Thick Concepts and Sociological Research

I

One way to classify concepts goes like this. Some concepts are descriptive, like yellow, furniture, and Uruguayan. Other concepts are evaluative, like goodness, wrongness, and beauty. ¹ One way to classify evaluative concepts goes like this. Some evaluative concepts are thin, like goodness, wrongness, and beauty. Other evaluative concepts are thick, like dignity, piety, cruelty, humanness, generosity, sleaziness, dogmatism, ostentatiousness, and garishness. ²

What’s the difference between thin and thick concepts? Thin concepts like goodness, wrongness, beauty, and appropriateness do an evaluative job, whose valence can be positive (pro) or negative (con). One key property of thin concepts is that you can apply them to anything you’d like. According to your evaluations, Becky’s piano is good, Huck’s action was wrong, and Tom’s painting is beautiful. While you applied these thin predicates to these particular piano, action, and painting, you could have applied them to any other. For example, you could have said that the Steinway Model D isn’t a good piano. If so, you would have made a mistake about musical instruments’ quality, but not a conceptual or semantic one. You could have said that concentrated animal feeding is a morally good thing. If so, you would have made a moral mistake, but not a conceptual or semantic one.

Thick concepts like compassion, cruelty, humanness, and elegance do an evaluative job, whose valence can be positive (pro) or negative (con). But they also do a descriptive job. That a practice is compassionate, a ruler is cruel, and a chess sacrifice and combination are elegant tells you something about what these things are like. Therefore, one key property of thick concepts is that you can’t apply them to anything you’d like. If you say that Oskar Dirlewanger and Augusto Pinochet weren’t cruel people, you wouldn’t understand what cruelty is (or, you wouldn’t understand the meaning of the English word “cruel”). If you say that concentrated animal feeding is a compassionate and humane practice, you wouldn’t understand what compassion and humanness are (or, you wouldn’t understand the meaning of the English words “compassionate” and “humane”). In both cases you’d be mistaken semantically or conceptually. You’d be mistaken in the same sense you’d be mistaken if you said, “Strawberries are yellow” and “Messi is Uruguayan.” ³ In this sense, thick concepts compassion and cruelty are like descriptive concepts yellow and Uruguayan, but unlike thin concepts good and wrong.

To be sure, there are legitimate disagreements as to who and what is cruel (and who and what isn’t), and who and what is compassionate (and who and what isn’t). Still, if X didn’t have certain basic properties, it wouldn’t qualify as a candidate for being a cruel act. If Y didn’t have certain basic properties, it wouldn’t qualify as a candidate for being a compassionate act. It’s not the case that anything goes. You aren’t free to speak of cruelty and compassion however you wish.

One way to classify thick concepts goes like this. Some thick concepts are ethical, like dignity, rudeness, nobleness, and pettiness. Some thick concepts are aesthetic, like kitschiness, daintiness, balance, and grace. Some thick concepts are epistemic, like open-mindedness, superficiality, conscientiousness, and dogmatism. Some thick concepts can be classified into more than one of these categories, like creativity (epistemic and aesthetic), elegance (epistemic and aesthetic), reliability (epistemic and ethical), integrity (ethical and aesthetic), and harmony (ethical and aesthetic).

This is one possible classification of thick concepts. There are very many thick concepts. Very many. ⁴ There’s a great variety of them. And there’s an infinite number of ways to classify them. Here’s one example. The English name of some thick concepts begins with the letter “d,” like dignity and decency. The English name of some thick concepts begins with the letter “c,” like cruelty and compassion. That’d be useful for a dictionary of thick concepts. As usual, the value of a classification is a function of what you intend to do with it.

All these concepts and distinctions are surely up philosophers’, linguists’, and cognitive scientists’ alley. But what do thick concepts have to do with sociologists’ empirical research? Why should sociology and social theory care?

II

The philosophical literature on thick concepts, terms, predicates, and properties has two founding mothers: Philippa Foot (1958) and Iris Murdoch (1970). Influenced by them, Bernard Williams (1985) showed that moral philosophy shouldn’t neglect the thick/thin distinction. ⁵ In recent times, work on thickness has blossomed in ethics, as well as aesthetics and epistemology (Kotzee and Wanderer 2008; Roberts 2018). ⁶ I draw on these philosophers to advance a twofold argument. First, sociological research can deliver not only rich catalogs of thick concepts, but also insights into their nature and workings. These insights can be obtained neither from the armchair nor in the laboratory. Second, sociological research on thick concepts can bring into view important aspects of social life, which thin concepts don’t provide access to. It may help sociology think afresh how social (or human) things are unique and how the investigation of social (or human) things is unique. ⁷

Think of utterances and beliefs that contain thick concepts. You might say to your friend: “That was generous of you,” “Your country’s president is small-minded,” or “That’s such a kitschy song!” You might believe that the movie’s plot was creative, but the lead actor’s performance was affected. You might worry about brutal dictators and fanatical groups; workers’ being exploited, women’s being objectified, and minorities’ being discriminated against. Psychologists tend to study individual-level phenomena. They may experimentally study how individuals use thick concepts, and what they show about cognition, judgment, and other aspects of human psychology. Sociologists tend to study social-level phenomena. Their units of analysis tend to be groups, fields, and societies. I’m interested in what a sociological approach to thick concepts should look like. In fact, there’s already quite a bit of sociological research on thick concepts and their social effects, even if the expression “thick concepts” isn’t there. In the sociologies of culture and art. Of religion. Morality. Gender. Valuation and evaluation (Lamont 2012). In many urban ethnographies. However, sociologists haven’t recognized thick concepts as such. Their distinctiveness and distinctive potential haven’t been grasped.

As I see things, a sociology of thick concepts starts with two overarching questions. First, what thick concepts are there in society S, group G, or field F at time t? Documenting societies’, groups’, and subcultures’ conceptual repertoires is already a staple of sociological and anthropological research. But there are still objections to be met. For example, fieldworkers often encounter unfamiliar concepts, which are hard to express in English (or French or German, as the case may be). Which are then argued to be inextricably intertwined with “their” way of life or “culture.” “We” just can’t get them. Or so some sociologists and anthropologists claim. Critics are uneasy about these lists of exotic concepts and their relativist interpretations. Why couldn’t we come to understand the concepts ninjō and giri in Japanese culture (not to be confused with the concept guiri in Spanish culture)? Or the concept última-chupada-del-mate in native Uruguayan culture? Or the concept desayuno in native Paraguayan culture—which is allegedly incommensurable with “our” breakfast? Critics argue that the idea of incommensurable conceptual schemes makes no sense; the very idea of a conceptual scheme makes no sense (Davidson 1984; Hollis and Lukes 1982; Quine 1960; Sreenivasan 2001).

