Let’s Cancel Habitus Before It Explains Something Else
Few theoretical concepts in sociology have enjoyed as much adoration as habitus. Its definition (“structured structures predisposed to serve as structuring structures”) is like an evil tongue twister for poor grad students socializing into the discipline.1 Its difficulty notwithstanding, however, the term virtually functions as a theoretical primitive for cultural sociology: it’s foundational (with every reviewer asking for it every single time), flexible, and, well, largely unfalsifiable.
1 Stephen Vaisey tried to “translate” the passage from The Logic of Practice here several years ago. It looks way better but I am just going to say that it’s still wildly terrifying.
A recent discussion in the American Sociological Review reminded me that we are, somehow, still acting as if habitus is doing proper theoretical work. I want to argue that it actually is not.
The ASR Debate on “Religious Habitus”
A recent article by Horwitz, Matheny, Laryea, and Schnabel (2022, from now on HMLS) brought habitus back into the discussion of cultural sociology, by arguing that girls raised in Jewish religious subcultures acquire a specific “religious habitus” that promotes elite educational attainment. The article explicitly claimed, in various instances, that
“religious subcultures also shape the habitus a child acquires [or] develop[s] within their family contexts [… which in turn] shapes aspirational differences (338).
The core of their argument is that culture matters for educational and professional attainment, and that habitus plays a key causal role in generating inter-individual differences.
More recently, Brady, Luft, and Zuckerman-Sivan (2025, alternatively BLZ) criticized the HMLS by asking whether we can really attribute group-level educational outcomes to cultural dispositions without specifying those exact cultural dispositions.2 If habitus is going to do any “explanatory” work, their argument goes, then it needs to specify portable cultural mechanisms that explicitly says what it is that makes culture “explanatory.” If not, we risk explaining variation by appealing to group identity itself, rather than identifying the precise mechanisms generating those differences.
2 A large part of the critique centers on class measurement and methodological choices, though I bracket them here for brevity.
Their core point, for the discussion of habitus, is this:
To have confidence in any claim that cultural socialization and internalization matter for attainment, it is necessary to specify the beliefs, desires, and skills that are associated with the parents’ culture [and] to specify that their children share these habitus features in common with their parents (10, emphasis added).
Put differently, while “habitus,” in itself, is significant, we need to make sure that we actually specify what concrete beliefs, desires, or skills we are talking about.
The Trouble with Spooky Concepts
Let’s pause for a second and ask the obvious question: what does even habitus do?
The problem is not that it tries to capture something irrelevant. In fact, habitus invokes exactly the kinds of things we should care about in culture: beliefs, preferences, and endowments.
The problem is that habitus does not really precisely specify what those beliefs, preferences, and endowments might be. Most of the time, we use it as a catch-all term for dispositions and quietly move on. The result is what I would call a spooky concept: a latent construct that supposedly shapes outcomes from behind the curtain, but never quite showing up for a proper greeting.
 |
|---|
| Bourdieu Reacting to Godard |
And the most important trouble with spooky concepts is that they often operate as explanatory placeholders. They stand in for concrete causal mechanisms we suspect are important, but have not quite identified, theorized, or properly tested. In that sense, they are relatively comforting. They allow us to say that something matters, without having to say how it matters. They create the illusion of causality while evading the responsibilities of specifying a full-fledged model.
To be clear, spooky concepts aren’t always bad. They sometimes serve a productive role in theory building by giving us a provisional placeholder. For instance, a “field” can be a useful construct if it offers a relational image of a cultural domain, so that we can orient our attention to certain practices until we specify the actual details. Likewise, utility in economics can function as a black box when we are agnostic about what, exactly, individuals are trying to maximize. The problem is when we start pretending that these placeholders are explanatory after all. That’s when a term like habitus starts to degenerate—when a tolerable theoretical placeholder becomes suddenly alive.
Making Culture Work as Explanation
To be fair, HMLS makes a strong effort to concretize habitus, more so than what BLZ claims: they point to “self-conceptions” (what girls want, how they see themselves, and the importance of being a mother or having a career). That is useful. Yet, if that is the argument (i.e., kids’ self-conceptions matter), then let’s just talk about those things. There is no need to smuggle the spooky machinery of habitus into this discussion, whose theoretical weight only comes at the expense of clarity.
One good example here comes from Vaisey (2010), who uses the same data as HMLS to evaluate whether educational aspirations motivate kids to have higher attainment. In a sense, the paper does not need Bourdieu at all,3 as “educational aspirations” is in itself a measurable, portable, and specific cultural mechanism that needs no placeholder. We thus know what we are really arguing about.
3 He is mentioned 19 times in the article, though as Vaisey said in another occasion, he is like a shield for sociologists.
In short: we need to decompose culture. To make sense of how it works, we need to stop invoking latent spooky constructs. Instead, as BLZ notes, we need portable and convertible mechanisms.
 |
|---|