Democratizing Social Scientists’ Words and Distinctions: Where and Why Moral Reasons Are Decisive

Abstract

Sitting at your desk, working on your social science paper, you take yourself to be free to use key words and distinctions as you see fit. I argue that these decisions aren’t up to you. Instead, they should be made collectively, by the social science community you’re embedded in, talking to, and contributing to, along with its stakeholders. Such democratic and deliberative processes will be driven by the common good, as opposed to benefiting particular social scientists, schools, research programs, methods, or regions of the world. My book, Words and Distinctions for the Common Good, is an invitation: I invite social science communities to democratically work on how to use their key words and distinctions—and which epistemic and moral goods these uses should help attain.

1. Words And Distinctions Matter

It’s 2024 and the methodological and epistemological foundations of mainstream social science appear wobblier than they used to be. ¹ ‘The replication crisis’ and ‘the pre-registration revolution’ are conspicuous signs of this. Concerns have been raised, too, about the role of ‘theory’ in the time of big data; value freedom; tensions between epistemic and non-epistemic values; causal claims’ centrality; the assumptions and limitations of measurement, construct validity, and randomized controlled trials; and the neglect of social science produced outside the U.S. and Western Europe. Meanwhile, philosophers of science have been long poking holes in the traditional scientific model, which rests on untenable premises about objectivity, truth, and the relationship between theory and evidence.

In my book, Words and Distinctions for the Common Good, I focus on a particular aspect of social scientists’ methodological and epistemological uncertainties. Social science literatures won’t stop disagreeing over their key words and distinctions. Be it ‘intelligence,’ ‘gender,’ ‘work,’ ‘institution,’ or ‘poverty,’ there are numerous overlapping definitions, concepts, or conceptualizations of it, but no clear method and criteria to determine their relative worth. Nor is there a clear method to determine the acceptability of distinctions and categories: for example, how to group things in scientific models and explanations. Definitions, concepts, categories, constructs, and measures proliferate. Researchers and labs often talk past one another. Their findings aren’t commensurable. Results have to be communicated using air quotes: they are about ‘happiness’ in diverse cultural contexts, as defined in this paper (but there are other recent papers that define it differently). This is confusing for policymakers, the public, and the media.

Seeing as this state of affairs persists, many social scientists throw their hands up in the air. There can’t be shared communal criteria to evaluate words and distinctions. Everyone is free to use them as they see fit, provided they openly communicate their choices.

I have a better plan. It involves you.

* * *

What is an organization? What is an institution? What’s diversity, leadership, entrepreneurship? Sustainability, work, trust? What is capitalism? You’d think that a sociologist of work should know what work is (and isn’t), a criminologist should know what a criminal gang is (and isn’t), and a psychologist should know what cognition and emotion are (and aren’t)—just like a planetary scientist is expected to know what a planet is (and isn’t). Who else could you ask what a planet is or what crime is? Similarly, an organization theorist should know what an organization is, shouldn’t they?

I don’t think they should. They’d be better off bypassing questions of the form ‘what is F?’ altogether. Sometimes social scientists explicitly make what-F-is claims, sometimes they’re assumed to lie in the background of a field or literature, as a sort of conceptual bedrock. Either way, they lead to blind alleys, since it’s unclear what makes one claim better than another; what criteria and methods could settle a disagreement. I argue that a different approach is called for. Not about the nature of social objects, but about the uses of words and distinctions.

Social scientists do research on organizations, institutions, gender, populism, and race and ethnicity. They make descriptive and explanatory claims about empathy, intelligence, neoliberalism, and power. They advise policymakers on diversity, work, digitalization, and religion. And yet, they can’t agree on what these things are and how to identify them. ² Many social science literatures have never-ending disputes over what F is (sometimes referred to as ‘conceptual’). I argue that these disputes are misconceived, so it’s no wonder that they never end. You can’t establish which of the competing views is true, because there’s no truth to be had here. There’s nothing for capitalism, trust, or entrepreneurship to be. Contrast with the truth-aptness of social scientists’ empirical claims. If you put forward an argument about the economic and organizational effects of diversity policies, empirical research will help determine its truth or falsity. But no amount of empirical research will tell you what diversity is.