Second, given the conceptual repertoire of society S, group G, or field F at time t, what work do particular thick concepts do in social and cultural practices, the public sphere, organizations, and the law? This question should be asked on a case-by-case basis, one thick concept at a time (or one small network of interrelated thick concepts at a time). And it should be asked about specific practices and domains. To what extent, in which ways, and to what effect does concept c figure in practice P in society S at time t? How does c help people and organizations do things in the world: understand it, represent it, exploit it, cope with it, change it? Methodologically, proceed as usual. Rely on representative data (if it suits your research design). Collect representative data (if it suits your research budget). Beware of anecdotal evidence, impressions, and intuitions. Have catholic tastes: surveys, statistical analyses, content analyses, experiments, ethnographies, and historical research can all come in handy. Make sure your article’s writing style is properly social scientific: formulaic, dry, and boastful (Billig 2013; Cowley 1956; Erikson 2008; Wallace and Garner 2013).

Suppose you find that thick concepts tc₁, tc₂, and tc₃ are important in particular domains in society S₁ (or in societies S₁ and S₂). For instance, you examine cruelty, compassion, and humanness in present-day U.S. society: how they’ve contributed to the work of animal welfare organizations, the development of “animal law,” and attendant social practices; what roles they’ve played in the history of society’s relationships to nonhuman animals, education, art, the ethics and regulation of scientific research, and capitalist food production and commercialization. In these processes, wrong and bad and impermissible and inappropriate and not okay haven’t been absent. Yet, cruelty, compassion, and humanness have been central: culturally, politically, and legally. ⁸

Your research demonstrates that tc₁, tc₂, and tc₃ are important in particular domains in S₁. They play significant roles in social practices and the law. People and organizations do things with tc₁, tc₂, and tc₃; they help get things done; they have social consequences. At this point, a tricky issue kicks in. Tricky and exciting. Thin concepts can be applied to anything and can exist anywhere. I don’t know if good, bad, permissible, and impermissible are universal, I’d guess they aren’t, but they could be. What matters is that they could.

By contrast, thick concepts can’t be applied to anything and can’t exist anywhere. As a matter of fact, thick concepts aren’t culturally universal. That’s easy to show. Anthropologists, historians, and sociologists have found many concepts that “we” don’t have. Similarly, not all societies have had exploitation, perversion, dignity, moderation, objectification, daintiness, dogmatism, humane, bestial, natural, and dehumanizing, which Western societies have put to various uses. These thick concepts reflect ideas, distinctions, and categories that are specific to particular societies’ concerns, problems, and historical trajectories. Therefore, you can always empirically ask what thick concepts exist in society S at time t. And you can always empirically ask what’s on particular societies’ and groups’ conceptual menus, catalogs, or repertoires.

However, I’m heading elsewhere. My point isn’t that, as a matter of fact, tc₁ exists here but not there; tc₂ exist there but not here; and tc₃ exists neither here nor there. These would be valuable things to learn about a society, epoch, group, subculture, or organizational field. But I wish to ask where it’s possible (and impossible) for tc₁, tc₂, and tc₃ to exist at all. And why: what makes it possible (or impossible) that tc₁, tc₂, and tc₃ exist at all; what needs to be in place in a given society for particular thick concepts to be possible.

Be possible? This sounds like a strange kind of relation between thick concepts and social facts. It doesn’t sound like a causal relation. What is it, then?

III

One key property of thick concepts is that they both describe and evaluate things. They do descriptive and evaluative work at the same time or in the same breath. One key debate about thick concepts is whether description and evaluation are inseparable.

According to the separability thesis, you can factor a thick concept into two parts. You can isolate its descriptive component, so it’s evaluation-free—say, the descriptive component of dignity, cruelty, exploitation, humanness, generosity, sleaziness, dogmatism, tolerance, garishness, mawkishness, and elegance. Given a “thick term or concept,” is there an “independently intelligible purely nonevaluative description with the same extension” (Väyrynen 2013:12)? Separability advocates say “yes.” ⁹ No evaluative stuff is needed to pick out exploitative, humane, or garish things. Then, you’d end up with ordinary descriptive concepts, like yellow and furniture, to which you’d attach positive or negative “prescriptive flags” (Williams 1985:141) or “gold stars” (Korsgaard 1996:71).

In the case of negatively-valenced concepts, you end up with something like this (where X is a “purely nonevaluative description”): X—that’s not good; X—I disapprove; X—I don’t like; X—avoid!; X—boo!; X—yuck! In the case of positively-valenced concepts, you end up with something like this (where Y is a “purely nonevaluative description”): Y—that’s good; Y—I approve; Y—I like; Y—approach!; Y—yay!; Y—clap, clap.

According to the inseparability thesis, this “disentangling manoeuvre” necessarily fails (McDowell [1981] 1998:201). This is either because a thick concept is an “indissoluble amalgam of two elements” (Roberts 2013a:82), or because there aren’t two distinct elements or parts to speak of, but just one whole. For example, there’s no way to say what cruelty is (or how to use “cruelty” or cruelty) without evaluative stuff creeping in. As Putnam (2002:38) argues, “it certainly is not the case that the extension of ‘cruel’ (setting the evaluation aside, as it were) is simply ‘causing deep suffering,’ nor . . . is ‘causes deep suffering’ itself free of evaluative force. ‘Suffering’ does not just mean ‘pain,’ nor does ‘deep’ just mean ‘a lot of’.” Inseparability advocates ask you to try it yourself. Analyze any other member of the class of thick concepts—say, discernment, decency, lewdness, rudeness, crassness, or infantile. You’ll obtain the same result: no satisfactory nonevaluative description.

The debate between separabilism and inseparabilism has itself become quite thick—and quite technical. ¹⁰ Much is at stake. It has implications for traditional metaethical disagreements between cognitivism and non-cognitivism. It has implications for traditional metaphysical dichotomies between facts and values, because thick concepts seem to challenge the idea that there’s a world of facts, there’s a world of values, and never the twain shall meet. It has implications for the very distinction between thick and thin concepts, whose character is controversial.

