press. ———. 2006. “Evolutionary Social Science Beyond Culture.” Behavioral and Brain Sciences 29, no. 4: 356–356. ———. 2007. “Functional Explanation and Evolutionary Social Science.” In Philosophy of Anthropology and Sociology, edited by Stephen P. Turner and Mark W. Risjord, 213–247. Amsterdam: North-Holland. Kitcher, Philip. 2011. The Ethical Project. Cambridge, MA: Harvard University Press. Kusch, Martin. 2009. “Testimony and the Value of Knowledge.” In Epistemic Value, edited by Adrian Haddock, Alan Millar and Duncan Pritchard, 60–94. Oxford: Oxford University Press. ———. 2011. “Knowledge and Certainties in the Epistemic State of Nature.” Episteme 8, no. 1: 6–23. ———. 2013. “Naturalized Epistemology and the Genealogy of Knowledge.” In Contemporary Perspectives on Early Modern Philosophy: Nature and Norms in Thought, edited by Martin Lenz and Anik Waldow, 87–100. Dordrecht; New York: Springer. Kusch, Martin, and Robin McKenna. 2018a. “The Genealogical Method in Epistemology.” Synthese. doi:10.1007/s11229-018-1675-1. ———. 2018b. “The Genealogy of Relativism and Absolutism.” In Metaepistemology: Realism and Anti-Realism, edited by Christos Kyriacou and Robin McKenna, 217–239. New York: Palgrave Macmillan. Lewis, David. 2002. Convention: A Philosophical Study. Oxford: Blackwell. Millikan, Ruth Garrett. 1984. Language, Thought, and Other Biological Categories. Cambridge, MA; London: MIT Press. ———. 1989. “In Defense of Proper Functions.” Philosophy of Science 56, no. 2: 288–302. ———. 2005. Language: A Biological Model. Oxford: Clarendon Press. Neander, Karen. 1991. “Functions as Selected Effects: The Conceptual Analyst’s Defense.” Philosophy of Science 58, no. 2: 168–184. ———. 2017. A Mark of the Mental: In Defense of Informational Teleosemantics. Cambridge, MA: MIT Press. TL: Nietzsche, Friedrich. 1979. “On Truth and Lies in a Nonmoral Sense.” In Philosophy and Truth: Selections from Nietzsche’s Notebooks of the Early 1870’s, edited and translated by Daniel Breazeale, 79–97. Atlantic Highlands, NJ; London: Humanities Press International. GM: ———. 1998. On the Genealogy of Morality, translated by Maudemarie Clark and Alan J. Swensen. Indianapolis, IN; Cambridge: Hackett. Nozick, Robert. 1974. Anarchy, State, and Utopia. Oxford: Blackwell. Pettit, Philip. 1996. “Functional Explanation and Virtual Selection.” British Journal for the Philosophy of Science 47, no. 2: 291–302. ———. 2000. “Rational Choice, Functional Selection and Empty Black Boxes.” Journal of Economic Methodology 7, no. 1: 33–57. ———. 2018. The Birth of Ethics: Reconstructing the Role and Nature of Morality, edited by Kinch Hoekstra. Oxford; New York: Oxford University Press.

218  Matthieu Queloz Queloz, Matthieu. 2017. “Nietzsche’s Pragmatic Genealogy of Justice.” British Journal for the History of Philosophy 25, no. 4: 727–749. ———. 2018. “Williams’s Pragmatic Genealogy and Self-Effacing Functionality.” Philosophers’ Imprint 18, no. 17: 1–20. ———. 2019a. “Genealogy and Knowledge-First Epistemology: A Mismatch?” The Philosophical Quarterly 69. doi:10.1093/pq/pqy041. ———. 2019b. “The Points of Concepts: Their Types, Tensions, and Connections.” Canadian Journal of Philosophy 49, no. 8: 1122–1145. ———. Forthcoming-a. “From Paradigm-Based Explanation to Pragmatic Genealogy.” Mind. doi:10.1093/mind/fzy083. ———. Forthcoming-b. “How Genealogies Can Affect the Space of Reasons.” Synthese Online First. doi:10.1007/s11229-018-1777-9. Rosenberg, Alex. 2016a. “Functionalism.” In The Routledge Companion to Philosophy of Social Science, edited by Lee McIntyre and Alex Rosenberg. London: Routledge. 147–157. ———. 2016b. Philosophy of Social Science. Boulder, CO: Westview Press. Searle, John. 2010. Making the Social World: The Structure of Human Civilization. Oxford; New York: Oxford University Press. Tieffenbach, Emma. 2011. “Invisible-Hand Explanations.” PhD diss., University of Geneva. ———. 2013. “Invisible-hand Explanations: From Blindness to Lack of Weness.” Social Science Information 52, no. 3: 450–470. Ullmann-Margalit, Edna. 1978. “Invisible-Hand Explanations.” Synthese 39: 263–291. ———. 1997. “The Invisible Hand and the Cunning of Reason.” Social Research 64, no. 2: 181–98. Williams, Bernard. 2002. Truth and Truthfulness: An Essay in Genealogy. Princeton, NJ: Princeton University Press. ———. 2007. “Truth and Truthfulness.” In What More Philosophers Think, edited by Julian Baggini and Jeremy Stangroom, 130–146. London: Continuum. ———. 2011. Ethics and the Limits of Philosophy. Routledge Classics. London; New York: Routledge. Wimsatt, William C. 1972. “Teleology and the Logical Structure of Function Statements.” Studies in History and Philosophy of Science Part A 3, no. 1: 1–80. Wright, Larry. 1973. “Functions.” Philosophical Review 82, no. 2: 139–168.

11 Social Organisms. Hegel’s Organizational View of Social Functions Daniel James

Introduction Analogies can be quite charming and, indeed, seductively so. Not only can they serve as powerful cognitive tools in exploring and explaining the nature of hitherto unfamiliar things in light of more familiar ones, as well as arguing for the working hypotheses we thus formulate about them. Moreover, they can serve as powerful rhetorical tools in making certain features of the thing in question more salient and, therefore, persuading others of the hypotheses and the arguments we formulate about them. However, to the extent that the cognitive and rhetorical effects of analogies can come apart, they can seduce us into accepting working hypotheses or arguments about things that are not founded by the relevant features they are taken to have in common. This seductiveness is what makes analogies, despite their undeniable charm, somewhat tricky.1 When it comes to both the charm and trickiness of analogies, social scientists are about as seducible as anyone else. Case in point: social functionalism. Part of what has made Jon Elster such a killjoy for all those social scientists eager to make use of functional explanations for their theoretical endeavors is that he has made them look like they have been not just charmed but, indeed, tricked by an analogy: the analogy between social and biological phenomena that underpins the very idea of functional explanation in social science. He traces this idea back to the emergence of social science as an autonomous discipline in the 19th century and distinguishes two versions of this analogy: first, the analogy between biological organisms and societies, both maintaining themselves; second, the analogy between biological organisms and firms, both struggling for survival (see Elster 1989, 75). Elster is quick – too quick, in my opinion – to dismiss the former version because it suggested “pseudoexplanations,” which gave rise to “pseudoproblems,” in turn (Elster 1989, 75). However, it’s easy to see why he’d be so quick to dismiss it. In the popular historical narrative, Darwin’s revolutionary achievement has not only eclipsed but thoroughly discredited previous approaches for their theological or speculative underpinnings (see Lenoir 1982,  3–4). With natural selection, so the story goes, Darwin

220  Daniel James had identified a mechanism that rendered the vitalistic and teleological notions on which these approaches relied obsolete. And, because these are the very approaches that, in turn, appear to underpin those versions of organismic analogy prominently advanced by the early sociologists Elster has in mind, it seems natural to consider it a non-starter. 2 In this chapter, I will challenge Elster’s dismissal of the organismic analogy. I will do so by examining a crucial episode in the history of the analogy he, in my view, rightly takes to underpin an influential account of functional explanation in social science. This is the very same episode within which he places the origin of the biological analogy, and which I consider to be a period of momentous conceptual innovation in social theory: the emergence of social functionalism in the 19th century. More specifically, I want to investigate what I take to be a central character within the backstory of social functionalism: G.W.F. Hegel. Thus, my investigation into his social theory will amount to a case study of the use made of the organismic analogy as a means of concept development. My central claim will be that Elster’s dismissal was premature and was so for two reasons. First, to claim that the organismic analogy figured among the premises of an analogical argument and, thus, as explanans in an explanation misses the point of the analogy. Hegel did not make an argumentative use of the analogy. Instead, he uses the organismic analogy to model the apparent close cooperation among the parts of the state and, thus conceptualize its characteristic structure in terms of organization: the mutual dependence among distinct structures generated by the state as a whole. Thus, it served as a means for the formation of the concept of social structure.3 Second, the organizational understanding of social functions suggested by the organismic analogy has not obviously been made obsolete by the triumph of the Darwinian method. Instead, reconsidering the organismic analogy in light of contemporary philosophy of biology puts an account of social functions that has fallen from view (and from grace) in the contemporary discussion back in focus: the organizational account. I will proceed as follows: first, I will outline Hegel’s theory of organisms, with a particular focus on his organizational understanding of functions, and spell out some of the metaphysical implications of this theory for both organisms as a whole and their parts. Second, I will reconstruct how he brings this understanding to bear on his account of the state. Third, I will discuss what I take to be the point of the organismic analogy. I will conclude my discussion by briefly considering its charm and trickiness for Hegel.

