Naturalism and Anti-Naturalism

The Naturalist versus Anti-Naturalist Debate By Mark Bevira and Jason Blakelyb (for The Routledge Handbook of Interpretive Political Studies, eds. Mark Bevir and R.A.W. Rhodes) aDepartment of Political Science, University of California, Berkeley, CA 94720-1950. Email: mbevir@berkeley.edu bDepartment of Political Science, Pepperdine University, 24255 Pacific Coast Highway, Malibu, CA 90263. Email: Jason.Blakely@pepperdine.edu 1 The Naturalist versus Anti-Naturalist Debate Contrary to widespread belief, the naturalism and anti-naturalism debate in political science is about philosophy not methods. Because this conflict is philosophical and not methodological, political scientists are free to embrace whatever methods suit their research purposes. That is, method pluralism can reign supreme over political science. But this method pluralism does not imply that the naturalist/anti-naturalist debate is safely forgotten or somehow resolved. On the contrary, the dueling philosophical assumptions facing off in this debate remain of the greatest importance to the study of politics today. What are these dueling assumptions? Naturalism is the assumption that explanations in political science should be formal, ahistorical, and invariant like those often found in the natural sciences. The philosophic roots of naturalism are found in the Vienna Circle, logical positivism, British empiricism, and early analytic philosophy (Ayer, 1952, p. 48; Carnap, 1929, p. 331; Neurath, 1931, p. 48). By contrast, anti-naturalism is the view that human beliefs and actions are expressive of meanings, making political inquiry incompatible with the naturalist quest for formal, ahistorical, and invariant explanations. The philosophic roots of anti-naturalism are found in German Romanticism, phenomenology, idealism, and post-Wittgensteinian analytic philosophy (Dilthey, 1976; Collingwood, 1946, pp. 285-288; Husserl, 1936; Winch, 1958). The purpose of this chapter is to survey the naturalist/anti-naturalist debate as a specifically philosophical conflict, with particular attention paid to how this conflict has generated anti-naturalist critiques and alternatives to the study of politics. To this end, we have divided the chapter into three parts. In the first part we survey the basic philosophical differences between naturalism and anti-naturalism, looking at how each engenders rival ways of thinking about methods and 2 explaining political reality. In the second part of the chapter we will see how these philosophical ruptures lead anti-naturalists to reject naturalist concepts. Specifically, where naturalists fashion concepts meant to capture the reified and essential features of political life, anti-naturalists rebuff any such reification and essentialism as distortions. The third and final part of the chapter then explores how, although unified in their rejection of naturalism, anti-naturalists nevertheless splinter into various schools of philosophic thought. Two of the most important are poststructuralism, as exemplified by Michel Foucault, and hermeneutics, as exemplified by Charles Taylor. One caveat, however, is necessary before commencing. Perhaps few political scientists today would wish to identify as naturalists. Indeed, where once naturalism was openly championed by the leaders of the discipline, the vogue today has shifted towards multi-methods, as is evident in political science methodologists like John Gerring (2001; 2011), Gary Goertz (2012), Robert Goodin (2009, pp. 9-12), Janet Box-Steffensmeier, Henry Brady, and David Collier (2008, pp. 29-30), and Ann Chih Lin (1998) among others. In one sense, we strongly support this shift towards method pluralism. However, an important proviso must be added: namely, often multi-methods advocates assume that naturalists and anti-naturalists can carry out a peaceful, untroubled coexistence so long as they adopt an ecumenical stance on methods (BoxSteffensmeier, Brady, and Collier, 2008, p. 30; Gerring, 2001, p. 9). Unfortunately, however, this sunny vision of an untroubled peace misunderstands the basic distinctions of the naturalist/antinaturalist debate. For as we now hope to show, these two groups remain at odds at the most fundamental philosophical level. 3 1. Naturalism and Anti-Naturalism Naturalism and anti-naturalism are incompatible on the most basic philosophical level. Where naturalists assume that explanations in the social sciences should be formal, ahistorical, and invariant, anti-naturalists believe that these explanations must instead involve the reconstruction of historically contingent meanings. Because the dividing issues in this debate are philosophical and not methodological, naturalist political scientists may legitimately employ what are sometimes thought of as “anti-naturalist” methods like ethnography, open-ended interview questions, and close textual analysis. Likewise, anti-naturalists commit no faux pas when they avail themselves of what are frequently considered “naturalist” methods like mass surveys, polling, or large-N statistics. Yet despite being able to share methods, naturalists and anti-naturalists also have very different views of the very methods that are in principle available to both. In the case of naturalists, the assumption that explanations are formal, ahistorical, and invariant often leads them to treat methods as if they were logics of discovery. A method becomes a logic of discovery whenever it is assumed that this method is a prerequisite to the practice of good political science. For example, political scientists turn methods into logics of discovery when they assume that a certain kind of statistical analysis or polling is necessary in order to conduct sound research. A logic of discovery thus consists of the belief that a formal, invariant procedure, rigorously carried out, secures scientific results. By contrast, anti-naturalists do not think of methods as necessary procedures of discovery. Instead, anti-naturalists tend to treat methods as heuristic techniques informing a skilled craft. Consider the craft of fishing. A fisherman may use whatever techniques he prefers 4 for fishing provided the results are acceptable. Thus he may use a fishing pole, a net, a trap, a spear, or even his bare hands. No matter what technique is used, the craft of fishing is developed within the context of some tradition that provides heuristic techniques and rules of thumb. In other words, fishermen can draw on a stock of knowledge that guides them in their craft. Similarly, anti-naturalists hold that the practice of political science is foremost a craft. This means political scientists