My sociological arguments draw on and are informed by these philosophical disputes, but luckily I don’t need to wait for them to be resolved. (Phew!) Nor do I need to take sides in them. (Phew!) Sociologists’ empirical research about thick concepts may make a pragmatic move. Whether or not separabilism ends up being philosophically viable, they can still draw a distinction between thick and thin concepts, as a convenient convention, with no metaphysical aspirations. Whatever the philosophical standing of the class of thick concepts turns out to be, Williams’s basic idea can be sociologically exploited. ¹¹ Moreover, instead of sorting evaluative concepts into two types, thick and thin, sociologists can take thickness to be a matter of degree (Dworkin 2011:182; Scheffler 1987). There’s a continuum from very thin to very thick; there are relatively thick (or thicker) and relatively thin (or thinner) concepts. ¹²

I argue that thick concepts have a special kind of relation to the societies in which they exist. They have a special kind of relation to a specific set of social structural, cultural, and institutional facts. No doubt, thin concepts, too, have important relations to the societies in which they exist. For instance, a comparison of Uruguay and Paraguay shows that the predicates “is good,” “is wrong,” “is permissible,” and “is beautiful” are used in very different ways and come up in very different contexts. Uruguayan laws’ and norms’ permissible is Paraguayan laws’ and norms’ impermissible. Uruguayan business practices’ right is Paraguayan business practices’ wrong. Montevideo Times art critics’ trash and ugliness is Asunción Times art critics’ treasure and beauty. In all likelihood, these aren’t mere correlations, but causal relations. Sociologists can causally account for thin concepts’ use and, more generally, how they figure in social life, institutions, and culture.

Thick concepts are also used in very different ways and come up in very different contexts in Uruguay and Paraguay. If you brought them together, Uruguayan and Paraguayan legislators wouldn’t agree on what counts as cruelty to animals. Uruguayan and Paraguayan rating agencies wouldn’t agree on what it is for a movie to show perverted sexual behavior. Uruguayan and Paraguayan business associations wouldn’t agree on who deserves the annual integrity, honesty, and corporate social responsibility award. Montevideo Times and Asunción Times art critics wouldn’t agree on what literary works are turgid, pompous, high-flown, or orotund. In all likelihood, these aren’t mere correlations, but causal relations. Sociologists can causally account for thick concepts’ use and, more generally, how they figure in social life, institutions, and culture.

The application of both thin and thick concepts varies and these variations can be accounted for. However, thick concepts raise an additional research question. There are specific social structural, cultural, and institutional facts on which their very existence and character depend. These facts might be referred to as “enablers” or “conditions of possibility”: without them, a thick concept couldn’t be, or couldn’t be what it is. Sociologists’ task is to empirically identify, describe, and explain these enablers; find out what it is about a society, group, or field that makes particular thick concepts possible. What social structural, institutional, and cultural arrangements enable dignity, honor, glory, fanaticism, and exploitation? How about gentlemanliness, masculinity, femininity, purity, and juvenile? How about the aesthetic concepts affected, trashy, classy, and grotesque? How about the epistemic concepts dogmatism, education (Siegel 2008), and knowledge (Kyle 2013b)? Under what conditions would these concepts be impossible or make no sense?

Scanlon (2003:276–77) argues that “[t]he thinness of thin concepts such as right, wrong, duty and obligation lies first and foremost in the abstractness, hence relative emptiness, of the ethical ideas that they involve, by contrast with concepts such as coward, lie, betrayal, brutality, honor, and gratitude, whose greater content reflects the distinctive character of particular social worlds.” ¹³ I agree that coward, lie, betrayal, brutality, honor, and gratitude “reflect the distinctive character of particular social worlds.” Unlike the abstractness and relative emptiness of right, wrong, duty, and obligation, thick concepts “involve” a “rich” “evaluative perspective” (see also Sreenivasan 2001; Väyrynen 2013). ¹⁴

However, I wish to go beyond Scanlon’s formulation. Take one of his examples, honor, on which there’s plenty of anthropological and historical research: medieval knights’ honor codes, Bushido honor codes, honor killings, the culture of honor in the South (of the United States or of Spain), early-modern Russia, Bedouin society, Japanese society, and mafia organizations. ¹⁵ The particular character of honor norms and practices is a consequence of the particular character of these social worlds and evaluative perspectives. These are ordinary causal relations.

Yet, in addition, the particular character of these social worlds and evaluative perspectives is such that honor can exist there at all. However it’s applied, whatever its character, causes, and effects, in these places and times you can have honor. What is it about them, then?

IV

Think of contemporary Western societies. Think of their characteristic economic structures, economic institutions, and economic-cultural understandings. Because you’re reading this, I’m afraid you’re likely to live in one of them—in which case, look around you.

Looking around me, I’m afraid a quick-and-dirty list includes: private property; capital; money; credit; financial instruments and markets; corporations; labor; institutionally-, culturally-, and legally-sanctioned distinctions between labor and non-labor, or work and non-work; (some) markets’ having (some) institutional and legal autonomy; (some) markets’ having (some) cultural legitimacy; the capitalist state; formal and informal rules of the game; international trade and trade law; cultural/moral understandings about production, profit, and markets; cultural/moral understandings about human beings, their motivations, values, and preferences.

While this list might be too quick and too dirty, it’ll do for present purposes. It lists some historically contingent facts about some contemporary societies, and specifically about their economic sphere or economic life. Call them “EF.” These things are contingent in the sense that, though relatively widespread today, not all societies and groups throughout history have had them. Not that all societies and groups have them today either.

Given EF, you can do many things. Differently put, EF enables many practices. For instance, here in the U.S. some people choose to buy stocks, get loans, register trademarks, and incorporate their businesses. Most people pay with either rectangular pieces of green paper or rectangular pieces of plastic. In many other places and times, that couldn’t have happened: one or more necessary conditions for buying stocks, getting loans, registering trademarks, incorporation, paying with dollars and debit and credit cards, and paying at all were missing. One of them is that there were no such things as stocks. (Duh.)

EF has another sort of consequence as well. If your friends are anything like mine, they might lament that more and more workers are being paid unfair wages and exploited. These days more and more things are getting commodified. Corporations lack integrity. Businesspeople are materialistic and “would do . . . anything for money,” even profiteering during emergencies, disasters, and wars (Putnam 1981:139). Successful and wealthy people are stingy, never give to charity, and never help the poor. To top it all off, more and more people have bourgeois attitudes and new rich tastes.

These utterances, statements, judgments, or speech acts use the thick concepts unfairness, exploitation, commodification, integrity, materialism, profiteering, successfulness, stinginess, bourgeois, and new rich. These concepts might also figure in societies’ laws and other written and unwritten norms, e.g., the ethics codes of trade associations and companies. They might also figure in societies’ practices, ideas, culture, and institutions, e.g., political and public policy projects, advertising campaigns, job descriptions, awards, rankings, categories and classifications, education, movies, and art.

Trivially, for your judgment about the prevalent commodification of public goods, education, sex, friendship, hospitality, and organ donation to be possible, you have to have access to commodification; commodification needs to be in place. For your judgment about rampant materialism to be possible, you have to have access to materialism; materialism needs to be in place. (Duh.) More interestingly, for commodification and materialism to be possible or make sense, certain structural, institutional, and cultural conditions need to be in place. These conditions are something like EF (to be precise, only part of that list is needed). In more metaphysically loaded terms, commodification and materialism are ontologically dependent on (part of) EF: they couldn’t exist without it. In turn, society S’s existing repertoire of thick concepts enables specific economic practices; concepts make things possible, too.

