Commanding a Clear View: Words, Concepts, and Social Science

Patrick Thaddeus Jackson

American University

[**Original document page numbers preserved in brackets for citation purposes.**]

[Start Page 48]

Patrick Thaddeus Jackson is Professor of International Relations and Associate Dean for Curriculum and Learning in the School of International Service at American University. He can be reached at .

To lay my cards on the table at the outset: I am broadly sympathetic to Frederic Schaffer’s overall campaign in favor of conceptual elucidation: “investigating the ways in which the social world is built up linguistically and the ways in which social actors deploy concepts to pursue their goals.”1 On numerous previous occasions I have been, like Schaffer, decidedly critical of scholarly efforts to “fix” the meaning of a concept (like the West or civilization) and then to use that scholarly reconstruction as a base from which to legislate appropriate and inappropriate practical claims using that concept—as though our task as scholars were to correct the social world rather than to explain and understand it. So Schaffer’s careful explication of techniques for elucidation, grouped under the headings of “grounding,” “locating,” and “exposing,” provides a refreshing alternative to the sort of advice about concept analysis one typically receives from scholars engaged in the kind of project I think rather problematic.

That said, in my view Schaffer’s book also illustrates— practically and performatively if not deliberately—an important liability of his approach to concepts. The version of “interpretivism” that emerges from his account, while grounded in how people in the field conventionally use the word, obscures rather than clarifies important philosophical distinctions between theory, methodology, and method, and shores up philosophically misleading but practically operative dichotomies opposing “interpretivism” to “positivism” as if those were coherent intellectual packages. Despite conventional use, I do not believe that “positivism” and “interpretivism” name such coherent packages. For that reason I do not believe that the only alternative to reconstructing a concept so that it can be inserted into a statistical study as an independent or dependent variable—which is the so-called “positivist” strategy—necessarily means taking on all three of Schaffer’s “sets of interpretivist questions”2 about ranges of meaning, linguistic and historical specificity, and political context. By adhering

[Start Page 49]

too closely to contemporary scholarly convention, Schaffer reproduces a simplified picture of our choices, and tacitly insists that we choose a side. Despite his admirable caveat about the book being “more akin to a collection of recipes written for an adventurous cook,”3 the overall picture he paints is still dichotomous.

The difficulties begin on the first page of the book, where Schaffer introduces positivism and interpretivism as “two different conceptions of social science or, more precisely, two different methodologies that the scholar may bring to the study of the social world.”4 He takes positivism to be a stance about the mind-independent existence of entities in the social world (which we might call a mind-world dualism); this gives rise to the task of formulating propositions “about those entities based upon the identification and measurement of regularities within and between them.” Interpretivism, by contrast, maintains that “there are no ‘real’ social entities, only culturally mediated social facts,” and that this stance, which we might call mind-world monism, means that the scholarly task is to “shed light on how shared meanings and their relation to power inform or structure the social world and the study of the social world.”5 There is a lot crammed into this distinction: philosophical ontology (also known as “methodology”)6 concerning the hook-up between the mind and the world, scientific ontology (also known as “theory”) concerning the character of social entities and the nature of social action, and a series of concrete research tasks that supposedly follow or flow from those positions. Theory and methodology are also linked, although the links are not made especially clear; we aren’t told precisely why a commitment to mind-world dualism on the part of the researcher both implies a task of looking for nomothetic empirical generalizations and directs researchers away from investigating the role of shared meanings.7 Nor are we told why an appreciation of “culturally mediated social facts” means that researchers have to prioritize the role of shared meanings in their scholarly investigations.

In my view the ambiguity here arises from the conflation of philosophically distinct registers with the unfortunate consequence of disciplining thinking space into a binary either/or choice between “positivism” or “interpretivism.” The two registers of theory and methodology speak to different aspects of scholarly claims; the former concerns the content of a claim about the world, while the latter concerns the status of that claim. And they can in principle be combined in a variety of different ways, despite the fact that we conventionally reproduce only certain combinations in our existing research practice. There is no philosophical problem at all involved in taking the theoretical position that shared meaning shapes the world, and simultaneously adopting a research strategy intended to disclose systematic cross-case correlations between variations in shared meaning and variations in social arrangements or outcomes. Similarly, there is no philosophical problem at all involved in taking the methodological position that scholarly knowledge is not a representation of a putatively mind-independent reality but is instead tightly co-constitutive of that reality, and simultaneously adopting a theoretical perspective that focuses on social structures or arrangements of practical activity rather than on shared meaning.