may use whatever methods they wish so long as the results are acceptable. Because skilled craftsmanship can always be creatively supplemented by ingenuity or luck, anti-naturalists believe that no method can simply be treated as a mechanical procedure or recipe for success. That is, in contrast to naturalists, anti-naturalists argue that no invariant, formal set of steps can permanently replace the work of exercising a skill. Rather, anti-naturalists hold that political scientists, like fishermen, also draw on traditions which instruct them in how to skillfully employ whichever techniques suit their purposes (from reading texts to conducting statistical research). Thus, even though naturalists and anti-naturalists share methods, their philosophic differences lead them to treat those methods differently. Where naturalists envision political science research as guided by formal logics of discovery, anti-naturalists instead see it as consisting of the heuristic techniques of a craft. But the philosophic difference between these two camps runs much deeper than even this basic incompatibility. In this section we survey two further ruptures between naturalism and anti-naturalism. First we will see that while naturalists treat political reality as comprised of isolated atoms, subject to measurement and correlation, anti-naturalists instead view it as composed of holistic webs of meaning. Second, we will 5 contrast naturalism’s search for explanatory causal laws with anti-naturalism’s attempt to construct historical narratives. i. Atomistic Units versus Holistic Meanings One of the most important philosophic differences between naturalists and anti-naturalists is that while the former atomize political reality into basic units that can then be correlated and measured, the latter instead treat it as comprised of holistic webs of meaning. Let us look at each of these two competing orientations in greater detail. Naturalist political scientists often assume that inquiry consists of a process of isolating units of political reality and correlating them to other such units. This process of atomization allows naturalists to single out bits of reality that might then be found repeated across vastly different contexts. For example, in their classic of behavioral political science, Gabriel Almond and Sidney Verba (1965, pp. 307-315) treated certain beliefs—e.g., how much “pride” the Italians versus the British felt for their respective governments—as isolated units that could be measured, correlated, and compared across widely different cultures. As we will see below, this picture of political reality as built out of basic, atomistic building blocks is directly tied to the naturalist quest for formal, invariant laws (MacIntyre, 1978). But for now the key point is that naturalist inquiry often consists of this attempt to breakdown political reality into measurable, isolated units. This stance puts naturalism directly at odds with the anti-naturalist view. For antinaturalists hold that human beliefs and actions should not be broken down into atomistic units, 6 but rather must be contextualized within wider networks of meaning (Taylor, 1985b). Political reality, on the anti-naturalist view, is not comprised of atomistic properties but of holistic webs of belief. Beliefs are holistic because they are always supported by further beliefs. Antinaturalists thus tend to treat human beliefs similar to the way literary scholars treat a passage from a novel—that is, they interpret and contextualize it in light of the further meanings and beliefs that inform it. The importance of this difference for political science will perhaps be illuminated by an example. Consider the case of an anti-naturalist who wishes to give an account of the political slogan “el silencio es salud” (silence is health) which was commonly expressed by individuals with authoritarian sympathies during Argentina’s Dirty War (Feitlowitz, 1998). At the most basic level, the very meaning of el silencio es salud, requires some background knowledge of the words and grammar of the Spanish language as well as the relevant historical and social context—in this case, things like Argentine authoritarianism in the 1970s, practices of state terrorism, the Cold War, and so on. Of course, so far naturalists like Almond and Verba might well agree that such background knowledge is necessary. However, what distinguishes the anti-naturalist approach is the key claim that the very meaning of the slogan is dependent upon the particular webs of belief that inform it. This crucial claim implies that meaning is not present in an isolated unit but is the function of a holistic context. So, a man might repeat the slogan “silence is health” because of his various beliefs about Argentina’s national security, the rowdiness of subversives, and the need for the state to silence dissenters. By contrast, another individual might adopt this slogan for very different reasons—for example, as expressing the ironic and cowardly view that 7 resisting evil is almost always futile and that surviving state terrorism in good health means remaining silent. In this way, anti-naturalists conclude that explaining and understanding particular meanings requires interpretively reconstructing the right context of beliefs and not isolating and correlating atomistic facts. Moreover, anti-naturalists hold that what is true of beliefs is also true of human actions— namely, that they must be unpacked within a holistic context of belief. Here anti-naturalists defend the thesis that actions are expressive of beliefs, a point often argued by drawing a basic distinction between actions and movements (Geertz, 1973, pp. 6-7; MacIntyre, 1964, pp. 56-66). A mere physical movement, anti-naturalists note, might be identical to an action. For instance, the involuntary twitch of an eyelid may be indistinguishable from the wink of an eye. Yet the difference between the two is vast because winking, as an action, is expressive of beliefs. A wink might be a sign of flirting, a warning, an agreed upon code, a joke, or a salutation. Thus, properly identifying actions requires linking them to the relevant web of beliefs that inform them. Furthermore, anti-naturalists conclude that because political reality is made up of human beliefs and actions, it is essentially expressive of holistic meanings. In other words, political reality itself must be interpreted in a way analogous to the study of a text, and not treated as a collection of isolated, repeating atoms. This marks a basic philosophical incompatibility between these two camps. Where naturalists attempt to subject political reality to a process of atomization, correlation, and measurement, anti-naturalists instead contextualize and interpret beliefs and actions within holistic webs of meaning. But this basic difference is in fact integrally related to a further incompatibility—namely, how political scientists ought to explain political reality. 