Besides EF, contemporary Western societies have characteristic political structures, political institutions, and political-cultural understandings. Call them “PF.” I won’t try to exhaustively list them, but think of, e.g., their constitutions, periodic elections, electoral systems, political parties, political organizations, state and legal apparatuses, rights and duties, the idea of a right and the idea of a duty, citizenship, public policy agencies, the public sphere, political journalism, moral/cultural understandings about the nation, and moral/cultural understandings about rights and justice.

Given PF, you can do many things. Differently put, PF enables many practices. For instance, here in the U.S. some people choose to vote, participate in public demonstrations, and invoke constitutional amendments. In many other places and times, that couldn’t have happened: one or more necessary conditions were missing. One of them is that there were no such things as constitutions. (Duh.)

PF has another sort of consequence as well. If your friends are anything like mine, they might criticize fanatical and extremist groups, totalitarian and tyrannical regimes, paternalistic policies, bureaucratic procedures, and discriminatory organizations. Journalists might debate the importance of human rights and freedom (vis-à-vis other social and political desiderata). Pundits might praise egalitarian organizations, diverse organizations, moderate political leaders, tolerant attitudes, and truly democratic decisions. Though not so common where I live, some people take solidarity to be the most fundamental political value, and call their fellow citizens to follow suit. Others take the most fundamental political principle to be loyalty to their party, nation, or ruler, and call their fellow citizens, subjects, or comrades to follow suit.

These utterances, statements, judgments, or speech acts use the thick concepts fanaticism, extremism, totalitarianism, tyranny, paternalism, bureaucratic, discrimination, human rights, freedom, egalitarian, diversity, moderation, tolerance, democratic, solidarity, and loyalty. These concepts might also figure in societies’ laws and other written and unwritten norms. They might also figure in societies’ practices, ideas, culture, and institutions. ¹⁶

Trivially, for it to be possible that a New York Times editorial discusses what counts as a diverse organization, a truly democratic decision, or an extremist group, they have to have access to diversity, democratic, and extremism; diversity, democratic, and extremism need to be in place. (Duh.) More interestingly, for diversity, democratic, and extremism to be possible or make sense, certain structural, institutional, and cultural conditions need to be in place. These conditions are something like PF. In more metaphysically loaded terms, diversity, democratic, and extremism are ontologically dependent on (part of) PF: they couldn’t exist without it. In turn, society S’s existing repertoire of thick concepts enables specific political practices; concepts make things possible, too.

Besides EF and PF, contemporary Western societies have certain religious structures, religious institutions, and religious-cultural understandings (though there’s much variation across and within societies). Call them “RF.” In the absence of RF_C, you couldn’t be baptized or receive the communion. Nor could you have and use the Christian thick concepts sanctity, saintliness, providence, love, piety, charity, celibacy, humility, blasphemy, apostasy, sinfulness, and so on. They’d be impossible, make no sense, or not be what they are. ¹⁷ In the absence of RF_J, you couldn’t keep kosher or go to the mikvah. Nor could you have and use the Jewish thick concepts rachamim, tzedakah, teshuva, anava, tzniut, chesed, kabod, and so on. ¹⁸ They’d be impossible, make no sense, or not be what they are. In turn, society S’s existing repertoire of thick concepts enables specific religious practices; concepts make things possible, too. ¹⁹

Likewise, contemporary Western societies have characteristic gender/sex structures, gender/sex institutions, and gender/sex-cultural understandings (though there’s much variation across and within societies). Call them “GF.” In the absence of GF, you couldn’t answer the call of nature in a gender-neutral restroom, you couldn’t get married or divorced, or a woman couldn’t marry another woman. Nor could you have and use the thick concepts masculinity, femininity, manly, girly, gentlemanly, effeminate, housewife, man of the house, queerness, bisexuality, genderfluidness, agenderness, objectification, womanizer, promiscuity, sluttiness, faithfulness, perversion, lewdness, and so on. They’d be impossible, make no sense, or not be what they are. In turn, society S’s existing repertoire of thick concepts enables specific gender/sex practices; concepts make things possible, too.

You get my drift.

V

Goffman (1983:3) says that “one can try to work backward from the verbal consequences of presuppositions to what is presupposed, allowing the direction of analysis to constrain what one examines.” ²⁰ My suggestion for sociological research goes in the same direction. Don’t start your investigation with society S₁’s structures, institutions, and cultural understandings, and then ask what thick concepts and properties they enable. Instead, start with thick concept and property tc, and then ask what it’s enabled by or dependent on. Or start with a small network of interrelated thick concepts and then ask what it’s enabled by or dependent on. (You may start with a network of concepts because you suspect that their enablers are the same. What makes them a network must be something else, though, lest your reasoning be circular.) If you don’t find tc in society S₂ at time t, if it seems not to exist there, ask if one of S₁’s enablers is missing. Or maybe other things are such that tc is impossible in S₂. Or, rather, tc doesn’t exist in S₂, it didn’t come into being, even though it could have.

This is a distinct kind of sociological approach. If you’re fond of fancy-sounding names and illustrious associates in the history of philosophy, you might call it “transcendental” (Abend 2014b). ²¹ Less pompously, and less demandingly, the fruits of your labor might be called “what-makes-it-possible arguments.” Regardless of their name, sociologists have occasionally made arguments of this sort, since the early days of the discipline and up until the present (Abend 2014a; Eyal et al. 2010; Krause 2014; Miller and Rose 2008; Navon and Eyal 2016; Spillman 2012). But they’ve never occupied center stage. Unlike causal arguments, their logic hasn’t been dissected. Textbooks have ignored them. Their distinctiveness and distinctive potential haven’t been grasped.

This sociological approach involves two steps. The first step is an empirical account. Go to Uruguay and observe how thick concept tc occurs there. Look for tc-speech acts, tc-institutions, tc-practices. Examine tc in action, in interaction, in organizational fields, in culture, in politics. Examine its cultural and material manifestations. Ask people what tc is, where they see it, and what they use it for. Assess if that’s actually what they use it for. Find out what it co-occurs or correlates with, as statisticians would. Trace its history and changes over time, as conceptual historians would.