Organisms and Their Functions Hegel addresses the nature of organisms in two closely connected places within his system: first, in the context of his theory of “life” and, more

Hegel’s View of Social Functions  221 specifically, of “the living individual” – that is, Hegel’s term for an organism – presented both in the Science of Logic and in the Encyclopedia Logic; second, in the context of his philosophy of nature. However, these are not the only contexts within which the notion of an organism is relevant to him. Although Hegel’s discussion of “the living individual” within the Science of Logic and the Encyclopedia Logic in many respects echoes his theory of “the animal organism” within his philosophy of nature, he takes the “living individual” to realize a more general structure which he calls “logical life” (SL 12.180). Hegel takes “logical life” to be a structure of which some biological systems are but one realization and social systems another (see PR §§ 270–273, 279, 309). Thus, for Hegel, “natural life” as well as “life of spirit as spirit” (SL 12.181) are realizations of the same structure. As such, both kinds of systems share the general features he determines in his discussion of “logical life” (among which only those features characterizing the “living individual” will be relevant for my purposes). According to Hegel, a “living individual” is distinguished from other kinds of systems by realizing a structure he calls “self-determination.” As I will show in this section, self-determination is a structure characterized by two features. First, organizational differentiation, that is, the generation and maintenance of distinct structures which each contribute in a distinct way to the realization of a particular end – the organs or, as Hegel calls them, “members” [Glieder] of an organism – brought about by the organism as a whole; second, organizational closure, that is, the mutual constraint and, thereby, maintenance of an organism’s organs, such that it constrains and thereby maintains itself (see Mossio, Saborido and Moreno 2009).4 Thus, for an organism to determine itself is for it to maintain itself through generating distinct structures that mutually constrain and thereby maintain each other. On Hegel’s account, organizational differentiation and closure jointly give rise to means-end relations between the organism as a whole and its organs, which we can identify with their respective functions. Thus, these two notions also underpin Hegel’s account of functions in general and social functions in particular. Although he does not explicitly develop a theory of functions, we can reconstruct such a theory based on his account of the kind of systems the parts of which can bear functions: self-maintaining, self-organizing systems.5 Therefore, to understand how Hegel conceives of functions, we must have a closer look at how the notions of organization and self-maintenance figure in his account of the structure of organisms. As I will show, Hegel’s theory of organisms implies that functions are teleological, explanatory, and normative. He accounts for each of these features in terms of the self-determination of the organisms. To account for the nature of “the living individual” (SL 12.182–186; E § 218), Hegel draws on Kant’s analysis of the concept of a “natural end”

222  Daniel James [Naturzweck] and on his understanding of the “intrinsic purposiveness” characteristic of such objects (see KU § 63–66). It is therefore instructive to begin by having a brief look into Kant’s account of these notions. In his Critique of Judgement (KU), Kant introduces the term ‘natural end’ to describe things that meet the conditions for a teleological judgment of the form ‘M serves E’ to be true about them (see Kreines 2005, 275). He provisionally glosses this term by stating that “a thing exists as a natural end if it is (though in a double sense) both cause and effect of itself” (see KU § 64). Roughly, on Kant’s account, for a system to constitute a “natural end,” it must meet two conditions, which both concern the nature of the means-end relationship between the parts and the whole of a system claimed in teleological judgment (see Kreines 2015, 101 ff.). First, the beneficial effects of the parts for the whole explain the existence, structure, and arrangement of the former (see KU § 63). Second, the existence, structure, and arrangement of each part is explained by all the other parts and thus by the system as a whole (see KU § 65). A  “product of nature” (KU § 64 – cf. § 65) that meets both of these conditions would be “an organized and self-organized being” (KU § 65), or an “an organized natural product […] in which every part is reciprocally both end and means” (KU § 66). On Kant’s account, these mutual instrumental relations among the parts of a “product of nature” amount to the “intrinsic purposiveness” characteristic of “natural ends.” Hegel adopts these concepts from Kant to account for the characteristic structure of organisms (see SL 12.182–186). As he puts it explicitly, “the purposiveness of the living being is to be grasped as inner” (SL 12.184). Likewise, he draws on Kant’s concept of a Naturzweck when he characterizes living beings as objects in which “all members are reciprocally momentary means as much as momentary purposes” (E § 216). For Hegel, as much as for Kant, such objects are distinguished from mere aggregates by the fact that the occurrence, nature, and arrangement – in a single word: the “organization” (see E § 352 add. – cf. E § 198 add.) – of their parts is explained in terms of their respective beneficial effects for one another and thus for the organism as a whole. Taking his cue from Kant, Hegel first spells out his take on the “inner purposiveness” of an organism in terms of a particular kind of instrumental relation between the organism as a whole and its organs: an explanatory relation. As he puts it, the organism is “the means and instrument of purpose, fully purposive,” but he adds that “precisely for this reason this means, and instrument is itself the accomplished purpose in which the subjective purpose thus immediately closes in upon itself [unmittelbar mit sich selbst zusammengeschlossen ist]” (SL 12.184). In light of Kant’s concept of the Naturzweck discussed above, I tentatively interpret Hegel’s redescription of the inner purposiveness of organisms as follows. An organism is “fully purposive” in the sense that each of its organs, in virtue of their respective structure and their

Hegel’s View of Social Functions  223 arrangement, brings about distinct beneficial effects for the organism as a whole, such that they jointly serve as a “means and instrument” for the purpose in question. When Hegel furthermore characterizes the arrangement of such organs as “the accomplished purpose,” he suggests that it is the result or “product” (SL 12.165) of a goal-oriented “activity” [Thätigkeit] (E § 218) or process brought about by the organism as a whole.6 Lastly, the fact that Hegel speaks of the (singular) purpose of an organism suggests that it is the very same purpose which, on the one hand, each organ serves, and, on the other, the “product” of which they are insofar as they compose an organized whole. This means that, as its organs, the parts of an organism do not merely serve the purpose in question; instead, this purpose explains their organizational differentiation insofar as it is brought about by an “activity” (E § 218) that is directed towards the realization of that very purpose. To anticipate my discussion below, the end in question is the self-maintenance of the organism as an individual whole of a particular kind. Therefore, what makes a composite system an organism is that its parts are, jointly, not only “means and instrument” in that they, by virtue of their organization, serve the end of its self-maintenance, but also “the accomplished purpose” in that their organization is directed towards the realization of that very end. Thus, Hegel takes the organization of the “living individual” to be a particular kind of process that is brought about by the organism as a whole and directed at its self-maintenance. At various places, Hegel refers to this process as “the self-determination of the living being” (SL 12.185 f., 12.182, 12.183, 12.187, E § 355 add.). My central interpretative claim in this section will be that such a process constitutes self-determination by virtue of organizational differentiation and closure. To argue for this claim, I will first show how these notions each figure in his theory of organisms. In his Encyclopedia Logic, Hegel describes the organizational differentiation of a system as “the process of the living within itself, in which it divides itself in itself and makes its corporal condition [Leiblichkeit] its object, its inorganic nature” (E § 218). By the “inorganic nature” of an organism, he means the chemical and mechanical processes through which the distinct structures it generates and maintains – as he puts it, “this inorganic side, as the relatively external, enters into the difference and opposition of its moments” (E § 218) – contribute to its self-maintenance. In his discussion of the animal organism, Hegel calls the process generating and maintaining such structures “structuring” [Gestaltung] or “articulation” [Gliederung]: Structuring [Gestaltung], as alive, is essentially process, and it is, as such, abstract process, the structural process [Gestaltungsproceß] within structure itself in which the organism converts its own members into a non-organic nature, into means, lives on itself and

224  Daniel James produces its own self, i.e. this same totality of articulated members [Gegliederung], so that each member is reciprocally end and means, maintains itself through the other members and in opposition to them. (E § 356 – translation altered) As Hegel makes clear here, the organization of the parts of an organism into an individual whole of organs—a “totality of articulated members”— is a process of “articulation” [Gliederung]. As he puts it in the Science of Logic, it is a process of “shaping itself inwardly” (SL 12.187), a process that is brought about by the organism itself. Here, Hegel moreover suggests that this process involves organizational closure. He claims that, within this process, the “moments” or organs of an organism “reciprocally surrender themselves, the one assimilating the other to itself, and preserve themselves in the process of producing themselves” (E § 218). This claim suggests that, in this process, each of the organs of an organism is subjected to constraints exerted by the others and that the characteristic activity of each organ is maintained through these mutual constraints. In this sense, the organs of a living being are subservient to each other and, thereby, to the living being as a whole. It is also the sense in which, as Hegel puts it in a deliberately Kantian turn of phrase, “all members are reciprocally momentary means as much as momentary purposes” (E § 216). Thus, he characterizes the relations among the organs of a living being as relations of mutual dependence among distinct structures that mutually constrain and thereby maintain their activity. In his discussion of the “animal organism,” he states this characterization even more explicitly: Each abstract system permeates, and is connected with, them all, each displays the entire structure […]; and this gives interconnection to the organism, for each system is dominated by the others with which it is interlaced and at the same time maintains within itself the total connection. (E § 355 add.) Hegel describes both the organizational differentiation and closure among a living being’s organs as a process of “systems uniting to produce a general, concrete interpenetration of one another so that each part (Gebilde) of the structure contains these systems linked together in it” (E § 355). He illustrates this idea by way of the example of the cardiovascular system’s capacity to circulate blood and thereby facilitate the transport of nutrients to and waste away from cells. For, on his account, the cardiovascular system only has this capacity because of its interaction with the other organ systems, for instance, because it is controlled by the (vegetative) nervous system. He generalizes this idea when he claims that “the other systems actually exist in each: blood and

Hegel’s View of Social Functions  225 nerves are everywhere present, and everywhere, too, here is a glandular, lymphatic substance, that which constitutes reproduction” (E § 354 add. – cf. Stern 1990, 101–102). In describing the process through which a living being brings about the organizational differentiation and closure of its organs as one in which “the subjective purpose […] immediately closes in upon itself”, Hegel introduces a second condition for the “inner purposiveness” of organisms, which accounts for how the organization of its parts to a whole of organs or “members” is explained in terms of their beneficial effects for the organism as a whole. This characterization suggests that the goal-directed process which explains this organization must, in some way, relate to itself. I take Hegel here to point to a particular kind of circular causal regime, in which the effects brought about by the organization of a system contribute to the maintenance of that very system itself, understood as an individual whole of a particular kind. In other words, he here points to the self-maintenance of an organism, the end state resulting from the characteristic activity of its organs. Along these lines, he states in the Science of Logic that in the “process of the living individuality” (SL 12.185) of an organism – its self-maintenance – through the constraint and thus maintenance of its organs’ activity – “the product, being its essence, is itself the producing factor” (SL 12.185). In the corresponding section of his Encyclopedia Logic, he states in the same spirit that the “activity of the members […] is only one activity of the subject, the activity into which its productions go back, so that through that activity only the subject is produced, i.e. it merely reproduces itself” (E § 218). These statements suggest that, through organizational differentiation and closure, an organism continually ‘reproduces’ or maintains itself. This is, in my view, what Hegel has mind when he speaks of “the self-determination of the living being.” What crucially distinguishes self-determination from other kinds of circular causal regime is that this process is, as Hegel puts it in the Science of Logic, a “purpose unto itself [Selbstzweck]” (SL 12.187, 12.176 f.). He thus distinguishes this causal regime in teleological terms. In his discussion of the “animal organism,” he echoes this characterization when he claims that “[a]s its own product, as self-end [Selbstzweck], animal life is End and Means at the same time” (E § 252 add.). Thus, for a self-maintaining system to be self-determining, it must stand in an instrumental relation to itself (see McLaughlin 2000, 211): Its activity must contribute to generating and maintaining the very conditions of its own ongoing existence. Hegel spells out this relation as follows: End is an ideal determination which is already existent beforehand; so that, in the process of realization which must fit in with what exists determinately beforehand, nothing new is developed. The realization is equally a return-into-self. The accomplished End has the same content as that which is already present in the agent; the living