8 ii. Causal Laws versus Historical Narratives Naturalist attempts to atomize political reality are usually part of a larger quest to discover the formal causal laws explaining political life. In sharp contrast to this, anti-naturalists view political reality as expressive of holistic meanings that must be explained through historical narratives. Naturalists and anti-naturalists thus advance philosophically incompatible forms of explanation: the one causal, ahistorical, and invariant; the other narrative, historical, and contingent. The naturalist attempt to explain politics in terms of general causal laws is evident in a wide variety of schools, including not only certain versions of Marxism but also mainstream, behavioral political science (Marx, 1859, pp. 388-391; Gunnell, 1975, pp. 65-67, 84). In its early years, the behavioral movement often drew support for this view of explanation by appealing to analytic philosophers like Carl Hempel (1965, p. 240; 1948) and A.J. Ayer (1952, p. 48) who argued that every bona fide science required the establishment of general causal laws. Explanation, on the view of these philosophers, consisted of finding the necessary causal connections between two states of affairs, one antecedent and the other a consequent. Once a necessary causal mechanism had been identified between two atomistic bits of reality a general law could be formulated. Naturalists’ tendency to atomize political reality thus ties in directly to their quest for correlating states of affairs in the hopes of discovering general causal laws. This search for necessary, causal laws once again highlights the extent of the incompatibility between naturalism and anti-naturalism. For we have already seen that anti9 naturalists view particular beliefs and actions as the result of the contingent reasons and beliefs that support them. Yet viewing beliefs as the result of such contingent reasoning implies that individuals could have reasoned differently, leading to other outcomes. In other words, the antinaturalist view is that beliefs and actions are the result of contingent reasons and not necessary causal links. This means that as long as beliefs are understood as existing in light of other beliefs, human behavior cannot be subsumed under determinate causal laws (Davidson, 2006). Instead of causal laws, viewing beliefs as the result of contingent processes of reasoning strongly implies that political science explanation is a form of historical narration. This is because reconstructing contingent beliefs and reasons involves telling narratives or stories. Many anti-naturalist philosophers have argued that human action is structured by narrative sequences (Heidegger, 1953, pp. 292-304; Ricoeur, 1983-1985). The basic point here is that, human beliefs and actions are not carried out in atomistic segments, but rather are integrated into a context of goals, beliefs, and desires that relate the past, present, and future to one another. In this way, human agency exists in a narrative stream that orders temporally remote episodes. Thus many anti-naturalist philosophers maintain that human agency actually embodies a narrative and that human beings are “a story-telling animal” that continually enacts stories (MacIntyre, 2007, pp. 216, 211-212; Taylor, 1985c). Political scientists must therefore fit beliefs and actions within sequences of past, present, and future. Or, put more simply, they must construct narratives (MacIntyre, 2007, p. 208). However, an immediate caveat must be added to this exposition of the anti-naturalist position. Namely, the anti-naturalist insistence on narratives does not negate method pluralism. Rather, it only requires that whatever methods a political scientist selects (be it statistics, mass 10 surveys, or individual interviews), he or she must ultimately evoke narratives of contingent belief and not necessary laws as explaining human behavior. Likewise, a similar caveat goes for the naturalist position: namely, a commitment to causal laws doesn’t bar the use of ethnographic research or close textual analysis. Thus we return to one of our major claims: the difference between naturalism and anti-naturalism is philosophical and not methodological. Despite their irreconcilable philosophic differences, both camps can avail themselves of whatever methods they choose. 2. Against Naturalist Concept Formation So far we have explored two philosophic ruptures between naturalism and antinaturalism, the first concerning atomism versus holism, and the second causal laws versus narratives. We will now examine how these basic philosophic differences help drive the antinaturalist critique and rejection of naturalist concepts in political science. Naturalists often either reify or essentialize political reality. This move reflects the basic naturalist mission. That is, reification and essentialism reflect the naturalist hope of discovering law-like explanations that are formal, ahistorical, and invariant. In what follows we look at reification and essentialism each in turn, drawing on examples from contemporary political science research. Approaching the topic in this way underscores our ongoing point that although the naturalism/anti-naturalism debate is about philosophy and not methods, it nevertheless remains of chief importance to the study of politics today. 11 i. Reification Reification occurs when political scientists treat beliefs, actions, and practices as nonintentional objects that can break free of a single context of intentional meaning. Instead of getting entangled in particular contexts of meaning, the appeal of reification is that it allows political scientists to conceive of political reality as comprised of movable parts that can migrate from context to context without distortion (a kind of movable furniture). Reification is thus clearly related to the naturalist form of inquiry we explored above, which seeks to isolate bits of reality that then form the basis for causal laws. There are two general types of reification that are common to political science today. The first is what might be called brute fact reification. This type of reification is the result of stripping units of political reality of intentional meanings almost entirely in favor of brute demographic, biological, or other such facts. The second kind of reification common to political