The second step is an armchair inference. Your goal is to establish what tc is enabled or made possible by. But this isn’t directly observable. ²² So, on the basis of your empirical inquiries, you’ll have to draw an inference, or, more likely, several inferences. Inferences to the best what-makes-it-possible claim, if you will. Sometimes this work is straightforward: an enabler might be entailed by the meaning-in-use of tc, which you empirically observed. Sometimes this work is less straightforward: it requires careful analyses of the contexts, varieties, and ramifications of tc-use, which you also empirically observed. Its conclusions will probably be debatable.

Hard cases might call for special tools and devices, such as thought experiments, counterfactuals, and comparisons. Ask, counterfactually, the absence or presence of what would have made tc impossible in Uruguay. Compare the present to the past and ask the absence or presence of what did make tc impossible at one point. Compare Uruguay, where tc exists, to Paraguay, where tc doesn’t exist. The products of these tools and devices aren’t demonstrative, but they will be consistent with some inferences and inconsistent with others. Then, you’ll have to put forward good old arguments as to what makes tc possible in Uruguay. Perhaps you can also put forward arguments as to what makes tc impossible in Paraguay and, if you feel in a generalizing and universalizing mood, what conditions are sine quibus non for tc to exist anywhere.

Here’s an example of a straightforward inference. Take bureaucratic in this sense: “Of or relating to bureaucracy or bureaucrats; spec. characterized by behaviour or practice regarded as typical of bureaucracy, as an excessive or obstructive concern with formal procedure.” ²³ bureaucratic couldn’t exist if there weren’t (something like) formal organizations. Which we happen to have, our lives happen to be structured by, but are socially and historically contingent.

Here’s a more interesting case. Take materialism to be something like this: “An emphasis on or preference for that which is material, at the expense of spiritual or other values; (now) esp. the tendency to treat material possessions and physical comfort as more important or desirable than spiritual values; a way of life based on material interests.” For materialism to be possible, values should come in kinds. One kind is material values and another kind is spiritual values. That a person treats her material possessions as more important than spiritual values requires that she can have material possessions, i.e., that things can be possessed by her. Which in turn requires private property in one way or another. Also, some kind of importance-understandings or importance-attitudes should be available.

Take any other tc. Sociologically and historically observe commodification-speech acts and practices. Study the extensive empirical literatures on it. Then tell us what economic facts need to be in place for it to be possible. Sociologically and historically observe human rights-speech acts and practices. Study the extensive empirical literatures on it. Then tell us what needs to be culturally and legally assumed about human beings’ nature for human rights to be possible or make sense, what metaphysical understandings and religious institutions human rights is compatible (and incompatible) with, what institutional and cultural facts make human rights possible (and impossible). Sociologically and historically observe honor-speech acts and practices. Study the extensive empirical literatures on it. Well, not so fast: there seem to be several honor concepts. In different places, times, and languages. Do they have common enablers? Or should you give an account of one of them only?

You may sociologically investigate thin concepts as well. By investigating very thick, relatively thick, relatively thin, and very thin concepts, you can ask if ontological dependence on structural, institutional, and cultural facts is a matter of degree. And if so, what exactly these degrees are degrees of. Is the thickness continuum correlated or even coextensive with an ontological-dependence continuum? Is it analytically true that the thinner a concept, the fewer structural, institutional, and cultural facts it depends on? Is that the very definition of thickness—e.g., rightness depends on fewer things than kindness, which depends on fewer things than diversity? (What’s logically true is that the fewer things a concept depends on, the more places it can exist in. Note: “can,” not “does.”)

As I see things, the focus of sociological research on thick concepts isn’t causes but enablers. Not what they’re caused by, but what they’re made possible by, or what they depend on. You might have been wondering what exactly this relation amounts to. Admittedly, I haven’t been sufficiently clear about it. It’s high time I was.

VI

What’s the character of the relation between a particular thick concept or a particular repertoire of thick concepts on one hand, and a particular set of structural, institutional, and cultural facts on the other? I can think of five candidates; five kinds of relation that may obtain between A and B:

Ontological dependence is a metaphysical relation: “[o]ne thing x will depend upon another y just in case it is necessary that y exists if x exists (or in the symbolism of modal logic, (◻ExEy))” (Fine 1995a:270). For example, metaphysicians “consider the relationship between a mouth and its smile to be a good illustration”: smiles ontologically depend on mouths (Koslicki 2012:188). Similarly, “you can’t have the hole without the doughnut” and you can’t have the holes without the Emmental cheese (Correia 2008; Varzi 1996:5). Holes depend on their hosts ontologically—as opposed to causally, logically, or conceptually.

Constitution relations are also common in metaphysics, e.g., in arguments concerning identity, mereology, explanation, and social ontology. For instance, a material constitution relation obtains between “the ten pound portion of clay” you bought this morning and “the statue of King Alfred the Great” you made out of it in the afternoon (Thomson 1998:149). In social ontology, a fundamental question is what groups, classes, institutions, social structures, and societies are constituted by, or consist in, or supervene on (Durkheim [1895] 1968; Epstein 2016; Schmitt 2003).

Less formally and sometimes somewhat metaphorically, constitution relations are appealed to in social and political theory, cultural studies, and the humanities. People may say that B is “built into” A. Or that A and B are “inextricably intertwined.” Or “mutually constitutive” (Healy 2017:119).

Presuppositions have had several uses in several disciplines. As applied to utterances, statements, or speech acts, they are investigated by linguists, conversation analysts, logicians, and philosophers of language (Frege 1892; Stalnaker 1973; Strawson 1950; see also Austin 1962; Goffman 1981, 1983 ²⁴ ). For example, “[i]f I say that the Queen of England is bald, I presuppose that England has a unique queen, and assert that she is bald” (Stalnaker 1974:47). More generally, “[t]he presuppositions of an utterance are the pieces of information that the speaker assumes (or acts as if she assumes) in order for her utterance to be meaningful in the current context. This broad characterization encompasses everything from general conversational norms to the particulars of how specific linguistic expressions are construed” (Potts 2015:168-169). Plus, presuppositions can help link language, institutions, and social reality: “the institution of buying and selling is presupposed to the description ‘sending a bill,’ as it is to ‘being owed for goods received’ . . . ” (Anscombe 1958:72; Searle 1964, 1969:51, 2010).

Going by the name of “preconditions,” “prerequisites,” or “assumptions,” presupposition relations also show up in social theory, history, and metaphysics. According to R.G. Collingwood’s Essay on Metaphysics, “all thinking . . . is an attempt to answer questions. Underlying every question is a presupposition which gives it purchase, itself possibly an answer to a prior question, behind which may lie some more fundamental presupposition” (Dray 1995:140). Then, metaphysics would be well advised to become “the study of the presuppositions which govern the various areas of experience” (D’Oro 2002:16). ²⁵

Conditions of possibility have had many takers, too. Kant and Kantian epistemologists ask how knowledge is possible and what makes knowledge possible (Brueckner 1996; Cassam 2007; Greco 2000; Strawson 1966). Their “transcendental arguments” hope to defeat the skeptic, as they “are supposed to demonstrate the impossibility or illegitimacy of [the] skeptical challenge by proving that certain concepts are necessary for thought or experience” (Stroud 1968:242). If this a priori tactic can be shown to work, the skeptic’s argument would be based on the very things she’s trying to cast doubt on. From which it’d follow that skepticism fails and all well-thinking philosophers can be happy and relax.