Social scientists’ uses of words and distinctions can’t be true or false, but they can be good or bad, better or worse. Goodness criteria are therefore needed, and they should be obtained through deliberative, democratic processes. These won’t be driven by what works for your papers and projects, but by everyone’s epistemic and moral good. Scientific communities and their stakeholders.

2. Disagreements, Good And Bad

Jacinta and her collaborators wrote a terrific paper about the chief features of organizations in France, Spain, and Germany, and their diachronic changes. It starts with their definition of ‘organization’: organization_J. To test their claims, they collected data about a large set of organizations_J. As it happens, Karen and her collaborators also wrote a terrific paper about the chief features of organizations in France, Spain, and Germany, and their diachronic changes. It starts with their definition of ‘organization,’: organization_K. To test their claims, they collected data about a large set of organizations_K.

Jacinta’s and Karen’s papers reached seemingly contradictory conclusions about the features of French, Spanish, and German organizations, their similarities and differences, how and why they have changed over time, and what hinders their efficient functioning. However, on closer examination, Jacinta’s and Karen’s conclusions aren’t contradictory, because they’re referring to different objects and have different explananda. ‘Organization’ doesn’t pick out the same entities in the two papers (despite many overlaps). Duh! How many children attended your friend’s party last Friday? The answer depends on the extension of ‘children,’ e.g., whether her two nephews count, who are 12 and 13 years old. Whether it’s a population of organizations, religions, artworks, or organisms, there’s no objective, automatic, algorithmic way of counting the entities your claims are about. Choices must be made. But on which grounds?

Unsurprisingly, if you asked her, Jacinta would insist that organization_J is a better and ‘more useful’ definition than organization_K. (Or, as it’s sometimes said, it’s a better conceptualization or conception of organization.) Unsurprisingly, Karen would claim just the opposite. A third scholar, Laura, argues that both organization_J and organization_K are OK, neither is superior, plus she applauds the multiplication of definitions. Is there a way to adjudicate this three-way disagreement? You’d think that social science methodology provides clear rules and mechanisms to assess competing definitions’ worth, usefulness, point, overall merit. (Or, as it’s sometimes said, what makes one conceptualization or conception of organization better than another.) I’m not sure that it does. ³

The history of the social sciences is rife with disputes that don’t get anywhere, due to confusions about the character of the issues at stake and how progress could be made. Not only is this unproductive and frustrating. Absent fair procedures and norms, Thrasymachus rubs his hands: the stronger can get away with pursuing their self-interest unimpeded (Republic 338c).

This isn’t to say that scientific disagreements are necessarily unproductive. It depends. Monica Prasad (2024) observes that there are many kinds of disagreements in social science and many kinds of things about which social scientists may disagree. I agree. Importantly, there’s good and bad: substantive disputes and merely verbal ones; epistemically beneficial and detrimental dissent; fruitful and fruitless consensuses; conflicts that contribute to knowledge and those that throw a monkey wrench in the works (Biddle & Leuschner, 2015; Chalmers, 2011; Dellsén & Baghramian, 2021; Mouffe, 1999). Prasad appreciates disagreements of the good sort. I agree. When it comes to practical problems, she points out that inaction isn’t an option and communities have reason to reach agreements. I agree. Prasad (2021a, 2021b) and her fellow ‘problem-solving sociologists’ have “created an infrastructure to help scholars interested in conducting research that tries to change the world.” They’ve been building institutional tools and mechanisms for communities to realize their epistemic and moral goals. Words and Distinctions for the Common Good is on the same page.

