Rethinking the Individualism-Holism Debate

The debate about individualism versus holism is one of the oldest and best known, yet it is also one of the most confused debates in philosophy of the social sciences. It might not be an exaggeration to say that there are as many versions of individualism as there are scholars committed to it. Isolated philosophizing, talking at cross-purposes, and failing to engage with actual social scientific practice have made the debate stagnant and a novel contribution particularly challenging. The book under review attempts to remedy this situation by confounding preexisting positions and putting the arguments and counterarguments in an effective context, which is dearly needed and thus most welcome. The self-proclaimed goals of the book are improving conceptual clarity of the major concepts and articulating what is really at stake; in short, rethinking the debate is what the reader is promised to get.

The book contains 12 papers, which are subsumed under two main headings, the debate about ontological individualism–holism and the debate about methodological individualism–holism. The first concerns issues around the ontological status of social phenomena and their relationships to individual agents, the second concerns questions around the extent to which social scientific explanations could and should focus on individuals and social phenomena, respectively (p. 1ff.). Both are certainly closely linked but nevertheless prioritize different questions and as such highlight distinct aspects. All 12 papers are written by a highly diverse set of prominent scholars (however—notably—the only contribution by a female scholar is the editor herself), who are known protagonists in the debate. As such, a collection of papers by this lineup of authors gives readers a very good overview of the major questions and positions currently under discussion and the main concepts and tools that are used.

The book starts with an introduction that lays out the conceptual basis and offers a set of distinctions between concepts, positions, and theories relevant in the individual-holism debate that are oftentimes mixed up in the literature. Relevant also is the editors’ reminder that the ontological and methodological individualism–holism debates link to two closely related debates, the micro-macro debate—revolving around the distinction between micro- and macro-level phenomena—and the agency-structure debate, which relies on a differentiation between agents and structures. While those interesting distinctions are unfortunately not analyzed in greater detail, they give a good entry to the papers that follow.

The first paper is by Brian Epstein who begins by taking a metaphysical perspective on the debate about ontological individualism–holism. He discusses the distinction between ontological individualism and anchor individualism as two metaphysical theses often conflated in the literature. Epstein defines ontological individualism as a thesis about social facts or phenomena, stating that they are exhaustively built out of, or dependent upon, individualistic ones (p. 18). Ontological individualism is a thesis about supervenience or the exhaustive determination of social facts by individual facts; ontological individualists accept a grounding relationship for social phenomena. Anchor individualism in turn is a thesis about the role of individuals in building the world. The thesis holds that social facts and phenomena are exhaustively anchored by individualistic facts (p. 19), whereby an anchoring relation is to be understood as a relation that determines the “putting in place” of social facts by individualistic facts (p. 18). According to Epstein, this “putting in place” relation is captured by a so-called frame principle. Such frame principles determine the conditions under which a social phenomenon exists or something has a social property. For such frame principles to hold, we as a society agree on a set of anchoring facts that fix the framing principles that ground a social fact.

Epstein’s point is that both theses—ontological and anchor individualism—are not only independent from, but are in fact in tension with, each other. The main difference between them is the point at which individualism has to hold. While the ontological individualist rejects non-individualistic facts as grounds of social facts, the anchor individualist rejects non-individualistic facts only as anchors but not as grounds that are determined by those anchors. By keeping anchors and grounds separate, anchor individualists can accept non-individualistic facts; social facts for the anchor individualist do not have to supervene on individualistic facts.

The second paper, by Dave Elder-Vass, casts doubt on methodological individualists’ tendency to reduce causal powers of social structure to those of individuals. Toward this goal, he offers an emergentist justification for the causal efficacy of social entities that aims at preserving the causal relevance of social structure in sociological explanations. First, he presents a set of arguments against what he calls residual Cartesianism in philosophy of mind, the idea that mental properties are fundamentally distinct from other kinds of properties and therefore exempt from the laws of causality (p. 41). He argues, first, against the shift of focus away from entities to properties underlying residual Cartesianism. Second, he argues against the view that mental properties are fundamentally different or collectively autonomous from physical properties, a distinction that commonly motivates doubts about the causal efficacy of mental properties in philosophy of mind. And third, he argues against the treatment of physical properties as being located on a lower level than mental properties.