What Schaffer identifies as “positivism” and “interpretivism” are combinations of theory and methodology that are not exhaustive of the combinatorial possibilities—and this matters precisely because Schaffer’s discussion presumes, albeit tacitly, that they are exhaustive. One is either a “positivist” or an “interpretivist,” and the whole of each package has to be accepted at the same time. His recommendation of three aspects of conceptual elucidation does not entertain the possibility that a scholar might not engage in all three aspects, and I submit that Schaffer cannot entertain that possibility be- cause the menu of choices has been framed too dichotomously.

To elucidate a concept, in Schaffer’s account, requires first grounding that concept in experience-near language (“commonplace words used in everyday contexts”8) by looking at how ordinary people use terms in their everyday lives, and then locating a term in its specific context to prevent a mistranslation of the relevant conceptual category into something in our vocabulary as though it seamlessly fit there. Grounding and locating terms in this way provides knowledge of concepts precisely because “a concept is constituted by the different ways in which a word is used,”9 and the scholar’s role is to call attention to that diversity rather than to identify commonalities across uses and contexts.

Why? Schaffer provides two broad answers, one involving criticizing concepts that “have become stabilized, naturalized, or neutralized in ways that obscure from view their histories of contingency and contestation,”10 and the other involving an effort to make certain that scholarly concepts correspond to the concepts in use among the people being studied: “we would need to know who each voter considers to be a member of his or her family in the context of ballot casting” in order to determine whether family members influence one another’s votes.11 Both of these ends reflect a skepticism about universals and a desire to provincialize scholarly discourse: if scholarly concepts are just one set of conventional uses of terms, there is no compelling reason to prefer the scholarly concepts over those operating in the contexts under investigation.

[Start Page 50]

We are thus enjoined to look at language-in-use in our explanatory theories, presumably because to do otherwise would be to commit the same sin that Sartori commits when he advances an abstract concept of the family designed to travel across contexts: “he reifies a particular and partial conception of what a family is” and thus “risks guiding in unexamined ways the thinking of anyone who adopts such a definition.”12 Sticking close to how people themselves use terms is the antidote to such a scholarly imposition.

But all of this makes very specific assumptions about the purpose and potential of scholarship. Knowledge about how a group of people conventionally uses a term might suffice as an explanation of an outcome if we had a theoretical reason for connecting conventional uses and outcomes—but such a connection is not an inevitable consequence of the close examination of conventional use per se. But the connection between conventional uses and outcomes is a consequence of the rejection of “positivism” understood as a more or less behaviorist quest for systematic correlations. The implicit argument seems to be that if we are to take people’s conventional uses of terms—their concepts—seriously, then we have no choice but to restrict scholarly explanation to an exercise of explication: an outcome or arrangement is explained when we disclose the concepts in terms of which that outcome is comprehensible to the people involved. In such a conception, operational concepts can’t be wrong, and people can’t be mistaken. It is difficult to see how we might explain a situation as produced by something people were doing without meaning to do it. People also cannot be unaware of the reasons for which they do things, affected by factors that lie outside of their conceptual universes. It is also unclear how a scholar might critique concepts, whether her own or those of the people under investigation, except by bringing to light their false universalism; there does not seem to be much space here for anything like a better concept.