226  Daniel James creature, therefore, with all its activities does not add anything to it. As the organization [of life] is its own End, so too it is its own Means, it is nothing merely there. (E § 352 add.) Hegel here identifies the “End” of the organism’s activity with “an ideal determination which is already existent beforehand,” that is, with the end-[state] that is maintained through constraining and thereby maintaining the activity of each of its organs. Given that the organism thus contributes to maintaining the conditions of its very own existence, the “realization” of this end is “equally a return-into-self,” in which “[t] he accomplished End has the same content as that which is already present in the agent.” In this sense, an organism is, for Hegel, “Ends and Means at the same time.” The basis for Hegel’s description of this kind of circular causal regime in teleological terms is its organizational closure, that is, the mutual constraint the distinct structures involved exert over each other, such that they mutually maintain their activity. Hegel takes this regime of mutual constraint among its organs to amount to the self-constraint of the organism as a whole.7 In other words, he takes the activity through which the organism as a whole constrains and thereby maintains the activity of its organs to be directed at constraining and thus maintaining itself. Self-determination is, therefore, the goal-directed process through which an organism constrains and thereby maintains the activity of its organs, such that it constrains and thereby maintains itself as an individual whole of a particular kind. Hegel calls self-determination, thus understood, “the initiating self-moving principle” (SL 12.183) of life. In light of this interpretation of how Hegel accounts for the selfdetermination of an organism, I can now explicate the understanding of functions implicit in this account. Broadly, functions are, on Hegel’s account, means-end relationships between the characteristic activity of each organ and the self-maintenance of the organism as a whole which obtain because of the self-determination of the organism (see DeVries 1991). What, thus, distinguishes functions from any other effect for the organism as a whole (including beneficial ones) is that they are distinct contributions to its self-maintenance, which are themselves maintained through organizational closure. Thus, a part M of a living being has a function just in case it is subject to organizational closure in an organizationally differentiated self-maintaining system S. This definition implies that M meets the following three necessary and jointly sufficient conditions:8 C1: M contributes to the maintenance of the organization O of S; C2: M is generated and maintained under some constraints exerted by O; C3: S is organizationally differentiated.

Hegel’s View of Social Functions  227 This organizational view has three crucial implications concerning the nature of functions. First, it implies an objective teleology, meaning that each organ possesses its function independently of whether we ascribe it or not. What grounds this objective teleology is the selfdetermination of the organism, that is, the activity through which the organism as a whole constrains and thereby maintains the activity of its organs that is directed at constraining and thereby maintaining itself. Second, like many contemporary theorists of functions in biology and sociology, Hegel takes functions to be explanatory in that they explain why the function bearer’s characteristic activity is maintained (see Wright 1973, 155–157). Third, Hegel’s view of functions also involves a normative dimension in the sense that he takes its bearer’s distinct contribution to the self-maintenance of the organism as a whole to be an effect it ought to produce. According to this view, the normativity of any given function just is what is required for a self-maintaining system to persist. Insofar an organism depends on its own activity – the constraint and thereby maintenance of its organs’ activity, such that they each contribute to its self-maintenance – for its persistence, each of its organs is, given this “self-end” of the activity of the organism as a whole, required to act such that it contributes to its realization. Hegel’s account of organisms and the functions of their organs in terms of self-determination has metaphysical implications for the nature of both organisms as a whole and their organs. Discussing these implications in any detail is beyond the scope of this chapter – but let me at least gesture at three such implications that will become relevant when I discuss Hegel’s account of the structure of the state below.9 First, organisms are, by their very nature, not things but, rather processes or, more specifically, hierarchies of mutually constraining and thereby maintaining processes.10 On Hegel’s account, part of what it is for a composite system to be an organism, as opposed to a mere aggregate, is for it to continually constrain and thereby maintain the activity of its organs and, thus, itself as an individual of a particular kind. Therefore, he partially identifies the organism with the very process of continually maintaining the activity of its organs (through their constraint) that is directed at its self-maintenance. More precisely, he identifies this process of self-determination, with the characteristic “form” or, to use a contemporary term, the structure of an organism (see E § 350 add.). Hegel’s talk of “form,” in my view, suggests that he here relies on some version of hylomorphism, that is, the view that the natures of entities are composites of matter and form. On this view, to account for the nature of any given object – that is, to explain why and how it possesses its characteristic features –, one must appeal to both their material and formal components (see Austin 2018, 311 – cf. Oderberg 2011).

228  Daniel James With regard to organisms, the process of self-determination is a part of the nature of an organism that is distinct from its material components insofar as this process persists through the continuous replacement of the matter out of which an organism is, at any given time, made up. Instead, this part consists of the entire hierarchy of mutually constraining and thereby maintaining processes these material components are subjected to, and that is directed at maintaining the organism as a whole. The mutual dependence among the organs of a living being, in turn, gives rise to instrumental relations of each of them to the others and of the entire organism to itself. It is, therefore, this teleological structure – the structure of self-determination – that makes an organism the kind of entity that it is. Second, that what makes a part of an organism its organ is that it is essentially dependent on it as a whole. For Hegel identifies the organs of a living being with those mutually constraining and thereby maintaining processes that compose the organism as a whole. However, other than the hierarchy of processes they compose, they do not maintain themselves. Instead, they depend for their persistence on the constraint and maintenance by the organism of which they are each part, such that they contribute to its self-maintenance. Thus, they could not persist as the processes they are if they did not possess this function (see Stern 1990, 104–106; DeVries 1991, 58). As such, organs have, to adopt an apt phrase from David Oderberg, “no life of their own” but only insofar as they have a particular function for – or are subservient to – the organism of which they are a part (see Oderberg 2018, 363). On Hegel’s account, this function thus makes each organ the kind of entity it is. Indeed, this is part of what it is for a part of an organism to be its organ. As such, he identifies the organs of a living being with the “moments” – or the subservient parts – of its “form,” the nature of which is each determined by their respective functions. In turn, it is the “form” or structure of self-determination that explains why an organism has its characteristic morphological features. Given that the nature of each of the processes that constitute an organism is determined by its function, we can thus account for its structure – the “formal” component of its nature – in terms of instrumental relations in which its organs stand to each other and the organism as a whole stands to itself. Insofar as Hegel takes the political state to be characterized by self-determination as well, he will bring this processual and teleological understanding of structure to bear on the nature of the political state. Indeed, thus conceptualizing the structure of the political state is part of the point of the use he makes of the organismic analogy.

The Organism of the State As I suggested above, Hegel believes that there are both biological and social (“spiritual”) systems which are likewise characterized by

Hegel’s View of Social Functions  229 self-determination. As such, they will also likewise display organizational differentiation and closure. To the extent that any given social system displays these processes, they will likewise jointly give rise to the instrumental relations between the system in question and some of its parts I identified with the respective functions of these parts. Therefore, Hegel’s account of the nature of organisms in terms of self-determination will also underpin his view of social functions. His theory of social functions thus rests on an analogy between biological and social systems of a particular kind, namely self-determining systems. For Hegel, there is one social system in particular that manifests self-determination as outlined in this section: the state. As such, the state is, for Hegel, the paradigmatic bearer of social functions. Thus, Hegel commits himself to an organizational view of social functions. Crucially for my purposes in this chapter, Hegel brings the metaphysical implications of this view to bear on how he accounts for the structure of the state as a particular process of self-determination, the parts of which, in turn, are characterized in terms of different functions. Hegel thus suggests two hypotheses about the nature of the state: first, that the state, too, just is a particular kind of process: the process of self-maintenance through self-constraint; second, that the organs of the state are also essentially dependent on it. Hegel will bring both metaphysical conclusions to bear on how he conceptualizes the characteristic structure of the state. Thus, he accounts for the social structure of the state in terms of his organizational view of functions. To reconstruct this view, we must understand what it means to regard the state as a selfdetermining system. In both the Philosophy of Right and the corresponding passages of the Encyclopedia, Hegel emphasizes, in line with his general characterization of the “living individual,” that “the individual state” is (at least under certain conditions) “a self-relating organism” (PR § 259, §§ 267, 269 f.), “the universal which has the universal as such as for its purpose,” “organized in itself” (Griesheim: 635 – my translation), and, indeed “a living unity” (PR § 272, remark), “a living spirit pure and simple” (E § 359) or, simply, “alive” (Griesheim: 638 – my translation). As such, the state is, much like the “animal organism,” distinguished from a mere aggregate insofar as its identity depends on its self-determination: Its activity of continually constraining and thereby maintaining the activity of its organs such that it constrains and thus maintains itself (see Wolff 1985; Sedgwick 2001). The notion of the “organism of the state” (PR § 267) has its systematic place in Hegel’s discussion of the “properly political state and its constitution” (PR § 267), that is, in the section on the constitution or “right within the state” (PR §§ 267–269). I will, therefore, focus my discussion on this section. Here, Hegel discusses the circular causal regime of social processes through which the “political state” constitutes a social organism, that is, a social system that is characterized by organizational

230  Daniel James differentiation and closure among its parts. This regime comprises the aptly labele constitutional organs of a constitutional monarchy: “the Legislative Power” (PR §§ 298–320), “the Executive Power” (PR §§ 287–297), and “the Crown” (PR §§ 275–286). As I will show in what follows, the constitution of the political state comprises both the organizational differentiation and closure among the constitutional organs, such that they each make a distinct contribution to the self-maintenance of the political state. As such, the political state is, on Hegel’s account, a self-determining system in the sense discussed above. Therefore, we can identify the distinct contributions its constitutional organs make to the self-maintenance of the political state of which they are part (by virtue of the organizational differentiation and closure among them) with their respective social functions. In his discussion of the “political state,” Hegel explicitly identifies its constitution with the “organism of the state” (PR § 267 – cf. § 269, 270, add.), and its powers with its organs. In doing so, he suggests that the “political constitution” (PR § 269) just is the particular way in which any given state is organized. In the same vein, he introduces the notion of the “constitution” in section 539 of the Encyclopedia as the “overall articulation of state-power [Staatsmacht]” (E § 359 – emphasis mine). In thus identifying the constitution of the state with its organization, he exploits an ambiguity in the German term for ‘constitution’: ‘Verfassung.’ For, while this term typically stands for the political constitution of a state (as it may be codified in constitutional law), it can – particularly in its adjectival form “(auf eine bestimmte Weise) verfasst sein” – also stand for the particular way something is structured. Thus, in adopting the notion of an organism to account for the nature of the state, Hegel draws on the organismic analogy to conceptualize its peculiar structure, a structure he calls self-determination. I, therefore, take conceptualizing the structure of the state to be the point of the organismic analogy in his theory of the state. He presents the outlines of this analogy in sections 269–271 of the Philosophy of Right, in which he discusses the “organism of the state” as well as its “political constitution.” Thus, in section 269, he states: This organism [of the state] is the development of the Idea to its differences and their objective actuality. Hence these different aspects are the various powers of the state with their functions and spheres of action [Geschäfte und Wirksamkeiten], by means of which the universal continually engenders itself […]. Throughout this process, the universal maintains its identity since it is itself the presupposition of its own production. This organism is the political constitution. (PR § 269) As Hegel first makes clear here, he identifies the political constitution of a given state not with a set of legal norms codified in constitutional law, but rather with a particular kind of goal-directed process, which he here