The idea behind his emergentist account is that social structure can be causally significant, if social entities have emergent causal powers. The account is based on a relational theory of emergence, meaning that causal powers are said to be emergent in that they arise from the interaction of social entities that stand in particular relationships with each other. He illustrates his account by introducing the concept of “norm circles” (p. 50). Norm circles are collectives committed to adopting and enforcing specific norms that causally produce social practices we want to explain. Only as collectives and only because they stand in a particular relationship with each other does their capability to causally influence people’s behavior toward normative institutions and social practices emerge from their interaction.

The third paper, by Daniel Little, makes the case for an actor-centered view in sociology and for an enriched theory of the agent inspired by the American pragmatist tradition of John Dewey. Little departs from an account of what he calls “methodological localism” hat is positioned between methodological individualism and holism. This account captures the idea that sociologists should focus on actors, yet emphasizes their “social-ness” by being socially embedded in (a) a set of social relations and institutions and (b) a set of values, commitments, emotions, social ideals, and ways to conceive the world. Sociological explanations, according to Little, should focus on detailing ways in which action and the context of such socially situated individuals bring about social phenomena. Little rejects what he calls the Aristotelian approach underlying for example expected utility theory, which characterizes action in terms of a means-ends relationship. He considers the approach to be a too simplified account of social action and social interaction. It seems, however, that he thereby falls short of giving a clear-cut argument. Inspired by alternative sociological theories of action in the New Pragmatist tradition—such as those of Neil Gross, Andrew Abbott, Mark Granovetter, and Hans Joas—Little concludes by arguing for a “more adequate” theory of the actor for sociology (p. 71) that pays attention to the “specifics of the actors whose thinking and actions constitute the social processes of interest to them” (p. 73).

In the fourth paper, Philip Pettit discusses three core issues in social ontology. First, he identifies the “individualism issue,” which is concerned with the question of whether social laws impose a top-down constraint on individual autonomy as entailed by social determinism à la Émile Durkheim. Second, he takes up the “atomism issue” concerned with the question of whether social interactions serve as part of the infrastructure of intentional autonomy. And third, he takes up the “singularism issue,” which is concerned with the question of whether groups can in fact rival individuals in that they can achieve intentional autonomy as collective—or what he calls corporative—agents (p. 77). Pettit argues in favor of an individualism of groups and offers arguments in support of anti-atomism and anti-singularism. In making his claims, he stresses the importance of intentionality and rationality characterizing human beings. Besides his view that understanding psychology is not only necessary but also apparently sufficient for social scientific explanations, Pettit argues that the major claims defended by anti-individualists would contradict the idea of human beings as intentional actors that equips us for interpersonal interactions in society. Giving up this “intentional . . . image” of ourselves seems to be an implausible view (p. 82). The methodological implications of his view are that only social laws and explanations should be sought that “make psychological sense”; that anti-atomism, however, establishes that psychological explanations should be grounded in patterns of conceptualizations—for example, norms—commonly established in a society; and that we should use the “intentional stance” to interpret individual action as one explanatory strategy to make sense of group action (p. 94). The normative implications would be that social arrangements should be assessed by prioritizing individual interests and ascribing rights and responsibilities to collective agents on the condition they serve individuals’ interests.