science is atomistic reification. Atomistic reification is the result of abstracting or otherwise isolating intentional meanings from their original, holistic context. Once they have been atomized, these concepts can be treated as independent variables. Let us look at each of these forms of reification more carefully in light of contemporary political science. Researchers of voter behavior often reify human identity by treating it as a brute fact. For example, many American political scientists today believe there is a strong correlation between the brute biological/demographic fact of age and the motivation to vote in presidential elections (Kernell, Jacobson, and Kousser, 2011, pp. 496-497). The two political scientists who have done the most to corroborate this claim are John Mark Hansen and Steven Rosenstone (1993, p. 128), who together have gathered an impressive store of data in order to indentify the “causes” and 12 “personal determinants” of voter participation. Surveying American presidential elections from 1956 to 1988, Rosenstone and Hansen found that voter turnout among the oldest citizens was twenty-nine percentage points higher than among the youngest. This led them to the general conclusion that “as people grow older, their involvement in American politics deepens” (Rosenstone and Hansen, 1993, pp. 136-137). Rosenstone and Hansen’s research exhibits several of the features of naturalism we have discussed so far. First, age is conceptualized as a reified, brute fact. This fact is then correlated with another brute fact like the act of voting. After enough data has been aggregated the hope is that basic causal or at least correlative patterns will emerge. In this way, the reified fact of age plays an essential role in the attempt to causally predict voter behavior. Anti-naturalists, by contrast, reject Rosenstone and Hansen’s reification of age on the grounds that it neglects the intentions and beliefs of particular agents. Anti-naturalists note that the political significance of age is always the result of a particular intentional context. Thus, they would argue that although there is indeed the brute biological/demographic fact of being a certain age, they would nevertheless insist that there is no brute fact as to the meaning or significance of this for the motives and purposes of the agents in question. Rather, the significance of age for voter behavior depends on the beliefs and intentions of particular actors. Thus anti-naturalists might point to the existence of enclaves of elderly Americans that are either highly apathetic or antagonistic to voting—for example, the last remnants of the “turn on, tune in, drop out” ethos of the 1960s. Similarly, they might point to segments of American youth that have become politically energized and enthusiastic voters. The purpose of such counterexamples would be to show that there is no mere social fact of “being sixty-five” or 13 “being eighteen” that determines voter behavior. Instead, what is relevant to politics is how the biological fact of age is interpreted by various different individuals and groups. That is to say, the interpreted significance of age is crucial to its role within political explanation. Of course, naturalist political scientists might respond to this line of critique by reifying the concept of age in an atomistic way instead of reifying it as a brute fact. In other words, they might adopt the second kind of reification discussed above in place of the first. Something like this is what Rosenstone and Hansen actually do when they consider various possible hypotheses for how aging determines voter turnout. After considering three possible hypotheses, Rosenstone and Hansen settle on treating the meaning of age on the “life-experience hypothesis,” which holds that “as people grow older … they accumulate information, skills, and attachments that help them to overcome the costs of political involvement” (1993, p. 137). This shift in analysis marks a turn towards atomistic reification because instead of reifying age into a brute biological fact, stripped of meanings, age is instead reified as a freefloating meaning. That is, instead of an analysis of the particular beliefs and intentions of the agents involved, Rosenstone and Hansen posit the “life-experience hypothesis” which endows age with a single, free-standing meaning. This atomistic meaning is then treated as the cause of higher levels of voter participation. The anti-naturalist objection to this atomistic reification should now be clear. For according to anti-naturalists, beliefs and actions are the result of particular intentions and reasons held by agents. Yet Rosenstone and Hansen’s analysis gives the distinct impression that such intentions can be neglected in favor of positing an atomistic and quasi-universal meaning to growing older in America. Of course, none of this excludes the possibility that anti-naturalist 14 might accept Rosenstone and Hansen’s finding that older Americans voted in higher numbers between 1956 and 1988. It is only that anti -naturalists would insist that all the work of explanation was yet to be done. This is because anti-naturalists maintain that explanation requires a narrative about the actual reasons why particular agents formed certain beliefs, and yet this is precisely what they argue is elided by the naturalist correlation of reified concepts. ii. Essentialism Reification occurs when political scientists subtract the intentions of agents from political concepts. Essentialism, by contrast, is the result of stripping political concepts of their contextual and historical specificity. Freed of this specificity, essentialist concepts endeavor to express historical and cultural constants, which in turn might be used as the basis for naturalist causal explanations. Generally speaking, essentialism in political science takes two forms—“strong” and “weak.” Where strong essentialism assumes a fixed core of common traits that must be present for the concept to apply, weak essentialism allows for degrees of variation. Again we will look briefly at each in light of contemporary political science research. Strong essentialism occurs whenever political scientists assume a fixed core of features in all cases where a concept might be legitimately applied (e.g., Sartori, 1984). This kind of essentialism is widespread in political science today, but we will limit ourselves to observing it in the study of “contentious politics” as conducted by Charles Tilly, Doug McAdam and Sidney Tarrow. This research is worthy of singling out not only because of its widespread influence but 15 also because the very field of “contentious