Many of these criticisms likely sound like the sort of thing a “positivist” might deploy against an “interpretivist,” since in Schaffer’s account those are the only alternatives. I do not agree. We are not faced with a dichotomous choice between a mind-world dualist strategy of producing universal concepts that populate nomothetic generalizations and a mind-world monist strategy of explicating diverse patterns of the conventional use of terms that results in enhanced awareness of the locality of meaning and a consequent privileging of the experience-near language of participants in our scholarly accounts. We are instead faced with a series of choices involving different philosophical registers, and a plethora of ways of making good use of Schaffer’s excellent technical advice about a variety of methods of grounding and locating concepts. One can act on Schaffer’s advice and pay attention to operational concepts understood as terms in conventional everyday use for a number of different reasons:

In the theoretical register, we might broadly distinguish13 between experience-near theories that focus on shared meanings as an explanatory factor, and social-relational theories that focus on patterns of transaction and the positions of actors relative to one another. The former would engage in grounding and locating in order to make sense of the cultural lifeworld of the actors being investigated, aiming to disclose the situationally-specific rules of, for example, how competence is defined and negotiated in a diplomatic setting14 or what makes for an acceptable public statement about foreign policy15 or an acceptable course of action regarding sovereign prerogatives.16 The latter would engage in grounding and locating in order to track how the deployment and invocation of particular concepts generate outcomes, or how available concepts are shaped by networks and specific histories; here the emphasis is on sketching transposable mechanisms like brokerage17 or legitimation18 and investigating how they play out in specific cases. In a way, the former is more interested in the content of a concept, while the latter privileges the form, either of the overall conceptual space or of particular types of rhetorical deployment, and while the former aims to produce “thick” locally specific knowledge, the latter aims to refine “thin” ideal-typical mechanisms and processes that can inform singular causal accounts.19

In the methodological register, we might broadly distinguish20 between at least four different ways of producing social-scientific knowledge: subsuming outcomes under general laws, identifying dispositional causal capacities, elaborating ideal-typical models, and reflexively grounding claims in the social position of the social scientist herself. Each of these four approaches would have a use for concepts understood as experience-near and grounded in conventional everyday usage, whether that was coding an operative meaning variable, abductively inferring a condition of possibility for a way of using terms, connecting the terms of a model to the local cultural context of the scholar herself, or dispelling the whitewashed universality of assuming that every place and every social situation is fundamentally identical and thus seamlessly translatable into one’s own default native language.

Also operating in the methodological register, we might broadly distinguish between descriptive knowledge intended to communicate facts, causal knowledge intended to convey the skills and capacities involved in making something happen, and a kind of knowledge intended to equip the recipient to go on appropriately—a type of knowledge I would personally prefer to call “interpretive,” with full awareness that this is a considerably narrower use of the term than Schaffer might want. As with the four social-scientific methodologies mentioned

[Start Page 51]

in the previous paragraph, Schaffer’s techniques for grounding and locating concepts can contribute to any of these three kinds of knowledge: we could engage in locating and grounding in order to answer questions about what some group of people mean by some term (descriptive knowledge), questions about which terms we might use ourselves in order to provoke a particular response (causal knowledge), or questions about how to use terms properly as a member of the community in question (in my lexicon if not in Schaffer’s, interpretive knowledge).21

By breaking apart Schaffer’s “interpretivism” and “positivism” into a series of commitments in different philosophical registers, we produce more thinking space, more possible combinations of substantive claims about the world and ways of evaluating those claims. We also disrupt the false dichotomization that would invariably connect examinations of how terms are used in practice with particular theories and methodologies, to the exclusion of other combinations. There is nothing whatsoever about studying how people use terms that commits us to any particular kind of theory or flavor of methodology. Looking at the everyday use of terms in order to ground and locate concepts does not lock us into theories that connect shared meaning to outcomes or obligate us to eschew causal explanations in favor of something like “understanding.” The fact that many self-proclaimed “interpretivists” do in fact combine a focus on everyday language with a rejection of causation does not mean that the combination itself forms some kind of seamless and compelling logical whole.