Hegel’s View of Social Functions  231 also calls “development” [Entwicklung (see PR §§ 270 add., 271)]. But what kind of process is the “development of the Idea to its differences and their objective actuality” to which these constitutional norms are subjected and with which Hegel here identifies the “political constitution”?11 In the passage cited above, Hegel drops two hints concerning the nature of this “development”: whereas the first concerns the organizational differentiation and closure among the constitutional organs of the political state, the latter concerns the self-maintenance of the political state as a whole. Together, these features amount to the selfdetermination of the political state. We can find the first hint in his claim that this process results in “[the Idea’s] differences and their objective actuality,” which he, in turn, identifies with “the various powers of the state with their functions and spheres of action.” In a later section, he reiterates this description when he characterizes the political constitution as “the organization of the state and the self-related process of its organic life, a process whereby it differentiates its moments within itself and develops them to selfsubsistence” (PR § 271). Both descriptions echo Hegel’s account of what I have above called ‘organizational differentiation’: the process through which an organism generates and maintains distinct structures which each contribute in a distinct way to the self-maintenance of the system of which these structures are each part. In case of the state, the structures generated and maintained by its political constitution are “the various powers of the state” or, in other words, the constitutional organs. As Hegel moreover makes explicit in his later discussion of the “division of powers within the state” (PR § 272), this process results not only in organizational differentiation but, indeed, in the organizational closure among the constitutional organs that constitute the political state. He brings out this point by way of a contrast with what he takes to be a fundamental misconception of the division of powers on the part of what he calls “the abstract understanding.” He characterizes this misconception in terms of three, in his view, mistaken assumptions (with the latter two being implications of the first). The first assumption concerns the nature of the constitutional organs themselves, the second the nature of their relation to each other, and the third the attitudes of the individual agents involved in the activity of the constitutional organs towards each other. For my purposes, only the first two of these assumptions are relevant. The first assumption amounts to “the false doctrine of the absolute self-subsistence of each of the powers against the others” (PR § 272). In the corresponding section of the Encyclopedia, he states this conception of the division of powers in terms of “their independence of each other in existence” (E § 541, remark). This statement suggests that, according to this doctrine, the constitutional organs exist and act independently of each other. Along the same lines, Hegel, in the addition to section 272 of the Philosophy of Right, speaks of “the monstrous error of so interpreting their [the powers of the state] distinction as to suppose

232  Daniel James that each power should subsist independently [für sich] in abstraction from the others.” (PR § 272 add.). Thus, on this view on the division of powers, the political state is merely an aggregate of constitutional bodies that exercise their respectively characteristic activities independently of each other. Moreover, this view implies the second assumption, which concerns the nature of the constitutional organs’ relation to each other. As Hegel puts it, the “abstract understanding” of the division of powers “interprets their relation to each other as negative, as a mutual restriction” (PR § 272, remark). This interpretation amounts to the view that the causal interaction among independently existing constitutional bodies is such that they mutually inhibit the manifestation of their respective powers. According to Hegel’s portrayal of this conception, “their function is to oppose one another” (PR § 272, remark). In other words, the causal interaction among the constitutional powers is – as he puts it metaphorically, echoing the characterization thereof as ‘checks and balances’ – a “strife” (PR § 272, remark), in which they act as “counterpoise [Gegengewichte]” and “dikes [Dämme]” (PR § 272, remark) for each other. As such, their interaction is an aggregative, mechanical process (cf. SL.12 140–142) among antagonistic powers – in Hegel’s expression: “the mechanism of a balance of powers external to each other” (E § 544) which, at best, can merely “bring about a general equilibrium, but not a living unity” (E § 544). By contrast, part of why the constitutional bodies constitute “a living unity” on Hegel’s account is that they are mutually dependent. As he puts it, echoing his account of the mutual dependence among the organs of a living being, “each of these powers is in itself the totality, because each contains the other moments and has them effective in itself” (PR § 272). In light of his general theory of organisms, I take this statement to imply that the constitutional organs constrain and thereby maintain, rather than inhibit each other’s activity. This causal regime of mutual constraint amounts to the organizational closure among the constitutional organs. In light of Hegel’s general claim that such a regime of mutual constraint among its organs is tantamount to the self-constraint of the organism as a whole, we can, moreover, take the individual whole they constitute – the political state – to constrain itself. Thus, the process of “development” that maintains the characteristic activity of the constitutional organs is brought about by the political state as a whole. This claim leads directly to Hegel’s second hint concerning the nature of the process of “development” with which he identifies the political constitution. For his characterization of the organizational differentiation of the constitutional bodies as a process “by means of which the universal continually engenders itself” suggests that it is directed at constraining and thereby maintaining the political state as a whole. In

Hegel’s View of Social Functions  233 light of my interpretation of Hegel’s theory of organisms, I take him to here describe the process through which the political state maintains itself through the characteristic activity of its organs or “the various powers of the state.” He states this understanding of the political constitution more explicitly in the Griesheim lecture transcript when he claims that “[t]he constitution is […] alive, active in this manner [auf diese Weise thätig]” insofar as it must “always be produced [hervorgebracht]” (Griesheim: 697, my translation) as the particular kind of process it is. This claim is again echoed in the addition to section 270 of the Philosophy of Right, where he reiterates that the “political constitution […] is produced perpetually by the state, while it is through it that the state maintains itself” (PR 270, add.). Thus, through the organizational differentiation and closure of its constitutional organs, the political state contributes to generating and maintaining the very conditions of its own ongoing existence. For this reason, Hegel also describes the political constitution as “the self-related process of its [the state’s] organic life” (PR § 271). As I have shown in the previous section, Hegel generally characterizes this kind of self-relation in teleological terms: as he puts it, it is a “purpose unto itself” (SL 12.187) or a “self-end” (E § 252 add.), insofar as it is brought about through organizational differentiation and closure among its organs. Indeed, this general teleological characterization is echoed in Hegel’s discussion of the state when he claims that its self-maintenance is “an absolute unmoved end in itself” (PR § 258 – I will return to this claim below). In line with his general view of organisms, Hegel regards this “self-related process” as the self-determination of the political state. Indeed, he explicitly characterizes this process in precisely these terms when claims that “[i]t is only the inner self-determination of the concept  […] that is the absolute source of the division of powers” (PR § 272 – cf. §§ 275, 278, 279). Hegel thus identifies the “organism of the state” (PR § 267) or its political constitution with a process of organizational differentiation and closure among its constitutional organs such that each makes a distinct contribution to the self-maintenance of the state of which they are part. In a single expression: It is the self-determination of the political state. In his lecture transcriptions on the Philosophy of Right, Hegel also calls the process through which the political state determines itself and which he identifies as the “main activity within the state [die Hauptthätigkeit im Staate],” to which each of its constitutional organs makes a distinct contribution, “governing [Regieren]” (Griesheim: 698 f.). Thus, “governing” is tantamount to what we might call ‘political life’ or ‘the life of the state.’ In the addition to section 274 of the Philosophy of Right, Hegel infers from the processual understanding of the state’s “political constitution” that “[a] constitution is not just something manufactured” and that

234  Daniel James “[n]o constitution […] is just the creation of its subjects” (PR § 274 add.). Instead, “it is the work of centuries, it is the Idea, the consciousness of rationality so far as that consciousness is developed in a particular people” (PR § 274 add.). Along the same lines, he claims in the corresponding section of the Encyclopedia that “[i]t is history and the indwelling spirit […] by which constitutions have been and are made” (E § 540). To understand these claims, it is, in my view, helpful to distinguish two senses in which Hegel uses the term “constitution” in these passages. On the one hand, it stands for the body of constitutional norms that are in effect in any given state at any given time. On the other, it stands for the very organization or structure of the political state of which these norms and habits are part. Thus, although the latter, processual understanding of the political constitution does not preclude that, at any given time, any given body of constitutional norms will also, in some sense, be part of the political state, it does imply that its “political constitution” is distinct from it (see Lagerspetz 2004, 229 f.). In line with Hegel’s general view of organisms, I suggest we relate these two understandings of the constitution by identifying the former with (part of) the matter that is subject to the very process that Hegel identifies with the “political constitution” of the state, and the latter process itself with its corresponding form. This reading is, in my view, supported by Hegel’s claim that the process of “governing” underlying the division of powers “is spirit knowing and willing itself after passing through the forming process of education [der durch die Form der Bildung hindurchgegangene, sich wissende und wollende Geist]” (PR § 270). More clearly in the German original than in its English translation, he here identifies the process underlying the division of powers with a particular processual form, which we may, more aptly, call formation (that is my favored translation of ‘Bildung’). Therefore, I take Hegel to claim that it is the constitutional matter of a state which cannot be enacted by any candidate sovereign, be it the crown or the assembly of the estates (see Lagerspetz 2004, 229 f.). Instead, it can only be enacted by the process of “governing” itself, a process that is distinct from any particular governing act it may involve. Thus, it is this process that acts upon the body of constitutional norms in effect in any given state at any given time. We can now turn to the centerpiece of this chapter: Hegel’s organizational view of social functions. Although he rarely uses the term ‘function’ explicitly to describe the characteristic activities of social entities (see, for instance, PR § 302, remark), I take a particular account of social functions to be implicit in his discussion of the division of powers. For he understands the division of powers in terms of their organizational differentiation and closure such that their activity is constrained and thereby maintained by the political state as a whole. In other words, he understands the division of powers in terms of the self-determination