politics” is framed in such a way that it presupposes a strong form of essentialism. The concept of contentious politics is strongly essentialist in that it is meant to capture the core features of otherwise heterogeneous political phenomena, including social movements, strikes, revolutions, nationalism, and democratization to name only a few. According to Tilly, McAdam, and Tarrow, what this seemingly disparate list of political phenomena share is a nucleus of fixed features that make them all cases of contentious politics. These core features include that they are “episodic,” “public,” and involve collective interaction between two claimants at least one of which is a government (2001, p. 5). Contentious politics is in this way defined using a logic of commonality, which allows political scientists to include or exclude particular cases on the basis of whether or not they exhibit the core features. And while McAdam, Tarrow, and Tilly express skepticism at the possibility of general laws, they nevertheless preserve the naturalist hope that identifying the essential properties of political contention will help them discover the recurring causal sequences behind political phenomena (2001, pp. 345-346). Take, for example, the criteria that all forms of political contention be episodic. True to the logic of strong essentialism, this core feature is used by Tilly and company to exclude all political phenomena that occurs with regularity. Thus, for instance, recurring actions like elections, parliamentary debates, and associational meetings are all placed outside the bounds of the concept of contentious political action. A similarly essentialist logic informs the criteria that politically contentious politics be public. In this case, what is excluded are any actions occurring inside “well-bounded organizations, including churches and firms” (McAdam, Tarrow, and Tilly, 16 2001, p. 5). And a similar point could be made about these political scientists’ criteria that all contentious politics involve two claimants, at least one of which is a government. Readers may have already guessed the anti-naturalist reasons for rejecting the strong essentialism upon which the study of contentious politics rests. Because anti-naturalists believe that actions are expressive of historically specific webs of belief, they are unlikely to accept a supposedly universal class of action like that of political contention that transcends historical and cultural meanings. Indeed, they would note that the attempt to posit such a universal class of action only comes at the expense of neglecting the historically contextual meanings that inform human beliefs. For anti-naturalists hold that human actions are expressive of contingent contexts of meaning, and therefore are not the product of some nucleus of essential features. Thus, once again, we arrive at the fundamental conceptual incompatibility between naturalist and antinaturalist approaches. Of course, political scientists of a naturalist persuasion might still respond to the above line of criticism by adopting a weak instead of strong form of essentialism. Indeed, weak essentialism in political science often reflects the attempt to rectify the excesses of strongly essentialism by allowing for variations of degree in terms of how present core features are in particular empirical cases (e.g., Collier and Mahon, 1993). So where strong essentialism struggles to account for the sheer nuance and complexity of political reality, weak essentialism attempts to allow for greater flexibility while still retaining the basic essentialist premise that human political life exhibits certain core features. Weak essentialism plays a central role in what is considered by many to be among the most solid scientific findings of modern political science—namely, democratic peace theory. 17 Democratic peace theory centers on the claim that two democratic regimes are dramatically less likely to wage war on one another than other pairs of regimes. In a highly influential article, the political scientists Bruce Russett and Zeev Maoz defended this thesis first by defining democratic versus authoritarian regimes using weak essentialist concepts. Specifically, Russett and Maoz adopted a numeric scale to create a continuum of how democratic versus authoritarian a given regime might be in reality. On the basis of certain core criteria, like competitiveness of political participation and constraints on the chief executive, Russett and Maoz devised a scoring system in which minus 100 represented the most authoritarian regime and plus 100 the most democratic (with plus 30 marking the lower threshold for classing regimes as democratic, and minus 25 the upper threshold for marking them as authoritarian). As Russett and Maoz put it, such a scale allowed that in empirical reality “a state can have mixed characteristics” where “some features may be democratic at the same time that others are highly autocratic” (1993, p. 628). Yet the concepts of democratic and authoritarian regimes were also essentialist in that they were defined by core features considered to migrate from context to context—an approach Russett has since repeated (Russett and Oneal, 2001, p. 45). In short, the numeric scale allowed for the complexity and nuance of actual empirical cases without giving up on the basic essentialist logic. Yet it is precisely the retention of this logic that renders weak essentialism vulnerable to the same anti-naturalist line of critique. Namely, anti-naturalists would argue that once again what drops out of weak essentialism is the particularity of historical context. But, anti-naturalists would note, democracy and authoritarianism are not natural types with a recurring nuclei of features. Rather, both democracy and authoritarianism are the result of historically contingent meanings expressed in the beliefs, actions, and practices of particular societies. In this way, 18 political regimes are historically specific forms of meaning and not natural types. Interpreting the meanings of particular regimes is thus much closer to studying literary or legal texts than to, say, breaking down complex chemicals into the periodic table of elements. The anti-naturalist critique of both reification and essentialism brings us to a further question. Namely, what exactly are the anti-naturalist alternatives to concept formation? For while we have seen that anti-naturalists believe explanation in the social sciences requires narratives, it still isn’t clear what sorts of concepts such a narrative approach to the study of politics implies. As we will now show, anti-naturalists are not entirely of one mind when it comes to concept formation. We will conclude by surveying the two most important philosophical schools of anti-naturalist thought. 