Conditions of possibility have also been mobilized in social theory, sociology, and history. You find them in Foucault and Bourdieu and their disciples (who frequently draw on Kant [Koopman 2013; Power 2011]). You even find them in scholars whose native language is neither French nor German. Besides academese, at times this terminology occurs in ordinary speech. You studied Arabic in college, which enables you to read Averroes’s The Incoherence of the Incoherence in the original. That there are planes makes it possible to get from Madrid to Montevideo in 13 hours. Under what conditions would it have been possible for Averroes to understand Aristotle’s Poetics (Borges 1949)?

Because of the polysemy and cumbersome genealogy of “conditions of possibility,” you might prefer to travel lighter. Which “to make possible” and “to enable” enable you to. Bonus points: you’d avoid the word “condition” and its own polysemy and cumbersome genealogy.

One problem is what makes it possible that there be A; another problem is what makes it possible that A be intelligible. Inquiries into intelligibility hark back to ancient epistemologists’ analyses of knowledge, cognition, and the connections between the mind and its objects. What makes intellection and intelligibility possible? Is a higher entity or transcendental source required, such as God and divine illumination, the Good, or “active intellect,” “agent intellect,” or “maker mind” (“nous poiêtikos”)? ²⁶

Contemporary phenomenologists and ethnomethodologists have produced other sorts of inquiry into intelligibility. They emphasize people’s subjectively experiencing and apprehending reality, and how things come to make sense to them. They highlight the necessary conditions for A’s being apprehended the way it is, and the necessary conditions for A’s making sense at all. They speak of “assumptions,” “the taken-for-granted,” and “presuppositions” (Garfinkel 1963:209–14). These issues are important, too, in sociological research on social interaction, and the shared assumptions or implicit rules of subcultures, scientific and religious communities, and organizational fields. Mutatis mutandis, they’re important in the late Wittgenstein and Charles Taylor’s (2004:23–28) “background understanding” (though he oscillates between relations (5) and (4): it “makes sense of any given act” and “makes [an] act possible”).

***

All in all, these five kinds of relation between A and B originate in different discussions and disciplines. ²⁷ Their similarities, differences, and overlaps are complex. There are lots of moving parts here, because A and B can be many kinds of things, e.g., the constitutive components of tables and statues, the presuppositions of utterances, and the conditions of possibility of practices. Even more moving parts: your ontological commitments about the nature of these things. Which is all the more tricky if, instead of tables and statues, you’re talking about social, cultural, and institutional stuff, and the practices of “self-interpreting animals” (C. Taylor 1985). Which is what I’m talking about.

VII

Placeholders A and B can now give way to the relata under scrutiny. Let tc stand for a thick concept and F for a set of social structural, institutional, and cultural facts:

I’m not going to argue for any of these five candidates here. ²⁸ Future work should look at each of them from the perspective of empirical sociological research, consider their pros and cons, alongside four questions. First, do you need to choose one and only one of these five, and then make a general claim about the relation between tc and F: what they (all of them) are? Or, rather, might this relation vary? If so, how does it vary? Second, given a particular tc and a particular F, do you need to choose one and only one of these five? Or might several relations obtain between them at the same time? Put differently, are all, some, or none of these relations exclusive? Third, I said that thickness is a matter of degree; tc is a simplification. Do these relations come in degrees, too? If so, how? Or, rather, do they start to apply at a certain thickness level? Fourth, given a particular tc, might there be several F-configurations that make it possible or intelligible?

Despite their differences, statements (1), (2), (3), (4), and (5) can serve similar sociological and methodological functions. They must also confront similar issues.

First issue: you can read statements (1) through (5) in two ways. One reading is that tc can’t exist without F. If there were no F, there could be no tc. If F weren’t there, tc couldn’t be there either. The other reading is that tc can’t be what it is without F. ²⁹ In other words, tc can still exist, but were it not for F, it’d have a different essence, identity, or features. Essentially, this is the distinction between existential and essential relations. ³⁰ In either case, thick concepts are analogous to “dependent objects”: they “cannot stand on their own” (Fine 1995b:463). To be sure, there are many kinds of modal statements and much turns on just what “can/can’t” or “could/couldn’t” you’re relying on. You can’t castle if you’ve already moved the king or the rook, Avicenna couldn’t chat with Averroes, bachelors can’t be married, you can’t fly nonstop from Montevideo to Vienna, and penguins and cassowaries can’t fly. Then, what exactly does it mean that “tc can’t exist without F”?

Second issue: statements (1) through (5) try to represent non-causal relations. B isn’t a cause of A; F isn’t a cause of a tc. This idea might be cashed out in various ways, only some of which are applicable to the social world. One example is Kant’s: “spatiality [is] presupposed by spatial representation” (Hatfield 2006:76). If you had no space concept, your mind couldn’t perceive or represent material objects, such as the avocado on your countertop. space doesn’t cause you to perceive the avocado. Rather, space makes it possible that you perceive the avocado. Another example stems from Plato (Republic VI, 508a–509b) and Aristotle (De Anima III.5, 430a16–18): the sun or light, i.e., there being the sun or there being light. You might argue that it’s a cause of your seeing the avocado. Or you might argue that it’s a condition for the possibility of your seeing the avocado. Which is it?

Distinguishing these five relations from causal relations isn’t free from trouble. For example, constitution and dependence have connections to downward causation, emergence, supervenience, and explanation. Most troublesome is relation (4)—making it possible, enabling, or being a condition of possibility. There being oxygen in the atmosphere made the Library of Alexandria’s burning to the ground possible. It made possible, too, that Trump win the 2016 presidential election and that I drink a cup of coffee this morning. It seems odd to claim that there being oxygen in the atmosphere caused these outcomes, even if it does pass a counterfactual test. Yet, it’s controversial how to draw the distinction between causes and conditions (or background or enabling conditions). It’s controversial whether it should be drawn at all (Broadbent 2008; Mackie 1965, 1974; Schaffer 2005; Skow 2016).