Hegel’s View of Social Functions  235 of the political state. Given my reconstruction of his general theory of organisms, this understanding implies that each of the constitutional organs bears a particular function for the political state, and is distinguished from the others by that function. But what is it for a part of the political state to bear a function? Recall that, in the previous section, I identified functions, on Hegel’s general account, with means-end relationships between the characteristic activity of each organ and the self-maintenance of the organism as a whole which obtain because of the self-determination of the organism. According to the definition Hegel’s account amounts to, a part of a living being has a function just in case it is subject to organizational closure in an organizationally differentiated self-maintaining system. In the here relevant case, the organizationally differentiated self-maintaining system in question is the political state, and the relevant parts are its constitutional organs. Given that the constitutional organs of the state meet the necessary and jointly sufficient conditions involved in my statement of the definition above, I take their functions to consist in the characteristic activities through which they each contribute to the self-maintenance of the political state as a whole. More specifically, given that the system to whose self-maintenance the constitutional organs each contribute – the political state – is social in nature, the same will be true of their functions. Thus, the constitutional organs each bear a particular social function: They are subject to organizational closure in an organizationally differentiated self-maintaining social system such that they contribute to the maintenance of its organization. So far, I have merely assumed that each of the constitutional organs makes a distinct contribution to the social process of “governing,” without further specifying what this process and the distinct contributions to it are. To do so, we must first determine what is distinctive of “political life” or the “life of the state,” as opposed to “logical life” more broadly. As noted above, Hegel calls the process through which the respective activity of the constitutional organs is constrained and thereby maintained, such that it makes a distinct contribution to the self-maintenance of the political state, a process of “education” or formation that results in “spirit knowing and willing itself” (PR § 270). Both where he discusses the division of powers and the nature of the state more broadly, he claims that it is the bearer of “the universal or substantial will” (PR § 258, remark – cf. §§ 257, 258) or, simply, “the will of the state” (PR § 261, add.). Crucially, he identifies “willing” as the characteristic activity of the state, which he takes to be “an absolute unmoved end in itself” (PR  §  258). This identification suggests that willing is the very end at which the activity of its organs is directed. Hegel makes this suggestion explicit in his discussion of the political constitution when he claims that the “very substantiality of the state is spirit knowing and willing itself

236  Daniel James after passing through the forming process of education” (PR § 270). Thus, he takes the process of formation through which the activity of the constitutional organs is constrained and thereby maintained to be directed at maintaining a particular both epistemic and volitional relation of the political state to itself. It is this relation that Hegel takes to be an end in itself. Addressing the nature of this relation would require an inquiry of its own – suffice it to say here that I think Hegel here identifies “governing” (or the political constitution of the state) with a process of collective will formation, determination, and implementation. In other words, the self-determination of the political state just is the process of forming, determining, and implementing its will. By virtue of the (both epistemic and volitional) self-relation resulting from this process, the constitutional organs constitute a corporate agent: the political state (see Steinberger 1988, 215–220; Lagerspetz 2004, 234–236). Indeed, in the corresponding section of the Encyclopedia, he explicitly ties the organizational differentiation and closure of the constitutional bodies as well as the self-maintenance of the political state as a whole to its agency: As a living spirit pure and simple, the state can only be an organized whole, differentiated into particular agencies, which, proceeding from the one concept (though not known as concept) of the rational will, continually produce it as their result. The constitution is this overall articulation of state-power [Staatsmacht]. (E § 539) As Hegel claims here, it is the same “concept […] of the rational will” from which the various powers, in their characteristic activity, are “proceeding” on the one hand and which they thereby “continually produce as their result,” on the other. In this claim, he again echoes his characterization of a self-determining system in terms of the instrumental relation in which it stands to itself: It is a “purpose unto itself [Selbstzweck]” (SL 12.187) or a “self-end, […] End and Means at the same time”, insofar as it is “its own product” (E § 252 add.). Recall that Hegel identified the “End” of the organism’s activity with “an ideal determination which is already existent beforehand,” to wit, the end-state that is maintained through constraining and thereby maintaining the activity of each of its organs. Through this activity, the organism contributes to maintaining the conditions of its very own existence. In other words, because of this activity, the activity of its organs is directed at maintaining the organism of which they are each part. In case of the political state, the “ideal determination” in question is “the one concept […] of the rational will”. Thus, when Hegel claims that the “particular agencies” that are each part of the political state are, in their characteristic activity, “proceeding” from this “concept,” I take him to mean that their activity is directed at maintaining the political state as an individual of a particular

Hegel’s View of Social Functions  237 kind, namely, as the bearer of a “rational will.” As such, the “rational will” both effects the organizational differentiation and closure of the constitutional organs and is effected by it. This is, in my view, the distinctively volitional manner in which the political state characteristically determines itself. To use a vitalistic metaphor: it is the will of the state that animates its organization. When Hegel, against the background of this account of the political state’s self-determination, identifies its “constitution” with “this overall articulation of state-power [Staatsmacht],” I take him to mean the particular kind of power that is the agency of the political state. The maintenance of this “state-power” or agency of the political state is the end to which its constitutional organs, through their respectively characteristic activity, each serve as a means. In the corresponding section of the Encyclopedia, he makes explicit that the political state’s self-determination involves its constitutional organs each contributing to the formation, determination, and implementation of its will in a distinct way: [The constitution] involves the determinations of the way in which the rational will […] firstly, comes to consciousness and understanding of itself and is found, and is, secondly, posited in actuality, through the agency of the government and its particular branches, and maintained in actuality, and also protected against the contingent subjectivity both of these governmental departments and of individuals. (E § 539) In this passage, Hegel distinguishes three distinct contributions to the process of forming and implementing the will of the state, which we can roughly map onto three stages of this process. In the first stage of formation, the legislative power weighs the state’s aims or policies (with a view to the interests of certain relevant social groups within its citizenry, which Hegel calls the estates [Stände]). In the second stage of determination, the crown determines these aims or policies, such that their subsequent implementation is “protected against the contingent subjectivity both of these governmental departments and of individuals.” In the third stage of implementation, the executive power applies the thereby determined aims and policies of the state to specific situations and cases. In the closing section of the introduction to his theory of the political constitution, Hegel identifies these three distinct contributions with what I have called the respective function of the constitutional organs. These are, first, the legislative branch’s “power to determine and establish the universal”; second, the executive branch’s “subsumption of individual cases and the spheres of particularity under the universal”; and, third, the crown’s “subjectivity, as the will with the power of ultimate decision—the crown” (PR § 273).

238  Daniel James I take the distinct contribution each of the constitutional organs makes to the process of forming, determining, and implementing the will of the state through its characteristic activity to constitute its respective function because this activity is constrained and thereby maintained by the political state of which they are each part. Indeed, it is the idea that each of the constitutional organs is the bearer of a function for the political state that underpins Hegel’s understanding of the division of powers. The notion of organizational closure, that is, of the mutual dependence among their parts, is particularly relevant to how he understands the division of powers. By drawing on this notion, he aims to model what he takes to be the nonaggregative, nonmechanistic causal interplay among the constitutional organs. It is this model of the political state that involves the functional explanation of its parts’ respective activity, an explanation he spells out in terms of the distinct contribution they each make to the self-maintenance of the political state of which they are part. On Hegel’s account, it is the respective function of each constitutional organ that determines its nature. Indeed, bearing particular functions for the political state is part of why they are its constitutional organs. In other words: Part of what it is for a part of the political state to be its organ is for it to be subservient to the self-maintenance of the political state, meaning that its activity is constrained and thereby maintained by the political state as a whole. As such, the constitutional organs are, in accordance with Hegel’s general theory of organisms, essentially dependent on the political state as a whole. I think that Hegel has this essential dependence of the constitutional organs on the political state as a whole in mind in his claim cited above, according to which “each of these powers is in itself the totality” (PR § 272). For, in addition to the mutual dependence among the constitutional organs of the political state mentioned above, he explains (“because” [dadurch, dass]) their nature by appealing the fact that they “remain utterly within its ideality and constitute nothing but a single individual whole” (PR § 272). Roughly, I take this to mean that the constitutional organs are each the kind of entities that they are – doing what they characteristically do – because their activity is constrained and thereby maintained by the political state of which they are each part. Thus, they remain “within its [the political state’s] ideality” insofar as they essentially depend on it. In this sense, Hegel also refers to them as mere “moments” (PR § 272) of the political state. This idea is reflected in Hegel’s introduction of the powers of the political state. As he puts it, through the process in which the state “divides itself into the distinct spheres of its activity,” they are “actually fixed determinations of the state, i.e. its powers” (PR § 270). Later, he restates this claim when he identifies the political constitution with “a process whereby it differentiates its moments within itself and develops them to self-subsistence” (PR § 271). As Hegel suggests in both statements, it is because of this

Hegel’s View of Social Functions  239 constraint and maintenance brought about the political state, such that they each make a distinct contribution to its self-maintenance, that they constitute its constitutional organs. Hegel thus makes use of the organismic analogy to account for the nature of the constitutional bodies themselves as well as their relations to each other. Crucial to this analogy is the notion of self-determination: a peculiar kind of circular causal regime, in which a system continually generates and maintains the conditions of its own ongoing existence through generating and maintaining distinct structures whose activity is mutually dependent, such that they each contribute to the maintenance of the system of which they are each part. Self-determination is, therefore, the general structure common to certain biological and social systems, a structure Hegel aims to make salient through his use of the organismic analogy. He draws on the notion of self-determination to conceptualize what he takes to be the characteristic structure of a particular kind of social system: the constitution of the political state. To understand the account of social structure to which the organismic analogy commits him, we must consider how the metaphysical implications of his general theory of organisms bear on this theory of the political state. As I will show in what follows, they amount to an organizational view of social structure as a particular kind of process the parts of which are distinguished in terms of different functions. Thus conceptualizing the structure of the state is, in my view, the point of the use he makes of the organismic analogy.

So, What’s the Point? Recall my brief discussion of the metaphysical implications of Hegel’s account of organisms and the functions of their organs in terms of self-determination for the nature of both organisms as a whole and their organs in the first section of this chapter. As I claimed above, the selfdetermination of organisms bears on their very identity: part of what it is to be an organism is to be a hierarchy of mutually constraining and thereby maintaining processes. The same is also true of the state. This implication is particularly relevant for my purposes because it directly concerns Hegel’s account of the social structure of the political state. For it suggests that the structure of the state is a particular process the parts of which are distinguished in terms of different functions. To cite another analogy, its structure is more akin to that of a piece of music than a piece of architecture. To see that Hegel, in accordance with this general theory of organisms, identifies the political state with a process, rather than a thing, recall the role that the notion of form plays in accounting for the nature of the political state. As I argued in the previous section, Hegel implicitly identifies the body of constitutional norms that are in effect as well as

240  Daniel James the corresponding shared habits that bring them into effect in any given state at any given time with (part of) the matter that is subject to the process of forming, determining, and implementing the will of the state. Correspondingly, he identifies this process itself with its form (to more adequately reflect his characterization of this form in processual terms, I spoke of ‘formation’). Above, I took Hegel’s talk of ‘form’ to suggest that his general theory of organisms relies on some version of hylomorphism, that is, the view that to account for the nature of an object one must appeal to both their material and formal components. This view is also reflected in Hegel’s account of the political state. For he takes the process of forming, determining and implementing the will of the state – or, as he puts it: “the forming process of education” (PR § 270) – to explain why and how the political state possesses its characteristic features. These features include, above all, having particular constitutional organs which each have the capacity for specific activities. Indeed, the view that we must account for the nature of the political state in terms of a particular processual form – its constitution – is reflected in his identification of “the organism of the state” with a specific “process,” namely, “the development of the Idea to its differences and their objective actuality” (PR § 269). Similarly, Hegel’s understanding of organisms as hierarchies of mutually constraining and thereby maintaining processes is reflected in his identification of the political constitution with “the organization of the state and the self-related process of its organic life” (PR § 271). These passages suggest that, as an organism, the political state is, by its very nature, a social system that is structured in a particular manner, namely as a process of forming, determining, and implementing the will of the state. Thus, on Hegel’s account, these three stages, to which each of the constitutional organs makes a distinct contribution, are each part of the social structure or the organization of the political state. I take this account of social structure to be a crucial part of the point of the use Hegel makes of the organismic analogy. For it suggests an account of the structure of the state as a particular kind of process, the parts of which are distinguished in terms of different functions. As I showed in the previous section, these functions are the formation, determination, and implementation of the will of the state. In a nutshell: through the organismic analogy, Hegel identifies the structure of the political state with its organization as a self-determining system.