Andras Szigeti’s paper questions the compatibility of individualism and anti-singularism underlying Pettit’s position. He takes up this issue by discussing the plausibility of “responsibility-collectivism,” the view that the moral responsibility of at least some collectives goes beyond the moral responsibilities of their individual members. By identifying “agency” and “having control” as two main requirements for attributing responsibility, Szigeti challenges the so-called “programming account,” which opens up the possibility of non-causal group control that often underlies responsibility-collectivism. Assuming that groups have control over whatever they are responsible for while “programming” their members to do whatever the group is responsible for, responsibility-collectivists preserve the idea that groups can be held responsible while at the same time not being committed to the idea of causal efficacy of higher level entities. Thereby, the programming account helps to avoid two unattractive positions: first, the problem of denying the exclusion principle, the principle that if a lower level property can be causally efficacious in bringing about some effect, no property supervening on that lower level property can be causally efficacious in bringing out that effect; and second, the problem of denying the relevance of higher order properties to causal explanations by granting them causal relevance but not causal efficacy. Szigeti argues, however, that non-causal group control is not acceptable as a basis for attributing moral responsibilities because it does not provide the relevant information to reconstruct the causal history behind whatever it is that the collective is held responsible for. This would, however, be required for being held responsible in the first place. For Szigeti, only causal group control is an admissible basis for responsibility attribution. But then, two problems emerge. First, we have to attribute causal efficacy to collectives in our causal explanations, which might entail an inconsistency with ontological individualism and a contradiction with the exclusion principle. This problem, he argues, can be solved by committing to a counterfactual account of causation. Second, granting group control and causal efficacy to the collective entails excluding causal efficacy on the individual level and compromises individual control. That, however, goes against the responsibility-collectivist’s original intention to find a notion of group control that preserves the individual’s control and individual responsibility. This puts the responsibility-collectivist, who presupposes anti-singularism in Pettit’s sense, in tension with individualism, a price of committing to anti-singularism that Pettit and others might not want to pay.

The final paper concerned with the ontological individualism-holism debate is by Petri Ylikoski. Ylikoski takes a welcome pragmatic perspective to the micro-macro problem. He starts by pointing out that the biggest problem of the debate is that its results are unproductive for the social scientific practitioner; he wants to remedy this situation by formulating an account of micro-macro relations that can be fruitful for social scientific modeling and explanation. First, he calls into question recent attempts to think about the micro-macro problem in the social sciences by drawing the analogy to how philosophers of mind think about the relation between the physical and the mental. Ylikoski is particularly skeptical about thinking about micro–macro relations in terms of levels that are presupposed by concepts such as supervenience or emergence. He offers three main arguments to reject the analogy between the problems of philosophers of the social sciences and philosophers of mind. First, the debate in philosophy of the social sciences rests upon different presuppositions than the debate in philosophy of mind. The central problem for philosophers of mind is how explanations given by psychological theories that contain mental concepts are related to explanations given by neuroscientific theories. While both sets of theories rely on a different vocabulary, they talk about the same things and thereby accept a comprehensive and exhaustive neural-level understanding of psychological processes (p. 120). The problem in philosophy of the social sciences is not to bridge the gap between a comprehensive and exhaustive individual-level understanding of psychological processes and a social description. Rather it is whether or not individual facts are causally sufficient to account for all social facts; it is about how facts about individuals and their interactions relate to large-scale facts about groups, organizations, and societies (p. 121). Second, the individual level in the social sciences is not comparable to the physical level in neuroscience. The difference is that the supervenience basis for macro-social properties frequently includes things such as institutions, organizations, social artifacts, buildings, and so on, something that the physical level does not include and that questions causal closure of the individual level as a meaningful way of thinking about the social world. Third, the social is also not comparable to the mental. Unlike mental properties that are constituted by a vocabulary that is discontinuous with the physical-level discourse, macro-social properties are aggregates of individual properties, are similar to micro-properties, or are anthropomorphic projections of personal properties.

Given that the analogy does not hold, Ylikoski proposes a new conceptual start for the social sciences that implies replacing the level-based approach with a scale-based approach. To motivate his account, two observations are crucial: first, macro-social facts are typically supra-individual; and second, the relevant relationships are part-whole relationships between macro-social entities and their constituent parts. Both observations suggest an account that allows for thinking about the difference between macro and micro as one between small-scale and large-scale social phenomena that provides a fruitful heuristic for the practitioner to think about micro-macro relationships while avoiding problems resulting from the levels-conceptualization. The scale-based account has three important implications. First, there is no such thing as a unique micro-level in the social sciences. It avoids a categorical distinction between micro and macro and instead allows the contrast between small and large scale to be relative, which is the way social scientists talk. Second, it acknowledges the context dependency of the micro-macro relation and the idea that what counts as micro and what counts as macro is relative to the social scientist’s explanatory goal. Finally, it accommodates the idea of a meso-level prominently introduced in fields such as analytical sociology and often coupled with a mechanism-based approach to theorizing. Ylikoski marries this account with what he calls the contrastive-counterfactual account of explanation, which pushes constitutive rather than causal explanation in the social sciences. As such, on the scale-based view, there is a continuum of scales between any specified micro- and macro-scales. Which scale is the relevant one is an empirical question and should not be settled on a priori ontological grounds.