Enablers or conditions of possibility involve distinct problems. Much like Aristotelian potentiality, a claim about possibility isn’t a claim about actuality. If B, then A can exist. Whether A does exist is a separate question. It might or it might not. There being oxygen in the atmosphere made my drinking a cup of coffee this morning possible (which I did). But it would have also made my drinking a cup of tea possible (which I didn’t). It’s a condition for the possibility of my flying tomorrow to New York (which I will) or to Montevideo (which I won’t). By contrast, causes must be actual causes and have actual effects (even if imaginary, counterfactual causes and effects may lend you a methodological helping hand).

tc is an actually existing concept, which F makes possible. The study of tc is what my sociological proposal is up to; that’s what sociologists can advantageously investigate. But you could object that these are only actual cases. Actuality entails possibility: if tc exists, it follows that it’s possible for it to exist. Shouldn’t an investigation about conditions of possibility also include that which isn’t actual but only possible? Shouldn’t it include nonexistent thick concepts that F also enables, which could have come into being, but as a matter of fact didn’t?

VIII

Ethology is the science of animal behavior. Typically nonhuman animals. But also human animals. ³¹ Plus, like Darwin, ethologists compare the two. Take decision-making (Farine et al. 2014; McFarland 1977; Seeley 2010):

Humans routinely make many decisions collectively, whether they choose a restaurant with friends, elect political leaders or decide actions to tackle international problems, such as climate change, that affect the future of the whole planet. We might be less aware of it, but group decisions are just as important to social animals as they are for us. Animal groups have to collectively decide about communal movements, activities, nesting sites and enterprises, such as cooperative breeding or hunting, that crucially affect their survival and reproduction. While human group decisions have been studied for millennia, the study of animal group decisions is relatively young, but is now expanding rapidly. It emerges that group decisions in animals pose many similar questions to those in humans. The purpose of the present issue is to integrate and combine approaches in the social and natural sciences in an area in which theoretical challenges and research questions are often similar. . . (Conradt and List 2009:719)

The analogy can go in both directions (Kalenscher and van Wingerden 2011; Santos and Rosati 2015). Studies about nonhuman animals may draw parallels to and use ideas and models drawn from studies about human animals. Studies about human animals may draw parallels to and use ideas and models drawn from studies about nonhuman animals. One methodological strategy is comparative cognition studies that use both human and nonhuman subjects. One theoretical strategy is to postulate “traits” and “capacities.” Humans’ traits and capacities are continuous with nonhuman animals’. Just like the brain is (Dehaene et al. 2005).

Analogies between human and nonhuman animals have a long history in both philosophy and science. They’ve always had fervent proponents and fervent opponents. Their debates have led to general, fundamental questions about humans’ place in the natural world. Is there such a thing as human nature? Which has larger causal effects, nature or nurture? Genes or culture? What’s the product of evolution and what isn’t? Are humans unique? Is there a difference in kind between human and nonhuman animals? Do nonhuman animals have language, culture, morality, beliefs, desires, and intentions?

Disputes about these questions are still common today; perhaps they’re more common than ever. They’re common both in academic forums and the public sphere. Unfortunately, for the most part they sound tired. They repeat themselves. They bring about two familiar camps. While both have access to new methods, data, and theories, most of their arguments have long been available. It’s easy to predict how this crowd will argue for p and how that crowd will argue for q. The composition of each camp is also familiar. A bit of information about someone’s background suffices to predict how she’ll position herself—e.g., a devout Christian, a devout Darwinist, and a devout cultural anthropologist. To make things worse, people talk past one another. It’s not clear what it’d be to get it right. Their disagreements end up being merely semantic or verbal: how to use the English words “language,” “culture,” “belief,” “intelligence,” “play,” and “morality” (Belleri 2018; Chalmers 2011; Jenkins 2014). ³²

I don’t wish to participate in these disputes. Not if they’re framed that way. Instead, I suggest that my arguments about thick concepts can contribute to our thinking about them.

Let Delmira be a social scientist who’s keen on similarities, analogies, and continuities. Be it groups, classes, nations, or individuals, she likes to examine what they share, why, and to what effect. She likes to put forward general claims about social structures and processes. That’s her preferred emphasis. Delmira’s experiments compare humans’ and nonhuman primates’ reactions to stimuli, and how their behaviors and neurobiological underpinnings are alike.

Now, Delmira isn’t oblivious to what nonhuman primates lack.

DELMIRA: “Humans are constantly making conscious or unconscious judgments about whatever appears before them: a family member, a person in need, a handsome date, a venomous snake, or a pepperoni pizza. Whether verbally or not, they judge the Torres García painting they’re looking at, the version of ‘Caribou’ they’re listening to, and the Leonard Susskind presentation they’re watching on YouTube. They may make more general judgments about the aesthetics of rock music, the value of string theory, and the benevolence and perfection of God. All of these judgments require language and sophisticated cognitive skills. However, they’re analogous to nonlinguistic reactions to stimuli, which human and nonhuman animals share. Like reactions, judgments can be positively or negatively valenced. ‘I like’ or ‘I don’t like.’ ‘Yay!’ or ‘boo!’ ‘Approach!’ or ‘avoid!’

Nonhuman primates don’t have science, art, morality, and religion. Yet, some of their traits, capacities, and behaviors are the building blocks, evolutionary precursors, antecedents, or roots of science, art, morality, and religion. ³³ They don’t have scientific theories and methods, but they hold true beliefs about natural phenomena. Like science, these beliefs are taught, learned, or culturally transmitted. Nonhuman primates don’t make aesthetic judgments, but they approach what’s pleasing and avoid what’s displeasing to their eyes and ears. They don’t make moral judgments, but they can be altruistic and inequality- or inequity-averse. They like (and sometimes reward) individuals who cooperate, reciprocate, and help. They dislike (and sometimes punish) individuals who aren’t prosocial and cheat. ³⁴

So, this is a continuum. Just like nonhuman animal species don’t have language, but they do have simpler communication systems.”

IX

Let Juana be a social scientist who isn’t persuaded.

JUANA: “I’m afraid I don’t buy your argument, Delmira. While it’s contentious what science, art, morality, and religion are, it’s uncontentious that they encompass many practices, norms, judgments, and speech acts. In them, you find both thin and thick concepts. Very thin and very thick ones. You find permissible- and bad-speech acts, institutions, and practices. And you find piety- and dignity-speech acts, institutions, and practices.

I don’t know if your analogy/continuum works for thin concepts and judgments, and the practices and institutions based on them. It might be useful for certain purposes. That said, you’d need a convincing transition from nonhuman animals’ approach!- and avoid!-reactions to human animals’ goodness-, inappropriateness- and wrongness-judgments. This transition is all too often declared, but not fleshed out and argued for.