Conclusion Let’s return to the theme with which I opened this chapter: the charm and trickiness of analogies. The charm of the organismic analogy for Hegel’s purposes stems from the use he can make of it to account for the nature and inner workings of a phenomenon which, at a time in which

Hegel’s View of Social Functions  241 the social sciences were yet to emerge as an autonomous domain of inquiry, was not well understood: the state. What, in my view, motivates him to draw on the putatively better-understood domain of biology is that both the source and the target phenomena – living beings and states – appear to share certain characteristics, such as a high degree of cooperation among some of their parts and maintenance of their identity through continuous change. The notion of self-determination is, in my view, meant to capture these shared features. It involves representing biological and social organisms more abstractly as systems that maintain themselves through maintaining the activity of their parts. The result of this abstraction – in which we include only those factors required to give rise to the phenomenon in question: self-determination – is the generic theoretical representation of “logical life” we encounter in the Science of Logic.12 Thus representing both biological and social organisms as systems of the same kind in terms of the generic theoretical representation of “logical life” suggests certain hypotheses about the features they are thus taken to have in common. These hypotheses guide Hegel’s inquiry into whether certain features of the source phenomenon – namely, all those that give rise to self-determination – can also be encountered in the target phenomenon. As such, Hegel’s generic theoretical representation of “logical life” acts as a model of “political life,” which highlights certain features of the target phenomenon while disregarding others. In case of the organismic analogy, these features include organizational differentiation and closure, as well as the self-maintenance they facilitate. The central task Hegel thereby sets himself in his account of the political state is to identify both organizational differentiation and closure in the political state. Accordingly, much of his discussion of the political constitution does, indeed, aim to fulfil this task by identifying various social processes that contribute to the self-maintenance of the political state. Where does this leave us with the organismic analogy? The kind of circular causal regime Hegel models using this analogy does not involve the social selection mechanism Elster misses in social-functional explanation. Therefore, Hegel’s organizational view of social functions is not vulnerable to the central objections Elster raises against the evolutionary view. But what about Elster’s complaint that the organismic analogy gives rise to “pseudo-explanations” and “pseudo-problems”? Whether and to which degree this is true depends on whether social systems do, in fact, display organizational differentiation and closure – and this is, in my view, an open, empirical question. Indeed, Hegel himself seems to be very much aware of this constraint on the organismic analogy and, therefore, on functional explanation in the social domain since he acknowledges that there are token states which do not display the selfdetermination required for some of their parts to be function bearers and, thus, support functional explanation. This suggests that Hegel’s

242  Daniel James account of the political state in terms of self-determination and the organismic analogy it rests upon is not merely an abstraction but also an idealization: a model of the state that disregards such inconvenient features of token states to show how the causal interaction among their constitutional bodies can give rise to state agency.13 Thus, this “purpose unto itself” (SL 12.187) or “self-end, […]” (E § 252 add.) towards which the activity of the political state is directed can be actualized to a greater or lesser degree. What lesson can we draw from my discussion of the use Hegel makes of the organismic analogy for our contemporary concerns? On the one hand, I think Elster’s dismissal of the organismic analogy rests on a misconception of what I take to be its point; on the other, I think his preference for the evolutionary analogy rest on an overly narrow understanding of functional explanation in general. For a look into the backstory of social functionalism yields an organizational understanding of social functions that, particularly given the recent revival of the kind of biological theory that informed it, strikes me as an underappreciated contender in the contemporary debate. However, whether such an understanding, indeed, fares better than its alternatives, remains yet to be shown.

Acknowledgments I would like to thank Dina Emundts, Till Hoeppner, Rolf-Peter Horstmann, Rebekka Hufendiek, Franz Knappik, Karen Koch, Thomas Meyer, Al Prescott-Couch, and Raphael van Riel for their helpful discussions and comments. I would also like to express my gratitude to the DFG for generous funding of the Network “Social Functions in Philosophy.”

Notes



Hegel’s View of Social Functions  243 As these authors point out, this notion traces back to Kant’s account of the ‘inner purposiveness’ characteristic of ‘natural ends’.





12 I am here drawing on ideas developed in Nersessian (2008).

References Austin, Christopher J. 2018: “A Biologically Informed Hylomorphism.” In Neo-Aristotelian Perspectives on Contemporary Science, edited by William M. R. Simpson, Robert C. Koons and Nicholas J. Teh, 185–210. New York: Routledge. Cartwright, Nancy. 1989. “Capacities and Abstractions.” In Scientific Explanation. Minnesota Studies in the Philosophy of Science, vol. 13, edited by Philip Kitcher and Wesley Salmon, 349–356. Minneapolis: University of Minnesota Press. Cheung, Tobias. 2010. “What Is an ‘Organism’? On the Occurrence of a New Term and Its Conceptual Transformations 1680–1850.” History and Philosophy of the Life Sciences 32, no. 2/3: 155–194. Collier, John D. 2004. “Self-Organization, Individuation and Identity.” Revue internationale de philosophie 228, no. 2: 151–172.

244  Daniel James DeVries, Willem A. 1991. “The Dialectic of Teleology.” Philosophical Topics 19, no. 2: 51–70. Dupré, John A., and Nicholson, Daniel J. 2018. “A Manifesto for a Processual Philosophy of Biology.” In Everything Flows: Towards a Processual Philosophy of Biology, edited by Daniel J. Nicholson and John A. Dupré, 3–45. Oxford: Oxford University Press. Elster, Jon. 1989. Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press. Govier, Trudy. 1989. “Analogies and Missing Premises.” Informal Logic 11, no. 3: 141–152. Hegel, G. W. F. 2004. Hegel’s Philosophy of Nature: Encyclopaedia of the Philosophical Sciences (1830), Part II. Oxford: Oxford University Press. ———. 2008. Outlines of the Philosophy of Right. Oxford World’s Classics. Translated by T. M. Knox. Revised, edited, and introduced by Stephen Houlgate. Oxford: Oxford University Press. ——— 2010a. The Science of Logic. Translated and edited by G. di Giovanni. Cambridge: Cambridge University Press. ———. 2010b. Encyclopedia of the Philosophical Sciences in Basic Outline. Part I: Science of Logic. Translated and edited by K. Brinkmann and D. Dahlstrom. Cambridge: Cambridge University Press. Kant, Immanuel. 2002. Critique of the Power of Judgment. The Cambridge Edition of the Works of Immanuel Kant. Translated by Paul Guyer and Eric Matthews, edited by Paul Guyer. Cambridge: Cambridge University Press. Kreines, James. 2005, “The Inexplicability of Kant’s Naturzweck: Kant on Teleology, Explanation and Biology.” Archiv für Geschichte der Philosophie 87, no. 3: 270–311. ———. 2015. Reason in the World: Hegel’s Metaphysics and Its Philosophical Appeal. Oxford: Oxford University Press. Lagerspetz, Eerik. 2004. “Hegel and Hobbes on Institutions and Collective Actions.” Ratio Juris 17, no. 2: 227–240. Lenoir, Timothy. 1982. The Strategy of Life: Teleology and Mechanics in Nineteenth-Century German Biology. Chicago, IL: University of Chicago Press. McLaughlin, Peter. 2000. What Functions Explain: Functional Explanation and Self-Reproducing Systems. Cambridge: Cambridge University Press. Montévil, Maël, and Mossio, Matteo. 2015. “Biological Organization as Closure of Constraints.” Journal of Theoretical Biology 372: 179–191. Moreno, Alvaro, and Mossio, Matteo. 2015. Biological Autonomy. A Philosophical and Theoretical Enquiry. Dordrecht: Springer. Mossio, Matteo, Saborido, Cristian, and Moreno, Alvaro. 2009. “An Organizational Account of Biological Functions.” British Journal for the Philosophy of Science 60, no. 4: 813–841. Mossio, Matteo, and Bich, Leonardo. 2017. “What Makes Biological Organization Teleological?” Synthese 194, no. 4: 1089–1114. Nersessian, Nancy. 2008. Creating Scientific Concepts. Cambridge, MA: The MIT Press. Oderberg, David S. 2011. “Essence and Properties.” Erkenntnis 75, no.1: 85–111.

Hegel’s View of Social Functions  245 ———. 2018. “The Great Unifier: Form and the Unity of the Organism.” In Neo-Aristotelian Perspectives on Contemporary Science, edited by William M.R. Simpson, Robert C. Koons, and Nicholas J. Teh, 211–234. New York: Routledge. Rosen, Robert. 1991. Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. New York: Columbia University Press. Sedgwick, Sally. 2001. “The State as Organism: The Metaphysical Basis of Hegel’s Philosophy of Right.” Southern Journal of Philosophy 39, no. 1: 171–188. Steinberger, Peter J. 1988. Logic and Politics: Hegel’s Philosophy of Right. New Haven, CT: Yale University Press. Stern, Robert. 1990. Hegel, Kant and the Structure of the Object. New York: Routledge. Varela, Francisco J. 1979. Principles of Biological Autonomy. The North Holland Series in General Systems Research, vol. 2. New York: Elsevier North Holland. Wolff, Michael. 1985. “Hegels staatstheoretischer Organizismus.” HegelStudien 19: 147–177. Wright, Larry. 1973. “Functions.” Philosophical Review 82, no. 2: 139–168.