3. What’s At Stake

I don’t love the word ‘concepts,’ because it can obscure an important distinction between words (henceforth ‘w’) and distinctions (henceforth d). The former: social scientists don’t agree on how to use key word ‘w’ in their literature or subfield. Different people use it differently. The latter: social scientists don’t agree on which distinctions to draw and accept, or, which is roughly the same, how to sort entities (objects, observations, data points) into classes (categories, groups, types). Different people draw and accept different ones.

In both cases, word ‘w’ and distinction d, adjudicating a dispute necessitates second-order criteria: what makes a good proposal good, goodness according to which principles and aims, and who gets to establish these. Unfortunately, in social scientists’ actual disputes, second-order criteria are fuzzy, unthematized, or unnoticed. For example, consider these situations:

• (a) You do research on work. Is housework work? More generally, are there different types of work? If so, how many types? And how many types of types?

• (b) You do comparative and historical research. How to distinguish between capitalist, proto-capitalist, and non-capitalist societies? Are there varieties of capitalism? Varieties on the basis of which property or properties?

• (c) Whatever your historical topic is, are your observations in present-day China and eighteenth-century Uruguay observations of the same phenomenon, practice, or institution, which is picked out by ‘w’?

• (d) Your models have social class and gender variables. How many classes are there? How many genders are there? How to decide?

• (e) Your dissertation is a case study. Your advisor says it’s excellent work, but it fails to elucidate what your case (phenomenon, relation, process) is a case of. Upon reflection, it strikes you that it can be a case of X, Y, or Z. Is one of them the correct answer? Or can you choose freely?

This is how I put the questions in Words and Distinctions for the Common Good. Words: given ‘w,’ how should it be used in social science community S? For instance, in which ways should ‘gender’ and ‘terrorism’ be used—by sociologists of gender and the interdisciplinary literature on terrorism, respectively? On which grounds? Alternative formulation: how should word ‘w’ be defined?

Distinctions: should d be accepted in social science community S? For instance, imagine that someone proposes a distinction between childhood, adolescence, adulthood, and old age. Or between ‘White,’ ‘Black or African American,’ ‘American Indian or Alaska Native,’ ‘Asian,’ and ‘Native Hawaiian or Other Pacific Islander.’ ⁴ Should it be accepted? On which grounds? Alternative formulation: should objects be classified according to distinction d?

These are hard questions. Hard and practical, not theoretical. My book isn’t trying to contribute to philosophy or ‘theory.’ It’s not about metaphysical conundrums, e.g., the real nature of corruption or the essence of capitalism. Instead, I’m talking about the ordinary practice of social science and the logic of research: what to do tomorrow at your lab, field site, archives, or computer; how to select and code your data; how to design your experiments; what to observe and sample; what generalizations you’ll be entitled to. We’re talking tomorrow morning.

We’re talking about what to do, about research practice, about action. Action can’t wait. You may remain undecided about the cogency of string theory or the causes of the fall of the Roman Empire. You may remain uncommitted in normative ethics as to whether you favor consequentialism, deontology, or virtue ethics. ‘I have to give more thought to the arguments, study the scholarship on them, and then I’ll see where I’m at.’ But this attitude doesn’t cut it for action.

You’re witnessing a nasty act right now right in front of you. Someone getting mugged or sexually harassed. Your boss implementing an immoral business plan, which will harm workers, customers, or the environment. Either you’ll do something about it or you won’t. There’s no comfortable abstention, postponement, neutrality for the time being. There’s no such thing as inaction either, since inaction is a kind of action: not doing anything to stop what’s going on would be to do something. I’ve given moral examples, but my point isn’t exclusively about morality. Action has always this sort of immediacy. The server is asking what you’d like to order. Your students are waiting for the lecture to get started. Something must and will be done; there’s no ‘pause’ button.