In either case, your analogy/continuum doesn’t work for thick concepts and judgments, and the practices and institutions based on them. Why? Because of the relation between tc and F, be it ontological dependence, constitution, presupposition, making it possible, or making it intelligible. F is (in some way) part of or built into tc. tc couldn’t exist if these facts didn’t happen to obtain in society S at time t. piety, dignity, and gentlemanliness are that kind of concepts. permissible, bad, and right aren’t. Nonhuman animals’ reactions and behaviors need no such facts. They have causes, but no presuppositions. They aren’t made possible by anything social structural, institutional, or cultural. They can exist anywhere. This isn’t a quantitative (in degree) but a qualitative (in kind) difference. Therefore, your tactic, Delmira, won’t take you very far. It won’t be useful. One key property of thick concepts, judgments, practices, and institutions will be missing.

This qualitative difference remains even if you view people as organisms that react to stimuli. Even if you view them as having ‘traits’ and ‘capacities.’ Even if you call their practices ‘behaviors.’ You may go ahead and do that if you wish. But science, art, morality, and religion have thick components that are qualitatively different from nonhuman animals’ building blocks. That’s not up to you or me.”

DELMIRA: “I came up with a forceful analogy/continuum between human and nonhuman animals. Now you come along, Juana, and contend that it doesn’t work and it’s not useful. What are you doing in the Porch of the King Archon? Just kidding; thanks so much for your feedback. But I’m afraid your objections fail, because you can’t establish in the abstract whether an analogy/continuum works or is useful. This is always a function of your perspective, interests, and aims. In using the expression ‘for certain purposes’ you’re acknowledging this yourself.

You claim that thick judgments are in some respects different from nonhuman primates’ reactions. That’s certainly the case. But they are similar in other respects. Some researchers and disciplines like to study the similarities, analogies, and continuities between X and Y. They like to advance general, nomothetic claims; they get excited about what their objects share. I’m one of those. Other researchers and disciplines like to study the dissimilarities, disanalogies, and discontinuities between X and Y. They like to advance particular, idiographic claims; they get excited about what makes their objects unique.

Both are legitimate paths for social science. It’s pointless to ask if, after all is said and done, X and Y are really similar or really dissimilar. It depends on the respect to which you’re asking. Don’t forget that any entity is similar and dissimilar to any other in an infinite number of respects. Neither similarities nor dissimilarities are more important, interesting, or useful simpliciter. It depends on what you’re trying to do. Even the laws of physics might be useless and lie (Cartwright 1983).

I agree that exploitation-judgments and practices are made possible by social facts in a way that wrongness-judgments and practices aren’t. However, either can be recruited by a project that focuses on similarities and continuities between human and nonhuman animals—provided it convincingly specifies similarities with respect to what and useful to do what.

In sum, if you’re making claims about usefulness, you have to bite the bullet: unlike weight, speed, or color, usefulness is always a relative property.”

JUANA: “Aren’t weight, speed, and color relative properties as well?”

X

Delmira and Juana disagree about human and nonhuman animals, their similarities and dissimilarities, and what thick concepts do and don’t show. In order for their dialogue to move forward, though, they need to find common methodological ground. What is it for an analogy to work, be useful, or be helpful? What are analogies’ success and failure conditions? How might an analogy be wrong? What makes a continuum a continuum? What’s the difference between a difference in kind and a difference in degree? Under what conditions is a disagreement merely verbal? Only then can Delmira and Juana turn to their substantive concerns: thick and thin concepts, and the analogy between human and nonhuman animals. Only then can they fruitfully discuss the case at hand, how nonhuman animals’ movements should be depicted, how judgments differ from reactions to stimuli, and what kind of differences are the differences between human and nonhuman animals (Boyle 2012; Penn, Holyoak and Povinelli 2008; Povinelli 2004).

My aim isn’t to adjudicate Delmira and Juana’s disagreements. That’d be hard work. But I agree with one of Juana’s points, whose sociological implications strike me as significant. Thick concepts raise these problems in a way that thin concepts alone wouldn’t have. This is because of the special kind of relation between thick concepts and society. Take any tc. Take humanness, barbarism, civility, public-spiritedness, progressiveness, perversion, or uptightness. Social structural, cultural, and institutional facts are constitutive components, presuppositions, or enablers of tc-speech acts, tc-institutions, and tc-practices—they aren’t just a context or “environment” (Perrin and Lee 2007). Were it not for these structural, cultural, and institutional things’ being the case or being in place, tc-speech acts, tc-institutions, and tc-practices would be impossible or unintelligible. ³⁵

Because of their special properties, thick concepts may shed new light on a traditional problem in social theory. These days many scholars are producing accounts of what people do, believe, want, and say, how they relate to one another, and how they organize themselves. Many scholars are producing explanations for why that’s so: why people do, believe, want, and say what they do, believe, want, and say. These scholars’ disciplinary backgrounds and epistemological inclinations are diverse. They use diverse methodological tools and technical terms. There are those who say “people” and those who say “humans.” There are those who try to explain “social life” and “social institutions” and those who try to explain “human traits” and “human behavior.” All of these projects might be referred to as “social science,” broadly understood.

If my arguments hold water, social scientists shouldn’t assume that evaluations are prototypically or paradigmatically about good, wrong, okay, beautiful, and appropriate. They should be clear about the distinction between thick and thin. Otherwise, their research would pay the price. For one thing, it’d be considering only a small and peculiar part of societies’ evaluative practices. For another, it’d be neglecting a crucial relation, A’s enabling B, which thick concepts bring into focus.

This isn’t a point about social ontology or metaphysics, but about empirical social science; its methodological standards and epistemological bases (see also FitzGerald and Goldie 2012; Gorski 2013; Sayer 2017; Thacher 2015). A social science of good, wrong, okay, beautiful, and appropriate stuff is a tempting prospect, because these concepts are methodologically tractable, translatable across domains and disciplines, and compatible with prevalent styles of scientific thought. Further, they make sense to external audiences, funding agencies, and the media. Tempting though this prospect might be, it’d miss things that shouldn’t be missed. At best, it’d take the part for the whole.

Human beings are creatures in whose lives there are thick concepts. They have tc-speech acts, tc-institutions, and tc-practices. I’ve argued for their sociological investigation. I’ve argued that this comprises the investigation of what they’re made possible by. These concepts couldn’t exist if particular social structural, institutional, and cultural facts didn’t obtain. More tentatively, I’ve suggested that a larger issue is at stake. In humans, social facts do both causal and enabling work. Scientific research on humans should concern itself with both. Scientific research on nonhuman animals has much to say about causal processes, e.g., by experimentally teasing out the causes of chimpanzees’, rats’, and bees’ behavior. Yet, their behavior isn’t made possible by social stuff in the relevant sense. In this sense, the science of human animals differs from the science of nonhuman animals.

Is this an important difference? What is it an important difference for?