Notes on Contributors

Amrei Bahr is a Postdoc at Heinrich Heine University Düsseldorf. Her main research interests lie in the philosophy of artifacts, with a special focus on artifact functionality and artifact copies, and in applied ethics, particularly the ethics of copying and the ethics of waste disposal. Rebekka Hufendiek is a Postdoc at the University of Basel. Her research interests lie in philosophy of mind and philosophy of science, particularly in empirical and ideological dimensions of research on cognitive and behavioral features. Her book Embodied Emotions: A Naturalist Approach to a Normative Phenomenon (Routledge 2015) provides a noncognitivist theory of emotions. Daniel James is a Postdoc at Heinrich Heine University Düsseldorf. His historical research concerns the intersection of Hegel’s metaphysics with his political philosophy. With a view to contemporary debates in social philosophy, he is interested in the concept of social power and its fruitfulness for social-scientific inquiry, as well as in social dispositions and their connection to social-structural explanation. Harold Kincaid is a Professor of Economics at the University of Cape Town, South Africa. He is the author of several books and many articles and book chapters in the philosophy of science, social science, and behavioral science. Heiner Koch  is a Researcher at the University of Duisburg-Essen. His research interests lie in social philosophy with a focus on philosophy of social science, social metaphysics, philosophy of technology, applied ethics, theories of power and domination, and Marxism. At the moment, his research focuses on machine learning. Daniel Little is a Professor of Philosophy and Sociology at the University of Michigan-Dearborn and UM-Ann Arbor. His research interests include the philosophy of the social sciences and organizational sociology. Recent publications include Disaggregating Historical Explanation (Routledge Handbook of Philosophy of Mechanisms, 2017) and New Directions in the Philosophy of Social Science (Rowman & Littlefield, 2016).

248  Notes on Contributors Mari Mikkola is an Associate Professor of Philosophy at the University of Oxford and Tutorial Fellow of Somerville College. She is the author of two books (The Wrong of Injustice: Dehumanization and Its Role in Feminist Philosophy and Pornography: A Philosophical Introduction, both with OUP) and of several articles on feminist philosophy, social ontology, and pornography. Andreas Müller is an Assistant Professor at the University of Bern. Most of his research concerns questions in moral philosophy. In his book Constructing Practical Reasons (Oxford University Press, 2020), he presents a constructivist theory of practical reasons that builds on a non-representationalist account of normative judgments and their truth. Matthieu Queloz is a Junior Research Fellow at Wolfson College and a Member of the Faculty of Philosophy at Oxford University. His book The Practical Origins of Ideas: Genealogy as Conceptual ReverseEngineering (under contract with OUP) examines a tradition of using genealogy to reveal the functions of concepts. Raphael van Riel holds a position as a Senior Lecturer at the University of Duisburg-Essen, where he directs a research group which focuses on theories of explanation. In his book The Concept of Reduction (Springer 2014), he offers a novel explication of reduction claims in the philosophy of mind and the philosophy of science. Charlotte Witt  is a Professor of Philosophy and Humanities at the University of New Hampshire. She is the author of The Metaphysics of Gender (OUP 2011) and has published extensively on Aristotle, Ancient Philosophy, Social Ontology, and Feminist Philosophy.

Name Index

Note: Page numbers followed by “n” denote endnotes. action 6, 14, 29, 30, 33, 35, 36, 38, 49, 124–128, 132, 140, 145, 149, 193, 194 adaptation 3, 4, 12, 31, 33, 41, 42, 95 adaptation, functional 29, 30, 42, 53 Afunction 123, 125, 128, 129, 131 agent 18, 36–38, 42, 101, 122, 124, 127, 129, 130, 144, 160, 171, 178, 184, 194, 203, 206, 225, 231, 236 analogy (between biology and sociology) 3–4, 29, 89n14, 219–220, 229, 242n2 analogy, organismic 4, 6, 48, 49, 54, 219–220, 228, 230, 239–242 analysis, functional 5, 7–8, 28, 48, 51, 53–55, 58, 60, 61, 67n17, 71, 75, 76, 78, 79, 81, 169–170, 179 anomie 9, 46, 48–49 anthropology 3, 7, 71, 95, 100, 114 argument, genealogical 142–145 arrangement 28–29, 30, 31, 33–35, 37, 39, 40–41, 130, 215, 222–223 artifact function see artifact, functionality artifact functionality 1, 32, 81, 130, 133n9, 146, 183–184, 190–191, 195–196 artifact functions, intentionalist accounts of 183–190, 193–196 benefit 2, 7, 8, 74, 187, 188, 190, 192, 204, 205 capacity 4, 58, 59, 99, 104, 107, 109, 129, 137, 150, 201, 203–205, 240 capacity for normative guidance 137, 150–151 capitalism 47, 63, 70–71 causal-role theory of functions 8, 94, 97, 99–100, 107, 117, 118n3, 186, 205

cause 21–24, 55, 58, 73, 83, 88n3, 97, 206 change 145–146, 149–152, 169, 170, 173, 175, 241; organizational 6, 31, 33–35, 37–43; social 5, 20, 45–47, 49, 51, 53–54, 64–65 claim, causal 21–23, 25, 94, 97, 98, 208 closure, organizational 10, 221, 223–226, 230–238, 241 cohesion, social 84, 85, 96, 191 commonality problems (for feminists) 161, 163, 164, 169, 176, 180 competition 52–53, 198n19 condition, human 94, 105, 111, 112 consequence, beneficial 29, 70, 89n8 constitution, political 230– 238, 240–241 constraint, mutual 221, 226, 232 constructivism, social (about gender) 161, 162, 179 context, social 76, 95, 99, 105, 107, 108, 129, 131 cooperation 7, 11, 20, 22, 104–107, 128–130, 137, 212, 220 Cummins functions 73, 80–81, 86, 90n15 debunking project 138, 140 dependence 10, 220, 224, 228, 231, 232, 238 dependency, causal 80, 82–83, 85–86 description, functional see analysis, functional design see: designer intentions designer intentions 1–2, 29–32, 34–37, 56, 60–61, 63, 87, 126, 147, 165–168, 184–189, 192, 203, 207 differentiation, organizational 9–10, 221, 223–225, 229–234, 236, 237, 241

250  Name Index directedness/Directional 30, 31, 53, 65, 72 disease 49–51, 54, 56 disposition see capacity division of powers 231–235, 238 domain, social 183, 184, 186, 190, 193, 195 dysfunction 9, 12, 32–34, 36, 46–67, 80, 102, 106, 116, 208 effect, beneficial 2, 3, 10, 20, 58, 70, 75, 82, 85, 97, 101, 107, 222, 223, 225; dysfunctional 9, 34, 45–48, 51, 54–59, 62–65; functional 1, 2, 7, 9, 14, 21–22, 30, 35, 63–65, 70, 71, 78, 79, 83, 86, 210; negative (see effect, dysfunctional); selected 14, 24, 58; unintended 48, 54, 60, 191–196 emancipation 46, 65, 151 emotion 93–96, 98, 99, 103–113, 117, 204 end, natural 221–222 entities, social 3, 24, 26, 29, 30, 45, 47, 50, 63, 66n10, 190, 196n2, 234 ergon 125–127 essence 94, 164, 175, 225 essence, historical 13, 14, 161, 162, 164, 165, 174, 177 essentialism, historical 162, 164, 170, 171, 173, 175, 176, 179 etiological theory of functions 4, 7–9, 12, 13, 48, 55–56, 61, 63, 79–81, 86–87, 94, 96–98, 100, 103–105, 108, 112, 117, 122, 136–137, 162, 164–164, 171, 208 evidence 10, 18–20, 22, 23, 25, 28, 94, 98–103, 105–108, 110–114, 117 evolution 7, 20, 24, 33, 65, 96, 101, 104, 105, 107, 111, 113, 116, 136, 138, 154, 169, 201; biological 18, 29, 51, 53, 75, 95, 97, 115, 141, 207; of morality 12, 136, 140, 141; social 4–6, 23, 29, 46–48, 52, 54 explanation, causal 10, 25, 82, 83, 86, 94, 96, 100–102, 110 explanation, functional 3, 10, 18–26, 46, 50, 64, 65, 70–87, 93–117, 138, 155, 160, 162, 171, 180, 190, 192, 193, 196, 207, 209, 211, 213, 219, 241, 242 explanation, genealogical 6, 86, 101, 136, 138, 142, 200–215 explanation, mechanistic 8, 9, 12, 85 explanation of social change 5, 20, 45–47, 49, 51, 53–54, 64–65 explanations as answers to whyquestions 102, 171

feedback loop see feedback mechanism feedback mechanism 31, 34, 35, 38, 40, 41, 43, 70–75, 78, 80, 81, 84–87, 106, 136, 138 feedback process see feedback mechanism feminism 159, 160, 162, 174, 175, 177, 180 form 227–228, 234, 239, 240 formation 234–237, 240 functionalism: social 3–6, 28, 29, 219, 220, 242; structural 48, 129 functionality 6, 33, 53, 58, 74, 95, 103, 104, 140, 141, 146–149, 151–153, 155, 183, 186, 195, 200, 203–210, 212 function ascriptions 1, 73, 87, 185, 188, 192, 195 function bearer 3, 6, 7, 46, 58, 59, 62, 70, 72, 78–84, 86, 97, 98, 107, 109, 136, 137, 227, 241 function, biological 93, 97, 104, 137, 140 function, manifest and latent 29, 64, 75, 76, 80, 83, 183 function of morality 135–140, 145, 146, 154, 155n1 function of shame and pride 93–111, 113–117 functions, causal-role theory of see causal-role theory of functions functions, etiological theory of see etiological theory of functions function, social 1–6, 13, 14, 28, 29, 93, 109, 183, 184, 186, 188–196, 200, 202–205, 212, 220, 229, 234, 235, 241, 242 functions, organizational theory of see functions, organizational view of functions, organizational view of 3, 4, 9, 219, 227, 229, 234, 239, 241 functions, proper see functions, etiological function statements 186–188, 194, 196, 197n15 function, teleological 165–168, 171, 173 function, unintended social 183, 184, 186–191, 195 game theory, game-theoretical 25, 26, 101–102 gender 13–14, 159–176, 178–180 gender as a natural kind with a historical essence 13, 161–162