3. Varieties of Anti-naturalism Varieties of anti-naturalism share a rejection of naturalist concepts as reified and essentialist. Likewise, they share certain philosophical affinities—such as holism and contextualism—that lead them to deny the naturalist quest for causal laws built on atomistic units of political reality. In this regard, there is considerable philosophical agreement among varieties of anti-naturalism, which comprise their basic break from naturalism. However, these similarities withstanding, important divisions nevertheless exist within the anti-naturalist response to the naturalist challenge. We will survey two particularly influential schools of anti-naturalist thought: poststructuralism and hermeneutics. Post-structuralism is part of a wider, post-modern turn which 19 encompasses theorists as diverse as Jacques Derrida, Jean-François Lyotard, and Jacques Lacan. The sheer heterogeneity of this school means that we will have a better shot of understanding if we restrict ourselves to one particularly influential strain. Foucault is probably the most impressive and influential advocate of post-structuralism to date. As we shall see, Foucault’s post-structural concepts are marked by hostility towards the human subject as well as quasistructural modes of explanation. By contrast, hermeneutics not only rejects Foucault’s poststructuralism, but also advances rival concepts that are more humanistic and historicist. Modern defenders of hermeneutics have included Heidegger, Gadamer, and Paul Ricoeur. But we will draw on the work of Taylor in particular, whose influence on these debates has grown significantly in recent years. Picking the state of the art in these debates is especially important because early varieties of phenomenology were not always free of naturalist commitments (e.g., Schutz, 1962). i. Post-Structuralism Post-structuralism is marked by hostility to the human subject. This hostility is clearly evident, for example, in many of Foucault’s central concepts, which deemphasize the power of individual agency. For instance, during his archaeological period, Foucault often argued that individual human beliefs and actions were determined by fixed structures of meaning called discourses or epistemes. As Foucault described it, the rules structuring a discourse or episteme did not operate “in the mind or consciousness of individuals, but in discourse itself,” limiting “all individuals who undertake to speak in this discursive field” (1972, p. 63). According to Foucault this meant social science explanations should focus neither on “an individual, nor … some kind 20 of collective consciousness,” but on “an anonymous field whose configuration defines the possible position of speaking subjects” (1972, p. 122). In other words, the concepts of discourse and episteme pointed towards a social science in which explanation consisted of synchronically locating individual beliefs and actions within the appropriate structures of meaning. Thus, unlike the law-like explanations of naturalists, Foucault instead evoked synchronic structures of meaning that purported to determine belief and action within their field. In this way, Foucault’s archaeological mode of explanation was both anti-humanist and neglectful of historical context. It was anti-humanist because it sought to eliminate the role of creative, deliberative agents in favor of the power of impersonal quasi-structures. It was neglectful of historical context because rather than attending to the particular webs of belief informing an individual’s actions, Foucault posited frameworks of meaning that transcended the particularities of individual beliefs. The latter critique, in any case, was what drove the rejection of post-structuralism by hermeneutic thinkers like Taylor. Taylor noted that the reduction of individual beliefs and actions to concepts like discourse, episteme, or regime of power rendered actual historical change very difficult to explain. Was Foucault suggesting that the disembodied discourses themselves exhibited a kind of agency? That is, did post-structuralists believe epistemes and discourses actually changed and evolved historically, independent of the actual individuals involved? If so, Taylor argued, it was not at all clear how such an impersonal “purposefulness without purpose” worked (1985a, pp. 169-170). For his part, Foucault tried to account for this type of hermeneutic criticism by arguing that discourses, epistemes, and regimes of power, were all characterized by their own kinds of 21 volatility and open-endedness. So he wrote that within a given regime of power there was constant “tension” and “activity” such that a regime was never fully fixed or realized (Foucault, 1975, pp. 26-27). But unfortunately this response remained vulnerable to Taylor’s objections. For even if one allowed that the concepts of discourse, episteme, and regime of power are inherently volatile, this does not resolve what sort of agency such entities have apart from that introduced by individuals. Indeed, in the final years of his life, Foucault himself appeared to increasingly appreciate the importance of this criticism and the need for some role for human agency in the study of social and political life (Foucault, 1983, p. 225). Post-structuralism therefore faces the dilemma that it must adopt a more humanistic view of agency if it is to explain historical change. Yet the latter would also require that it drop or significantly remake its distinctive conceptual vocabulary of discourses, epistemes, and regimes of power. That is, resolving this dilemma might mean abandoning some of the more original elements of the post-structural project. Regardless of how post-structuralists respond to this problem going forward, an alternative that allows for individual human agency is presently offered by hermeneutics. ii. Hermeneutics Hermeneutics conceives of humans as the creative source of the beliefs, actions, and practices that comprise political reality. Taylor (1985c), for instance, has argued that human beings are “self-interpreting animals” able to creatively articulate, modify, and change their beliefs. The fact that humans are self-interpreting means they cannot be reduced to explanation 22 by impersonal laws or quasi-structures. But neither does it mean that individuals are entirely autonomous when it comes to belief and action. Rather, hermeneutic theorists insist that individuals exhibit a kind of situated agency, in which they retain creative capacities while