Harold Kincaid’s paper opens the section on methodological holism-individualism. Kincaid discusses “dead ends and live issues” in the debate, on which he takes a naturalist perspective. His judgments about issues being dead ends and being relevant are motivated not by a priori philosophical concerns in conceptual analysis or standard ontology but rather by their relevance for social research and as such are oftentimes to be decided empirically. What we then learn from Kincaid is that a standard reductionist reading of methodological individualism, social mechanisms, eliminativism, and explanation are all dead ends. Regarding reductionism, it comes in at least two forms, namely, that “all social science explanations must be entirely in terms of individuals and that they can be entirely in terms of individuals” (p. 141; italics in original). Kincaid first argues that there are at least four open issues that require fixing. First, there are several open questions such as the following: What are the restrictions on referring to individual psychology? Do we only need psychology or rather cognitive science to give individualist explanations? And how should we think of the relations between individuals in purely individualist terms? Second, what we should take to be an explanation in the social sciences is open. Is it causal or nomological explanation or is an explanation something altogether that aims at understanding? Third, does “can” in the thesis refer to something feasible or something in principle possible? How far we can go should be decided by carefully reviewing social scientific explanations. Fourth, what is the rationale to argue for individualist explanations anyways in the social sciences?

Kincaid argues against the argument from composition common among analytical sociologists and prominent defenders of individualism as a reductionist thesis, such as J.W.N. Watkins. This argument roughly states that

if entities of kind S are composed of entities of kind I and their behavior is determined by the behavior of entities of kind I, then it follows that it is possible to explain what we know about S’s in terms of I’s. (p. 142)

Kincaid’s main argument against accepting the reductionist thesis is that we do not have the conceptual resources to explain complex wholes from information about their parts. This is a limitation that we do not only confront in the social sciences but for which we have evidence from quantum chemistry (which cannot be reduced to quantum mechanical explanations), from molecular biology (where we cannot reduce genes to sequences of DNA only), and from complex systems theory (where we cannot explain the behavior of complex dynamic systems from the nature of individual agents alone). This body of evidence from science leads Kincaid to conclude that the argument from composition generally fails, although local reductions might be possible and fruitful in some cases.

Next, Kincaid argues against the need for individualist mechanisms in the social sciences. His basic argument is that what a mechanism is and what it is supposed to do are so diverse that we need to be very specific about what we aim for and decide its fruitfulness on a case by case basis. And often, examples from science—say, from macroeconomics—tell us that we do not go as far as possible to the individual level to get evidential support. This just speaks against adopting individualist mechanisms to be justifiable as a general principle for Kincaid. So, how holist or individualist can or must we be in the social sciences according to Kincaid? Kincaid thinks that we cannot fully commit to either alternative. There is too much evidence against being entirely individualist in science. But at the same time, holism does not seem to be plausible. So, how much social explanation do we need? For Kincaid, this is an issue to be decided empirically. For some explanations, we have to have detailed psychological knowledge while in other cases we need to add social variables to complement individualist factors as they play a bigger causal role in some instances than in others. The message we get from Kincaid is that ultimately science decides when individualism is useful and when it is not and that efforts to integrate the individual and the social level in, for example, multi-level structural equation models should be the new focus of philosophy of the social sciences to foster better social science.

Jeroen Van Bouwel’s contribution also departs from the imperative that the methodological individualism-holism debate should ultimately be fruitful for the practitioner. Van Bouwel attempts to live up to this imperative by outlining a framework for explanatory pluralism that is heavily informed by the plurality of explanatory strategies that we find in the social sciences. The core element of this framework is a set of different explanation-seeking questions that all could potentially be asked when trying to understand a specific social phenomenon. As those questions are requests for information provided on the basis of different explanatory models, answering them in the best possible way requires trading off various epistemic virtues such as accuracy, adequacy, and efficiency. When we apply our models to answer a set of explanation-seeking questions, those virtues are traded off according to our explanatory goals. One consequence is that the established “winner takes all approach” of explanation, according to which there is only one best form and/or level of explanation, is wrongheaded and that a toolbox is needed that contains different levels and forms of social scientific explanation appropriate for the distinct goals (p. 161).