Name Index  251 gender kinds 13, 160–161, 164–168, 170–171, 173–176, 179 gender socialization 13, 108, 167, 168, 170–172, 175–176 gender system 13–14, 162, 166–171, 173, 176 genealogy, pragmatic 6, 200–202, 205–211, 213–215 group selection 24 group, social 103, 104 group traits 24 hierarchy, social 93, 104–109, 111–114, 167, 171 holism see holistic holistic 19, 93, 94, 96, 101, 111, 117 homeostasis see homeostatic homeostatic 164, 166–171, 173 homeostatic property cluster theory 164 hylomorphism 227, 240 idealization 202, 242 ideology, Ideological 94, 103, 111, 113, 114, 117, 150 implications, metaphysical 8, 13, 220, 227, 229, 239 institution, social 5, 41, 78, 87 intention 1, 2, 5, 6, 14, 18, 31, 35, 37, 38, 52–56, 58, 60, 61, 63, 71, 73–75, 80, 87, 101, 108, 116, 128, 183–187, 189–196, 213, 214 interests 2, 18, 30, 34–38, 40–42, 100, 101, 115, 191–193 interpersonal approach to emotions 107–110, 118n4 intimacy, nomadic 188, 189, 194, 195 kinds 13, 103, 108, 160, 161, 162, 164–179 kinds, natural 13, 14, 50, 123, 161, 162, 164, 171, 174–177, 179 kinds, social 13, 14, 50, 123, 132n3, 159–161, 163, 178 kind, teleofunctional 173, 174 laws of social functioning 75–76, 84, 85 learning, organizational 28 living being 3, 4, 49, 222–226, 228, 232, 235, 241 malfunction 47–48, 54, 56–57, 64, 66n7, 185–186 materialism, historical 46, 51, 53–54 matter 227–228, 234, 240 mechanism 95, 97, 100, 105, 106

mechanism, causal 4–5, 42, 94, 97, 111, 114, 117 mechanism, cultural see mechanism, social mechanism of selection see selection mechanism mechanism, selectionist see selection mechanism mechanism, social 30, 65, 167 Missing-Mechanism Argument (MMA) 64, 70–74, 77–78, 84, 86–87 model 24, 164, 241 model, dynamic 6, 200, 202, 205, 207, 212, 214 modeling, interpretive 10, 94, 100–103, 106, 110–111, 115–117 modeling, rational 94, 101–102, 115–116 naturalism 19, 139, 153–154 naturalism, reductive (about morality) 140–143 naturalistic challenge (to morality) 139–143, 146, 154 naturalist realism (about morality) 142–143, 155n5 nature, human 103, 111–113 needs 2, 80, 83, 107, 160, 201, 203, 208–209, 214 needs, social 21, 28, 202 normativity of social roles / social role normativity 122–125, 129–131 normativity, social 123–124, 131, 132n4 norms 20, 22, 49, 104, 109, 128, 130, 153, 234, 239–240 norms, social 4, 12, 96, 124–125, 129, 137 objectivism, social (about gender) 163, 171–172, 176 ontological assumptions see metaphysical implications optimality arguments 98–99, 105–106, 207 organism, biological 220, 242n3 organism of the state see organism, social organism, social 4, 48–50, 229–230, 233, 240–241 organization 6, 9–10, 21, 30–42, 50, 58–59, 61, 63, 220–224, 233, 240 organization, hierarchical 93, 111, 113–114 organs, constitutional 230–233, 235–240, 242

252  Name Index pluralism, alethic 148–149 position, social 122–125, 163 practice, social 11, 14, 20–22, 76, 83, 85–86, 102, 108–110, 113, 160, 194, 196 pride 93–97, 99, 103–107, 111–116 process, causal 18, 52, 98, 101, 106, 145 processes, ontogenetic 13, 161, 165, 168–169 process, evolutionary 30, 96 process, historical 13, 98 process, ontogenetic 13, 161, 165, 167–169, 172, 174–175, 181n6 process, social 2, 32, 54, 229, 235–237, 239, 241 progress, ethical 146, 150–152, 155 reasoning, practical 144–145, 155 relation, causal 21–23, 25, 97, 99 relation, instrumental 206–207, 209, 222, 225, 228, 236 replication 164–165 representation, political 173–176, 178 representation problem for (feminists) 161, 163, 174, 176, 179 reproduction 13, 55, 59, 63, 137, 163–165, 169, 171 resilience 6, 77, 200, 209–210 role, causal 7–9, 94, 97, 99, 107, 117, 118n3, 136, 165, 186, 205 role, social 11, 21, 118n4, 122–125, 127, 130–131, 132n2 science 26, 100, 129, 139 science, social 2, 5, 9–10, 18–20, 24–25, 48, 54, 77, 101–102, 219–220 selection, cultural 104, 136 selection, historical 97, 210 selection history 55, 99 selection mechanism 15n4, 73, 75, 77, 81, 83, 85, 87, 98, 105 selection mechanism, social 4, 105, 241 selection, natural 7, 29, 52, 75, 97, 101, 136–138, 219–220 self-constraint 226, 229, 232 self-determination 10, 221–231, 233, 236–237, 239, 241 self-end 225, 233, 236, 242 self-maintenance 10, 62, 221, 223, 225–227, 229–230, 233, 236

service function 183, 187–188, 190 sex vs gender 159–160, 166, 168, 170 shame 93–99, 103–115 shame and guilt cultures 109, 113–114 skepticism (about gender) / ‘gender skeptic’ views 159, 178–179 socialization see gender socialization society 4, 9, 18, 21, 24, 29–30, 41, 45–46, 49–50, 52–53, 63–65, 77, 82, 87, 109, 124, 130, 201, 209 sociology 3, 29, 34, 41, 70, 227 stability 2–5, 37–38, 46, 49, 54, 62, 70, 130, 165, 169, 209 standard, normative 11–12, 54, 56 standards of excellence 127–129 state of nature 200–203, 207, 213, 215, 215n2 state, political 228–242 status functions 12, 55, 104 status, social 93, 103–104, 106, 189 structure, social 24, 37, 50, 64, 70, 74–75, 103, 107, 116, 125, 163, 176, 220, 229, 239–240 survival 24, 29, 55, 58–59, 63, 81–82, 86, 103, 107, 137, 208 system, biological 8, 221, 228–229, 239 system, social 5, 8, 20, 50–51, 54, 59, 66n7, 79, 83–85, 109, 221, 229, 235, 239–241 teleofunction 165–166, 168, 170, 173–174 teleology 4, 10, 18, 20–21, 80, 215n6, 227, 243n7 traits, biological 3, 62, 81, 86 transition, progressive 145–146, 151–152 truths, moral 136, 138–140, 142, 145–146, 149, 152–155 type, social 49–50, 64 varieties of social functionalism: persistence functionalism, etiological functionalism, composed object functionalism, constituent functionalism, identity functionalism 79–84 virtue 127–128, 202, 204, 206–211 voluntarism vs ascriptivism (about social normativity) 124–125, 131

Subject Index

Note: Page numbers followed by “n” denote endnotes. Achinstein, Peter 183, 186–188, 190, 197n11 Alcoff, Linda 160, 162 Aristotle 11, 123, 125–131, 132n7, 133n8, n9, n12 Bach, Theodore 13–14, 160–166, 168–179, 180n2, 181n5, n6 Barney, Rachel 126, 127, 132n7, 133n8, n12 Benedict, Ruth 96, 114 Bourdieu, Pierre 2 Boyd, Richard 164 Broadie, Sarah 127 Cohen, Gerald A. (Jerry) 29, 45–46, 51–53, 64, 70, 73, 88n3 Craig, Edward 101, 200, 202, 206, 212–214 Craver, Carl 7–9, 72, 97, 100 Cummins, Robert 7, 9, 21, 55–56, 58–60, 66n11, 67n13, 73, 80–83, 86, 90n15, 97, 136, 205, n12 Darwin, Charles 5, 14n2, 19, 52, 117n2, 119–220, 136, 138–139 Defoe, Daniel 116 Dummett, Michael 211 Durkheim, Émile 4, 9, 18–19, 46, 48–49, 66n3, 79–80, 82, 87 Elster, Jon 5, 8, 10, 29, 32, 41, 64, 70–71, 73–79, 82–83, 87, 88n1, 98, 186, 190–193, 197n16, 198n17, 219–220, 241–242, n3, n4, n6, n7, n8, n18 Engels, Friedrich 52–53 Fessler, Daniel 95–96, 104–107, 116 Fligstein, Neil 37–38, 42 Frevert, Ute 109–110

Fricker, Miranda 6, 200, 202, 206 Frye, Marilyn 179–180 Geertz, Clifford 102, 114 Geser, Hans 188 Griffiths, Paul 8, 13, 81, 98 Guyau, Jean-Marie 48–49 Haslanger, Sally 14, 162–164, 170–172, 174, 176, 180n2, 193–194 Hegel, G.W.F 4, 87, 119–242, 243n5, n6, n9, n11 Hempel, Carl G. 7, 72, 74, 78, 84 Hobbes, Thomas 6, 201, 215n2 Hume, David 6, 152, 200, 202, 206–207 Jackson, Frank 71, 77, 89n9 Joyce, Richard 135, 137–138 Kant, Immanuel 221–222, 224, 243n4 Kilday, Anne-Marie 116 Kitcher, Philip 12, 135, 137–139, 144–155, 156n8, 201, n9, n12, n13, n15, n18 Knight, Jack 42 Krohs, Ulrich 80, 130 Kusch, Martin 202, 212 MacIntyre, Alasdair 11, 128–129 Malinowski, Bronislaw 2, 5, 29, 80, 83, 86–87 Mallon, Ron 122, 132n2, 164, 178–179, n3 McAdam, Doug 37–38, 42 McKenna, Robin 202 McLaughlin, Peter 8–9, 66n9, 72, 73, 90n15, 132n7, 190–192, 198n18, 225 Merton Robert K. 5, 9, 28–29, 45–46, 48–50, 53–54, 56, 58, 60, 62–65,

254  Subject Index 66n6, 67n17, 75–76, 80, 82–84, 87, 89n11, 183 Millikan, Ruth 7, 13, 55–56, 72–73, 80–81, 97, 164, 205, 208, 216n8 Moore, G.E. 141 Moreno, Alvaro 9, 58, 62, 64, 67n15, 118n3, 221, 242n4, 243n8 Mossio, Matteo 9–10, 58, 62, 64, 67n15, 118n3, 221, 242n4, 243n7, n8

Radcliffe-Brown 2, 5, 20, 28, 48–53, 56, 63–64, 66n4, 79–85, 87, n7

Nagel, Ernest 72, 88n6 Nash, David 116 Neander, Karen 7, 9, 97, 184, 205 Nietzsche, Friedrich 200, 202, 205–206

Taylor, Charles 102 Tilly, Charles 2, 20 Tracy, Jessica 93, 95, 103–106, 111–113

Oderberg, David 227–228 Parsons, Talcott 4–5, 20, 70, 80, 82, 87 Perrow, Charles 34–36, 38, 42 Peterson, Jordan B. 113 Pettit, Philip 6, 12, 14n4, 71, 77, 89n9, 135, 137, 139, 142–145, 154, 155n5, 156n6, 202, 209–210 Pierson, Paul 191–193, 195, 198n19 Plato 126, 128, 133n8

Saborido, Cristian 9, 58, 62, 64, 66n7, 67n15, 118n3, 221, 243n8 Searle, John 12, 55, 118n3, 155n3, 177, 206, 208 Spencer, Herbert 4, 70, 86–87, 141, 154 Street, Sharon 135, 137–139

Ullrich, Wolfgang 189 Williams, Bernard 200, 202–206, 208, 215n2, n5 Wright, Crispin 147–149, 155, 156n14 Wright, Erik O. 46, 65 Wright, Larry 7, 88n6, 97, 136, 197n15, 205, 227 Zack, Naomi 179