Indeed, moving beyond the “positivist”/”interpretivist” dichotomy makes it possible to articulate an approach to conceptual analysis that avoids the problems of objectivism, one- sidedness, and false universalism, which Schaffer rightly criticizes,22 but also serves causal-explanatory purposes. Such an approach might involve using the precise grounding and locating of concepts not so much to “destabilize” putatively “timeless, essential properties”23 as to show the relatively stable patterns that emerge in the course of actual lived experience. Ideal-typifying such patterns would allow us to investigate how situated social actors deploy terms and the concepts they carry with them, and how such deployments produce outcomes: through processes of legitimation, boundary-creation and boundary-maintenance, and the foreclosing of otherwise-possible courses of action. Proceeding in this way means that the patterns of use that figure into the resulting causal account are only analytical instruments; their apparent “universality” is only a logical generality, and not any kind of a categorical pronouncement about the essence of a concept or its limitless empirical applicability. And although an ideal- typified notion is by definition24 “one-sided,” this is what we might call a tactical one-sidedness for the purpose of explaining a specific outcome: what is extracted from lived experience are the broad outlines of how the group being investigated actually uses the relevant terms, and this is done not in order to elucidate one or another dimension of how a concept might be defined, but to explain how this outcome arises from this conventional use.

Ideal-typification thus provides a potential response to two of the problems Schaffer identifies: universality and one-sidedness. As for “objectivism,” another problem he identifies, Schaffer worries that treating concepts in anything but an “interpretivist” way means that scholarly treatments of a concept present “a seemingly value-free, objective definition that appears to be generated by a purely analytic set of operations (identifying essential properties, etc.) and detached from a broader context of political contestation.”25 This might be the case if we were talking about scholarship that intended to define a concept, and thus to tell audiences both inside and outside of the academy precisely what something means. But I very much doubt that this is the intent of much social-scientific scholarship, and I would argue that it shouldn’t be our intent when we are operating as social scientists or as any other kind of scientist. Our job—our vocation—when operating in a scientific idiom is to produce well-reasoned conclusions about matters of fact: descriptions and explanations that are systematic, public, and worldly insofar as they bracket the realms of the divine and the transcendent in order to focus on producing “disenchanted” accounts. 26 So I, as a social scientist, should never be pronouncing judgment on what, for example, “democracy” is; rather, I should be using a precisely defined notion of “democracy” in my descriptions and explanations, and explicitly refraining from using that definition to legislate appropriate and inappropriate use by others. In that way, a better awareness of the limits of the social-scientific enterprise functions to disrupt “objectivism,” and in so doing can address Schaffer’s concerns.

Of course, I am very aware that the philosophical limits of social-scientific inquiry are often blurred in practice, as political and social actors either look to scholarly definitions as guidelines for their own thinking, or deploy scholarship as part of a campaign to produce one or another practical outcome. Sometimes we scholars even aid and abet that process. Regardless, my point is that the problems of “positivism” that Schaffer identifies are not, in my view, philosophical problems so much as they are institutional and political problems. I think that Schaffer is quite correct that we can start to address these problems by bringing our technical language “back to the rough ground,” as Wittgenstein might put it,27 and regarding it as nothing other than ordinary language in a specific context.28 

[Start Page 52]

But I disagree that the only important contribution that grounding and locating scholarly concepts can make is to re- activate histories of contestation and make us aware that our scholarly notions do not drop intact from some rationalist heaven. Even if some of us need to be reminded of that from time to time, I very much doubt that pointing to past and ongoing social contestation about the meaning of a concept is going to do the trick. In my view, reflexive critique of our own scholarly concepts should start by clarifying how we as social scientists are supposed to use concepts—and that in turn comes from (to invoke Wittgenstein again) an investigation of our own “form of life” intended not to explain outcomes, but to normatively prescribe ways of appropriately “going on.” If we did that, in my view, we would be in much better shape to contribute to ongoing conversations in a complex and turbulent world.


Adler-Nissen, Rebecca. 2015. Opting Out of the European Union: Diplomacy, Sovereignty and European Integration. Reprint edition. Cambridge University Press.

Goddard, Stacie E. and Ronald R. Krebs. 2015. “Rhetoric, Legitimation, and Grand Strategy.” Security Studies vol. 24, no. 1: 5–36. Jackson, Patrick Thaddeus. 2006. Civilizing the Enemy: German Re-

construction and the Invention of the West. Ann Arbor: University of Michigan Press.

Jackson, Patrick Thaddeus. 2016. The Conduct of Inquiry in International Relations. 2nd ed. London: Routledge.