Social scientists do research day in and day out. Their projects must go on. Deadlines must be met. Papers and dissertations must be written, journal editors must accept and reject manuscripts, funding agencies must select grant and fellowship awardees, and recruitment committees must make job offers. Papers, dissertations, and grant proposals will draw some distinction d and use word ‘w’ in some way—rather than others. This isn’t a superficial or merely verbal issue, but substantively consequential (cf. Shepherd et al., 2019; Suddaby, 2010; Suddaby et al., 2017). Social scientists’ empirical claims, causal explanations, and theories are dependent on it: their truth-value, accuracy, coefficients’ size and significance, models’ fit, and so on. They depend partly on whether ‘religion’ includes Umbanda, Candomblé, Buddhism, and Confucianism; ‘terrorist organization’ includes the KKK and the Muslim Brotherhood; Trump, Modi, Perón, and Chávez are in the same class; and ‘children’ applies to your friend’s nephews.

Your ‘w’ and d are consequential, too, for your project’s policy implications and import in the public sphere. They partly specify which problem your research addresses and may solve, which problem instances or tokens belong together, and whether different problem-solving sociologists are addressing the same problem (Prasad, 2021a, 2021b). Plus, they affect the value of your laborious data collection, e.g., your comparative and historical dataset about corruption or about the effects of neoliberal reforms around the world. You don’t want it to include reforms that aren’t neoliberal. Obviously, this is a function of the extension of ‘neoliberal’ and the attendant distinction between neoliberal and non-neoliberal reforms.

Social scientists’ disputes over ‘w’ and d have epistemic consequences, but also tangible social and material ones. Consequences for social scientists’ own careers, for their stakeholders, and for society at large.

4. What Experts Can’t Do

Social scientists are experts. Sociologists of race and ethnicity have expert knowledge about the causes and consequences of racial and ethnic discrimination. Economic and business historians about the history of markets, economic institutions, and firms. The social psychology literature on altruism about the conditions under which altruistic behavior is more and less likely to occur. Social scientists are also methodological experts. They devise innovative methods and skillfully employ them.

Experts have good grounds to support their beliefs. Their claims and views are based on their special expertise. On their knowledge, understanding, experience, methods, and evidence. If a journalist, politician, or policymaker has a question about race and ethnicity in the U.S., they’d be well advised to contact a sociologist of race and ethnicity. Experts on X know a large number of facts about X and the causal relations X enters into: what it causes and what it’s caused by.

Yet, experts’ expertise doesn’t comprise what race really is. Nor what ethnicity is. Nor what discrimination is. Nor what altruism is. How could they know? There are no facts to be established about these things’ true being or what they really are. (‘True being’? ‘Really are’? Wrong question!) Contrast with a historical fact, such as ‘Charlemagne died in 814.’ Or with a relation among variables, such as ‘years of formal education is correlated with political participation,’ or ‘smoking tobacco is a cause of cancer.’ In these cases, scientists do get the last word.

Nor can experts determine how the words ‘race,’ ‘ethnicity,’ ‘discrimination,’ or ‘altruism’ should be used. This isn’t a question an individual could sort out. Not even a leading social scientist who’s been working in the field their whole life. It’s not factual or empirical, but a practical reason issue. It’s a practical reason issue for a community, along with its stakeholders. Put differently, the community must make decisions about its practices and norms regarding word uses and distinctions. It must also make decisions about the procedures through which it’ll make decisions about its practices and norms.

‘Practical reason’ sounds like abstruse philosophical jargon, but the basic idea is straightforward. It’s Saturday afternoon, and you’re hanging out with a few friends at Alejandro’s place. You guys intend to cook and eat dinner together around 8:30 p.m. ‘What should we cook for dinner?’ There are better and worse answers to this question. Alejandro and Colo are vegetarian, so pork chops would be a bad idea. Cholent would be bad twice over, as it must be cooked overnight. No doubt, there might be many good answers and there needn’t be one best answer. But the point is that goodness and badness aren’t in the eye of the beholder. Crucially, we’re trying to figure out what course of action would be good for everyone—as opposed to good for any one person. We’re aiming at the common good.