Van Bouwel discusses the implications of his framework for debates about emergence, reduction, and microfoundations for the social sciences. First, regarding the position of ontological emergence, Van Bouwel’s framework calls into question the recurring inference by emergentists from higher level explanations to ontological emergence as unnecessarily restrictive as regards the explanatory strategies available. Second, Van Bouwel defends (but does not really argue for) reductive explanations as offering an accurate, adequate, and efficient explanation, depending on the explanatory interests of the scientist, while acknowledging the idea that the amount and specification of levels in the social sciences is perspectival (p. 166). Third, while Van Bouwel seems to agree on the recent defense of microfoundations as playing a justificatory rather than explanatory role in macro-level explanations, his account leads us to ask in greater detail how exactly microfoundations allow for satisfactory macro-level explanations. Van Bouwel ends by noting the recent trend in the debate away from a dichotomous way to think about individualism and holism and the winner-take-all approach and more toward acknowledging pluralism. He himself defends what he labels “interactive pluralism,” a middle position between integrative and isolationist pluralism discussed extensively by Sandra Mitchell (2004), which on one hand does not demand integration of levels in explanation while on the other hand does not discourage interaction and integration of levels in social scientific explanation. According to Van Bouwel, it ultimately depends upon our explanatory interests. Because the commitment of many social scientists to pluralism has so far, however, been rhetorical only, what remains is the question of how actively explanatory pluralism as a normative endorsement of plurality can be promoted as a new mind-set in the social sciences.

Zahle’s contribution engages in detail with Elder-Vass’s version of the argument from emergence in support of methodological holism. By appealing to emergence, it is argued that we cannot account for significant events in the social world, unless we also use holist explanations (p. 178). Zahle reconstructs Elder-Vass’s version of the argument that relies on two kinds of ontological distinctions. First, we can distinguish between three kinds of (causal) properties that social entities can have, namely, non-emergent, resultant, and emergent properties. Non-emergent properties are properties that parts possess in isolation, that is, also if they were not part of a whole. Resultant properties are possessed by wholes but they result from the non-emergent properties of their parts if aggregated. And emergent properties are possessed by wholes; they can be of two kinds: emergent properties of wholes proper, that is, properties not possessed by their parts; and properties of parts that they, however, would not possess in isolation, that is, if they were not part of a structured set of components. It is the latter properties that are relevant for Elder-Vass’s argument for methodological holism. The need for holist explanation arises because those properties can causally explain the effects that emergent properties of wholes have and that are not reducible to the effects of non-emergent properties possessed by their parts. Therefore, holist explanations are indispensable and as such also widely used in the social sciences. Zahle then argues that methodological individualists have nevertheless reason to resist giving up their position, just because Elder-Vass’s argument relies on a rather narrow conception of individualist explanation and a clear-cut distinction between individualist and holist explanation that we do not necessarily have reasons to uphold. In fact, Zahle argues that Elder-Vass does not give good reasons for upholding the narrow distinction he draws between individualist and holist explanations. Addressing the question whether it is possible at all to draw a particular distinction between the two kinds of explanation, Zahle concludes by arguing that a good reason for it is that it is useful from the perspective of explaining the social world.

Finn Collin approaches the issue around individualism and holism from the perspective of Actor Network Theory (ANT), an approach that has recently gained prominence in social theory and been applied mainly in science and technology studies. While it places the actor at the center of social analysis, it suggests a highly innovative category of an actor. The extensive work of its main protagonist, Bruno Latour, serves Collin as a template through which he discusses the implications of ANT for methodological individualism and reductionism; prima facie, both seem to be methodological positions that could be justified through Latour’s metaphysical and methodological commitments. Collin introduces the major concepts in light of their larger context, thereby attempting to offer a coherent picture of the Latourian philosophy and methodology of the social sciences. A core concept is that of an “actant,” an entity with the sole characteristic of being active, that is, acting upon other actants. Not only is the category of actants fundamentally relational and dynamic; actants are also considered to be particulars as opposed to kinds having certain essences, which impedes thinking of them as subjective human agents that have to be contrasted with the objective natural environment. Rather, the social world is made up of hybrids bridging the gap between the social and the natural. On this picture, society consists of a network of highly heterogeneous actants; social structure is at best a working category for formulating the explanandum of sociology, which is not to be taken literally. Furthermore, sociology being a science of the particular only and committed to methodological particularism, ANT’s focus lies on detecting concrete mechanisms that can link the micro- and the macro-entities, which demands a micro-analytic perspective. At the same time, talk about levels and prioritizing a specific level as privileged are rejected by Latour’s strong resistance to reductionism. Every actor is irreducible in that it is unique in the kind of difference it makes on the outcome.