Jackson, Patrick Thaddeus and Daniel H. Nexon. 2013. “International Theory in a Post-Paradigmatic Era: From Substantive Wagers to Scientific Ontologies.” European Journal of International Relations vol. 19, no. 3: 543–565.

McCourt, David M. 2016. “Practice Theory and Relationalism as the New Constructivism.” International Studies Quarterly vol. 60, no. 3: 475–485.

Neumann, Iver B. 2012. At Home with the Diplomats: Inside a European Foreign Ministry. Cornell University Press.

Nexon, Daniel H. 2009. The Struggle for Power in Early Modern Europe: Religious Conflict, Dynastic Empires, and International Change. Princeton: Princeton University Press.

Patomäki, Heikki and Colin Wight. 2000. “After Postpositivism? The Promises of Critical Realism.” International Studies Quarterly vol. 44, no. 2: 213–237.

Pouliot, Vincent. 2016. International Pecking Orders: The Politics and Practice of Multilateral Diplomacy. Cambridge University Press.

Schaffer, Frederic Charles. 2016. Elucidating Social Science Concepts: An Interpretivist Guide. New York: Routledge.

Tilly, Charles. 1998. “International Communities, Secure or Otherwise.” In Security Communities, edited by Emanuel Adler and Michael Barnett. Cambridge: Cambridge University Press, 397– 412.

Weber, Max. 1999. “Die ‘Objektivität’ Sozialwissenschaftlicher Und Sozialpolitischer Erkenntnis.” In Gesammelte Aufsätze Zur Wissenschaftslehre, edited by Elizabeth Flitner. Potsdam: Internet-Ausgabe, 146–214. Online at:

494/html/ (last accessed 13 February 2017).

Weber, Max. 2004. The Vocation Lectures. Indianapolis: Hackett Press. Wittgenstein, Ludwig. 1953. Philosophical Investigations. Oxford:



1 Schaffer 2016, 7.

2 Schaffer 2016, 21.

3 Schaffer 2016, xv.

4 Schaffer 2016, 1.

5 Schaffer 2016, 2. Note that Schaffer does not use the terms “mind-world dualism” or “mind-world monism.” Nor does he distinguish between philosophical and scientific ontology. But I am not quibbling about terms here. Regardless of which terms one uses, I am arguing that these distinctions need to be a core part of our methodological ruminations.

6 On this expanded use of the term “methodology,” see Jackson 2016.

7 After all, mind-world dualism can also lead to a critical realist disclosure of real-but-undetectable causal powers, which is a far cry from a search for nomothetic generalizations. On this alternative, see Patomäki and Wight (2000).

8 Schaffer 2016, 2.

9 Schaffer 2016, 74.

10 Schaffer 2016, 83.

11 Schaffer 2016, 16.

12 Schaffer 2016, 14.

13 Here I invoke a distinction explored in Jackson and Nexon (2013).

14 Pouliot 2016.

15 Neumann 2012.

16 Adler-Nissen 2015.

17 Tilly 1998; Nexon 2009.

18 Jackson 2006; Goddard and Krebs 2015.

19 In actual scholarly practice, this is not a hard and fast distinction, and “practice turn” and “processual/relational” accounts frequently incorporate both types of theorizing—as did a strain of “constructivist” theorizing that differed quite significantly from the U.S.-mainstream version (McCourt 2016).

20 As I do in Jackson 2016.

21 Knowledge that criticizes particular concepts is also interpretive knowledge in the sense I am using the term here: our use of our concepts to critique what some other group is doing is an interpretation of their actions through the lens of our concepts and thus tell us how to go on in our community. Our use of our concepts to clarify what we ourselves ought to be doing is an interpretation of our own context that tells us how to go on in our community (as in Wittgenstein 1953). All of these are variants on the problem of how to use terms properly as a member of a community.

22 Schaffer 2016, 13–20.

23 Schaffer 2016, 86.

24 By Weber’s 1999 definition.

25 Schaffer 2016, 19.

26 “Disenchanted” in Max Weber’s (2004, 14–19) sense.

27 Wittgenstein 1953, sec. 107.

28 Schaffer writes: “In their own arena of practice, social scientists’ language is ordinary language and can be analyzed as such” (Schaffer 2016, 32).