Nobody can answer these questions by themself, whether it’s a group member or an external observer. Why can’t an individual establish, or find out, what to cook for dinner? To be sure, you can gather people’s desires, preferences, proposals, and reasons; what they feel like eating; how hungry they think they’ll be; their cooking skills; and how much effort they’d like to expend preparing and serving the food. But then, how will you integrate these data to produce a decision? What and how to count? People’s preferences, desires, and reasons are diverse, and there’s no incontrovertible integration or aggregation rule. Ariel and Lev have been going through a lot lately, and they’re in financial dire straits, so their enjoyment could be argued to take precedence. As it turns out, their preferences don’t coincide, so it’ll be aggregation time again.

Speaking of finances, the dinner will have to be paid for. Should people chip in in proportion to their financial means? (To each according to their needs, from each according to their abilities.) If so: in proportion to each person’s income, wealth, something else, a composite measure?

A group of friends’ discussing what to cook for dinner isn’t an exchange of information, which may eventually be plugged into a formula. Nor is it about discovering a factual truth. Rather, it’s a collective quest at whose heart is reason-giving. It’s a deliberative process through which they’ll come to see what the best thing to do is, all things considered. It’s a construction of sorts. However, that the outcome is constructed means neither that it’s made up nor that anything goes.

You and your friends plan on cooking and eating dinner together around 8:30 p.m. You don’t know what yet, but something will have to be cooked.

5. Democracy, Unfreedom, Usefulness

How should we use the word ‘well-being,’ ‘capitalism,’ or ‘leadership’ in our social science discipline or literature, so we can go on to make empirical claims about it? How about ‘trust,’ ‘creativity,’ or ‘responsibility’? Should our research projects draw a distinction between empathy and sympathy, Global North and Global South, liberal and conservative beliefs, socialist and social-democratic political programs, happiness and well-being? Should we accept and use the stratigraphic distinction between Holocene and Anthropocene, sociologists’ distinction between moral and etiquette norms, or tripartite classifications of sexual orientations into straight, gay, and bisexual? Or should we reject them and seek alternatives?

These questions are akin to ‘what should we cook for dinner?’; they’re all practical reason questions. Answers to them aren’t truth-apt. Rather, their goal is the good. The good as such and everyone’s good. Everyone’s good is important. Everyone’s reasons and views are equally important. The name of the game is democratic deliberation. Or so I argue in Words and Distinctions for the Common Good.

Nobody’s special. Nor am I, of course. My suggestions are just suggestions. What I’m doing is to extend an invitation. I invite social science communities to get together and make democratic decisions about key words and distinctions—about particular instances of ‘w’ and d and about procedures to make such decisions. The good news is, you’re invited. Less welcome news is that you won’t be free. Individual social scientists aren’t free (appearances to the contrary).

This is how things stand at present. In general, social scientists consider themselves entitled to define key words and draw distinctions as they see fit. Researchers, teams, and labs choose their preferred distinctions and word uses, such that it works for them, their models, papers, projects, and research programs. The idea is something like, it’s my work, so I can sovereignly make these decisions. Sometimes they provide a rationale, which typically invokes the usefulness of ‘w’ or d.

I think this state of affairs is infelicitous. For one, individuals’ presumed sovereignty results in the proliferation of incommensurable word uses, distinctions, and eventually findings and claims (cf. Anvari et al., 2024). Definitions of ‘w’ are cheap, so inflation ensues. They’re stipulative, i.e., not even constrained by natural language. Pretty much anything goes. The usual upshot is misunderstandings and normative and procedural uncertainties. Which, in turn, enable high-status and powerful actors to de facto call the shots.