Collin draws some methodological lessons from this complex philosophical picture. First, the actant as a more inclusive concept of a social agent prohibits a narrow understanding of methodological individualism. Second, by requesting a concrete mechanism behind social action and interaction in each case, Latour questions micro-analyses that take the conventionally conceived actor as the main stopping point. Committing to methodological individualism is at best a decision that is pragmatically motivated, as the social world should not be divided hierarchically into levels; Latour defends a “flat ontology.” Collin does well in qualifying the expectations that we might have with respect to the implications of Latour’s view for the individualism/holism debate. His views are still too complex, too divergent, and in parts too inconsistent to serve as a clear-cut philosophy and theory of society. However, we can take Latour’s views as a “laboratory in which novel methodological ideas . . . are being tested out on a body of concrete empirical material” (p. 216).

Mark Risjord takes up the issue of agency in the social sciences. His criticism in a nutshell is that the current notions of agency available—and he mainly discusses instrumental rationality, Pierre Bourdieu’s concept of habitus, and Antony Giddens’s notion of practical consciousness—cannot capture the relational dimension of agency in relation to social structure. Risjord begins by detailing an example of jazz improvisation, highlighting the relational dimension of creating, maintaining, and re-establishing the musical structure via mutual responsiveness in a dynamic environment. According to Risjord, such joint action goes beyond signaling and choice among the musicians. According to Risjord, agency requires reconceptualization to capture interaction in this case. All three cases—the rational chooser acting upon his preferences, the habitual agent acting according to her dispositions that are shaped in turn by social patterns, and Giddens’s account of agency in terms of various levels of consciousness produced by and reproducing social structure—have limitations as an appropriate account of agency in the case of jazz. The rational actor model, by focusing solely on choice, misses the responsiveness of the actor to the behavior of other actors and the thereby permanently changing environment. While Bourdieu and Giddens highlight the influence of social structure on the environment, the limitations of both accounts originate in the recursive relationship their view commits to between macro-structure and individual behavior. A recursive relationship—by focusing mainly on the reproductive power of social practices that agents internalize and thereby reproduce—cannot capture the recreation of a macro-structure after its breakdown.

Risjord suggests an alternative, what he calls an ecological account of agency as a threefold account that goes beyond treating agency as an individual phenomenon, that is, the elements of meta-cognition, attunement, and affordances—informed by known psychosocial cognitive capacities. Meta-cognition and attunement ensures the improvising maintenance of social structure via the capacity to be attuned to one’s environment. Affordances are possible actions available to the agent in a particular environment and capture the capacity to recognize one’s actions and actions of others within a specific environment. And lastly, we have the capacity to meta-cognitively reflect on one’s interaction with the environment and relate one’s ongoing performance to prior plans and propositional knowledge (p. 234f.). Given those elements, Risjord’s account emphasizes the relational character that is essential to the very idea of agency, which makes it distinct from the three accounts he discusses, yet makes it informed by several of their characteristics.