nevertheless existing against the backdrop of some tradition of inherited beliefs and practices. Thus the concept of tradition (unlike post-structural notions of episteme or power) allows for individual agency without denying the active role of a larger social inheritance (Gadamer, 1960, II.1; Taylor, 2011, p. 34). According to hermeneutic theorists, political scientists can explain individual beliefs and actions by referring both to traditions and the ability of agents to reason creatively in light of dilemmas. Traditions can be used by students of politics to explain continuities in belief and practice. So, political scientists might explain the beliefs and actions of individuals by arguing that they inherited them from a particular tradition. Dilemmas, by contrast, can help explain changes in belief. Because humans hold beliefs for particular reasons, those beliefs can change whenever the supporting beliefs are invalidated or in some other way challenged. Dilemmas, in this regard, are any beliefs that create pressure on an existing web of beliefs. Political scientists can thus employ dilemmas and traditions in tandem to explain how individuals creatively respond within a world of shared meanings, beliefs, and practices. It is important to add that the concepts of dilemma and tradition do not commit hermeneutic theorists to the view that all actions and beliefs are rational or must be taken at face value. To the contrary, according to Taylor, hermeneutic explanation is by its very nature critical as “there can’t be explanation without a judgment of truth or validity” (Taylor, 1995, p. 153). The reason for this is that if an agent offers reasons for acting that appear deluded or otherwise 23 distorted, then a hermeneutic theorist must search for an alternative explanation for that agent’s actions. Such alternative explanations might include evoking pathologies (e.g., disease or injury) but it might also require identifying cases of repression or self-deception. Hermeneutic theorists conceptualize such acts of self-deception in terms of unacknowledged or hidden motives. Such a hidden motive might consist of any repressed belief, desire, or need that leads an individual to form a web of inauthentic or rationalizing beliefs (e.g., Taylor, 1985d, pp. 4-6). A brief example will help illustrate how this works from a hermeneutic perspective. Consider the case of Nazi, “Aryan physics,” which was advanced by otherwise worldclass physicists like Philipp Lenard and Johannes Stark. Presumably some distortion of rational belief was going on in this case. Most likely, the development of Aryan physics by these two famous scientists was driven by an irrational unwillingness to accept the contributions of Jewish physicists like Einstein. So, where Lenard and Stark might have thought Aryan physics was simply another way of saying “good” or “proper” physics, a hermeneutic student of politics might argue that their beliefs and actions are instead explained by an irrational form of ideological distortion. Hermeneutics thus appears to offer an approach to the study of politics that is equally humanistic and critical. And yet this school of thought is often criticized for being unable to move beyond small-scale analysis of individual beliefs in order to explain politics at a general level (Box-Steffensmeier, Brady, and Collier, 2008). However, this latter criticism represents a basic misunderstanding of the hermeneutic position. For as we noted above (in part one), antinaturalist approaches are not tied to any one level of analysis. Taylor (2007), for one, has used a hermeneutic approach to offer an account of political life at the most general level—narrating 24 nothing less than the civilizational shift toward secularism in the modern era. Thus it would appear that in addition to being humanistic and critical, hermeneutic approaches are not limited to small-scale research. There is, in this sense, a certain playfulness in hermeneutic approaches. For, unlike naturalists, hermeneutic theorists need not privilege a single level of description. Rather, because they believe there are multiple true ways of describing human actions, hermeneutic students of politics don’t have to treat political reality as though it were reducible to a single level of truly scientific description (Davidson, 2006, pp. 24-25). This, of course, runs against the tendency we observed in naturalist political scientists like Tilly and Russett. For where the latter presuppose some privileged, essential core to, say, a politics of contention or democratic peace, hermeneutics insists that researchers can instead pragmatically choose their level of description given the wider purposes of their project. In short, hermeneutic theorists believe there are multiple true stories that can be told about political reality, not just one scientifically official account. This means the hermeneutic study of politics, like the analysis of a literary or legal text, can legitimately take on various forms depending on the purposes of the interpreter. This once again returns us to some of the fundamental differences between hermeneutics and the naturalist presuppositions informing much of political science. For, the entire logic of naturalism presses towards the discovery of a universal, invariant language of explanation, while hermeneutics seeks historically contingent stories. 25 Conclusion We have tried to show how the basic ruptures between naturalism and anti-naturalism are philosophical and not methodological. One such rupture can be seen in the naturalist tendency to treat methods as logics of discovery versus the anti-naturalist preference for treating methods as the techniques of a craft. Another set of ruptures is evident in naturalism’s inclination towards atomism and causal explanations versus anti-naturalism’s championing of holism and narrative explanations. We also saw that fundamental differences of philosophy are what fuel the anti-naturalist critique and rejection of naturalist concepts in political science. So, anti-naturalists draw on their philosophical assumptions in order to argue that naturalist political science is distorted by reification and essentialism. In place of naturalism, anti-naturalists have offered a number of philosophic alternatives. Two of the most important are post-structuralism and hermeneutics. Post-structuralism advocates an anti-humanist conception of political science, while hermeneutics favors a humanistic vision of the study of political life. Both reject naturalism. Because the conflict between naturalism and anti-naturalism is ultimately about philosophy and not methods, methodological pluralism can and should reign within political science research. None of this serves to keep the peace between these two battling factions. For in the naturalism/anti-naturalism debate the price of suing for a false peace is conceptual confusion and muddle—and the latter, whatever the outcome of future exchanges, is something no side should desire. 