Moreover, appeals to ‘usefulness’ are elliptical and uninformative. First: you say ‘useful,’ but what is it useful for? The desiderata of social science research are numerous. Parsimony, descriptive accuracy, predictive power, generalizability and scope, tractability, replicability, insight, policy implications, elegance and beauty, the list is very long . . . (Abend, 2023, pp. 252–255). Trade-offs are inescapable: what’s useful to accomplish A and B will thwart C and D. What’s the way forward? I argue that communities—not individuals at their desks—must make choices about desiderata and priorities. Fellow community members! Which aims should be served by our key words and distinctions? Which aims are more important to us, when, and where? ‘W’ and d should be useful for things the collectivity is after. Social scientist Robinson Crusoe may prioritize whatever he pleases. For everyone else, though, research is produced in the context of and is addressed to a community. A scientific community and in certain respects the broader society.

Second: you say ‘useful,’ but do you actually mean useful to you? What’s useful to you, your graduate students, and your lab isn’t useful to me, my graduate students, and my lab. Indeed, it actively harms us and our longstanding research program. The community wishes to advance its interests, to further its good, the common good. It doesn’t care about your career or mine or anybody else’s. So, ‘it’s useful to me, so my models work and my paper gets published’ isn’t a valid reason. Self-interest is beside the point. ⁵

Third: you say ‘useful,’ but what costs are incurred? Choices about ‘w’ and d have epistemic and non-epistemic costs. What do they put aside, obscure, leave out of the picture? What about opportunity costs? What’s the set of options among which choices were and are made? To what extent are costs unequally distributed, so only certain people will have to pay the bill?

Roy Suddaby (2014, 2024) has strong views about what theory is and isn’t. ⁶ My book shows why such views aren’t truth-apt; Suddaby can’t be correct in the way he intends to. However, he might well make suggestions about the uses of the word ‘theory,’ and these can be good or bad, in the same sense that ‘let’s make pork chops!’ or ‘how about cauliflower soup?’ can. This is all up to the community and its democratic deliberations. It may consider Suddaby’s suggestions, their epistemic and moral payoffs (useful for what?), who would be positively and negatively affected by them (useful to whom?), and in particular whether the main overall beneficiary would be the common good.

6. Equality And The Common Good

You might be thinking: deliberative democratic processes are nice in theory, but their Achilles’ heel is who gets to participate, whence the criteria for inclusion and exclusion. And insofar as arguments refer to a ‘community,’ who’s a community member, how to decide who’s a community member, how to decide how to decide who’s a community member. . . Point well-taken. This is admittedly a can of worms, actually more than one, which I have to skirt here.

As a first pass, though: the democratic process will include everyone and everything involved in and affected by research in which ‘w’ or d are central. People who do research for a living, but also various stakeholders; human, but also institutions, practices, and the environment. The claim is twofold. First, all of them will participate as equals. Second, the good of all of them counts; it matters how everyone and everything will fare. Committing to democratic procedures means no special privileges for high-status scholars, renowned professors, journal editors, and award and recruitment committee chairs. No more authority to well-to-do universities in the Global North than underfunded universities in the Global South. No more attention and deference to English-speaking participants than to speakers of other languages. No status differences: holders of fancy endowed chairs should get as much of a say as untenured assistant professors and inexperienced students. It’s not about how much you know about a topic or how talented a social scientist you are. Political scientists don’t get additional votes in democratic elections either; they’re just like any other citizen.

It doesn’t follow that communities’ democratic deliberations should disregard expertise. Experts can play productive advisory roles. They can supply pertinent facts, quantitative and qualitative information, insightful analyses and interpretations, and well-supported causal relations and predictions. “Look, if your community uses ‘w’ and accepts d in these situations, you should expect the following effects.” Such knowledge will improve our collective conversations, deliberations, decisions, strategies, and actions. But that’s how far experts should go. At the end of the day, decisions about words and distinctions will be democratically made by us all. Working together. Oriented toward the common good.

See you all at 5:30 pm in the agora.