The last paper, by Mathew D. McCubbins and Mark Turner, is narrower in focus. It discusses the rationality assumption underlying current individualist accounts of strategic interaction that are used to model large-scale events, mainly in game theory. The claim defended is that rational choice conceptualized as consistent choice is an unrealistic assumption and should therefore be put under heavy scrutiny before we use it to make generalizations about human behavior. The paper discusses various accounts that have empirically challenged the rationality assumption, such as the work on framing by Daniel Kahneman and Amos Tversky, the research on heuristics and biases by Gerd Gigerenzer, and empirical research on risk attitudes. While McCubbins and Turner acknowledge this research, they argue that for such results—subsumed under the heading of behavioral game theory—to enable behavioral generalizations for making predictions, behavior has to be assumed to remain consistent across contexts. Testing empirically for such consistency, the paper presents some interesting results from a set of experiments that—put together as a battery—support the claim that the consistency assumption of behavior is unwarranted. The idea underlying the design of an experimental setup is that if people behave consistently according to their preferences, they should do so across very similar game situations. What experimental evidence suggests, however, is that, whether testing for fair or egotistical or some other behavior entirely, the same sample of individuals does not behave consistently across games. The main conclusion the authors draw is that the social sciences should, rather than using the consistency assumption largely uncritically and unscrutinized to make predictions, continue critically testing the assumption to ultimately aim at recasting the foundations of the field (p. 250).

Zahle and Collin have collected a set of rich papers that together offer a multilayered perspective on the individualism–holism debate. As such, the collection provides an excellent overview of the relevant positions currently defended in the literature. Some contributions, such as for example Brian Epstein’s paper, make an attempt to clarify core concepts such as “individualism” that are key in this debate, however paying the price of mainly keeping it at an abstract metaphysical level where questions about the relevance and usefulness of those conceptual results for concrete social scientific practices remain unanswered. Other papers, such as the one by Harold Kincaid, directly question the relevance of most aspects raised in the current debate for making progress in the social sciences and in philosophy of the social sciences. To reconnect the debate with social scientific practice, scholars such as Ylikoski attempt to show the links between a set of philosophical concepts and some current trends in analytical sociology, which is dearly needed and would—if followed through—at least contribute to the debate’s “rethinking” that the editors envisage. The observation and call for reflection upon whether, and if so in which way, a debate that has become highly abstract and thus increasingly remote from scientific practice should be informed by actual social scientific practices is one of the more important insights of the volume.

The diversity in contributions is a virtue of the book but it also comes at the cost of opening up new avenues to rethinking the debate. While most papers are instructive taken as separate contributions, after reading the collection from beginning to end, the book leaves the reader unsatisfied and a bit lost. First, the volume remains disconnected from what has so far been discussed in the individualism–holism debate. While some papers offer new analytical tools, they remain fully isolated from the debate’s precursors. This is unfortunate because the history of this debate is extremely rich, contained a large set of fruitful contributions, and had been historically pushed by prominent scholars such as Max Weber, Karl Popper, Friedrich August von Hayek, among many others. Their contributions are fully ignored in this volume, yet their absence not even justified. That leaves the reader with curiosity of how those novel contributions are to be located. Second, each contribution stands largely on its own. Neither is there an extensive attempt made by the authors to draw upon each other nor do the editors make any connections explicit and discuss them toward their self-proclaimed goal. There is no conclusion that would offer the reader some guidance in synthesizing what has been presented. It then becomes clear that the introduction—while giving a good entry into the debate—is not particularly helpful. The conceptual background of the introduction could have been useful to discuss and, where applicable, synthesize the conclusions of the various contributions to locate them in the field, but the introduction remains isolated from the rest.

In closing the book, one thus wonders what has been achieved. As the commonalities and differences of the various contributions are not discussed, it is not clear how they help us to rethink the debate; diversity is what remains. The reader has to figure out the implications of this diverse set of arguments for herself, in addition to assessing their importance and relevance for the debate. How exactly do all of them link together? How does all of this connect to the fundamental and highly important contributions that have been made throughout the history of the social sciences? And, against this background, why and in which way exactly do those arguments help us to rethink the debate? Having left those questions open is unfortunate, given the importance of the project, the need to give some direction, and the impressive lineup of contributors.

The book has certainly taken up a highly important issue. One of the oldest philosophical debates in the social sciences is in dire need of rethinking and advancement to regain relevance. But such scattered, specialized, and abstract research as offered here will not suffice. Admittedly, the project of Rethinking the Individualism-Holism Debate might be one of the biggest challenges for philosophers of the social sciences. This book—while interesting and relevant in its own right—does not fully meet the challenge. Yet, it hopefully motivates further research that tries to establish connections, between philosophy and scientific practice, between the historical and contemporary contributions to the debate, and between the multiple yet separate paths along which this debate has been advanced in recent years.