26 Reference list Almond, G. and Verba, S. (1965) The Civic Culture. Boston: Little, Brown and Company. Ayer, A. (1952) Language, Truth, and Logic, 2nd edn. New York: Dover Publications. Box-Steffensmeier, J., Brady, H., and Collier, D. (2008) ‘Political Science Methodology’, in Box-Steffensmeier, J., Brady, H., and Collier, D. (eds.) The Oxford Handbook of Political Methodology. Oxford: Oxford University Press, pp. 3-34. Carnap, R., Hahn, H., and Neurath, O. (1929) ‘Wissenschaftliche Weltauffassung: Der Wiener Kreis’, in Sarkear, S. (ed.) The Emergence of Logical Empiricism. Reprint, New York: Garland Publishing, 1996, pp. 321-340. Chih Lin, A. (1998) ‘Bridging Positivist and Interpretivist Approaches to Qualitative Methods’, Policy Studies Journal 21:1, pp. 162-180. Collier, D. and Mahon, J. (1993) ‘Conceptual “Stretching” Revisited: Adapting Categories in Comparative Analysis’, American Political Science Review 87(4), pp. 845-855. Collingwood, R. (1946) The Idea of History. Oxford: Clarendon Press. Davidson, D. (2006) ‘Actions, Reasons, and Causes’, in Lepore, E. and Ludwig, K. (eds) The Essential Davidson. Oxford: Clarendon Press, pp. 23-36. Dilthey, W. (1976) Selected Writings. Translated by H.P. Rickman. Cambridge: Cambridge University Press. Feitlowitz, M. (1998) A Lexicon of Terror: Argentina and the Legacies of Torture. Oxford: Oxford University Press. Foucault, M. (1983) ‘The Subject and Power’, in Dreyfus, H. and Rabinow, P. (eds.) Michel Foucault: Beyond Structuralism and Hermeneutics, 2nd edn. Chicago: University of Chicago Press, pp. 208-226. Foucault, M. (1975) Discipline and Punish. Translated by Alan Sheridan. Reprint, New York: Vintage Books, 1995. Foucault, M. (1972) The Archaeology of Knowledge. Translated by A. M. Sheridan Smith. Reprint, New York: Vintage Books, 2010. Gadamer, H. (1960) Truth and Method. Translated by Joel Weinscheimer and Donald G. Marshall. Reprint, New York: Continuum, 2004. Geertz, C. (1973) The Interpretation of Cultures. New York: Basic Books. 27 Gerring, J. (2001) Social Science Methodology: A Criterial Framework. Cambridge, UK: Cambridge University Press. Gerring, J. (2011) Social Science Methodology: A Unified Framework, 2nd edn. Cambridge: Cambridge University Press. Goertz, G. and Mahoney, J. (2012) A Tale of Two Cultures: Qualitative and Quantitative Research in the Social Sciences. Princeton: Princeton University Press. Goodin, R. (2009) ‘The State of the Discipline, the Discipline of the State’, in Goodin R (ed.) Oxford Handbook of Political Science. Oxford: Oxford University Press. Gunnell, J. (1975) Philosophy, Science and Political Inquiry. New Jersey: General Learning Press. Heidegger, M. (1953) Being and Time. Translated by Joan Stambaugh. Reprint, Albany: State University of New York Press, 1996. Hempel, C. and Oppenheim, P. (1948) ‘Studies in the Logic of Explanation’, Philosophy of Science 15, pp. 135-175. Hempel, C. (1965) ‘The Function of General Laws in History’, in Aspects of Scientific Explanation. New York: The Free Press, pp. 231-243. Husserl, E. (1936) ‘Vienna Lecture’ and ‘The Attitude of Natural Science and the Attitude of Humanistic Science’, in Carr, D. (ed.) The Crisis of European Sciences and Transcendental Phenomenology. Reprint, Evanston, IL: Northwestern University Press, 1970, pp. 269-300, 315334. Kernell, S., Jacobson, G., and Kousser, T. (eds.) (2011) The Logic of American Politics, 5th edn. Washington, DC: CQ Press. McAdam, D., Tarrow, S., and Tilly, C. (2001) Dynamics of Contention. Cambridge, UK: Cambridge University Press. MacIntyre, A. (1964) ‘A Mistake About Causality in Social Science’, in Laslett, P. and Runciman, W.G. (eds.) Philosophy, Politics and Society. Oxford: Basil Blackwell, pp. 48-70. MacIntyre, A. (1978) ‘Is a Science of Comparative Politics Possible?’, in Against the Self Images of the Age. Notre Dame, IN: University of Notre Dame Press, pp. 260-279. MacIntyre, A. (2007) After Virtue, 3rd edn. Notre Dame, IN: University of Notre Dame Press. Maoz, Z. and Russett, B. (1993) ‘Normative and Structural Causes of Democratic Peace, 19461986’, American Political Science Review 87(3), pp. 624-638. 28 Marx, K. (1859) ‘Preface to A Critique of Political Economy’, in McLellan, D. (ed.) Karl Marx: Selected Writings. Reprint, Oxford: Oxford University Press, 1977, pp. 388-391. Neurath, O. (1931) ‘Physicalism: The Philosophy of the Viennese Circle’, in Cohen, R. and Neurath, M. (eds) Philosophical Papers: 1913-1946. Reprint, Boston: D. Reidel Publishing Company, 1983, pp. 52-57. Ricoeur, P. (1983-1985) Time and Narrative, 3 vols. Translated by Kathleen McLaughlin and David Pellauer. Reprint, Chicago: University of Chicago Press, 1984-1988. Rosenstone, S. and Hansen, J. (1993) Mobilization, Participation, and Democracy in America. New York: Macmillan. Russett, B. and Oneal, J. (2001) Triangulating Peace. New York: W.W. Norton & Company. Sartori, G. (ed.) (1984) Social Science Concepts: A Systematic Analysis. Beverly Hills, CA: Sage. Schutz, A. (1962) ‘Concept and Theory Formation in the Social Sciences’, in Emmet, D. and MacIntyre, A. (eds) Sociological Theory and Philosophical Analysis. Reprint, New York: Macmillan Company, 1970, pp. 1-19. Taylor, C. (1985a) ‘Foucault on Freedom and Truth’ in Philosophy and the Human Sciences: Philosophical Papers 2. Cambridge, UK: Cambridge University Press, pp. 152-184. Taylor, C. (1985b) ‘Interpretation and the Sciences of Man’, in Philosophy and the Human Sciences: Philosophical Papers 2. Cambridge, UK: Cambridge University Press, pp. 15-57. Taylor, C. (1985c) ‘Self-Interpreting Animals’, in Human Agency and Language: Philosophical Papers 1. Cambridge: Cambridge University Press, pp. 45-76. Taylor, C. (1985d) ‘Introduction’, in Human Agency and Language: Philosophical Papers 1. Cambridge: Cambridge University Press, pp. 1-12. Taylor, C. (1995) ‘Comparison, History, Truth’, in Philosophical Arguments. Cambridge, MA: Harvard University Press, pp. 146-164. Taylor, C. (2007) A Secular Age. Cambridge, MA: Belknap Press of Harvard University. Taylor, C. (2011) ‘Understanding the Other: A Gadamerian View on Conceptual Schemes’, in Dilemmas and Connections. Cambridge, MA: Belknap Press of Harvard University Press, pp. 24-38. Winch, P. (1958) The Idea of a Social Science. London: Routledge and Kegan Paul. 29