Beliefs – Page 2 – Culture, Cognition, and Action (culturecog)

Bourdieu as a (Hetero)phenomenologist

Toward the beginning of Pierre Bourdieu’s newly published 1999-2000 lectures from the College de France (Manet: A Symbolic Revolution), a perplexed art student asks Bourdieu the following question, one that the verbatim lectures show to have worried and preoccupied the great theorist of practice:

Your recent conferences have astonished me. For when you speak of developing a theoretical approach to art, taking note not of the intentions of the artist but of his dispositions, this implies that the artist cannot be the author of a theoretical work, as Kandinsky was, because he cannot be conscious of the dispositions that he has acquired unconsciously. As an artist, I find this point of view difficult to accept, for I see in it a risk of alienation that is heavy with consequences and, in this case, must I conclude that I am unable to write a thesis? (Bourdieu 2017: 60)

This question comes after Bourdieu makes claims like the 19th century French artist Edouard Manet included “explosive” qualities in his work “without necessarily being aware of it” (27), that if we were to ask about any Manet painting “Did Manet consciously want to do that and did he premeditate what he put into that painting?” the answer must be an unambiguous “no” (45), and finally, as Bourdieu claims, “What I am describing here is … not the conscious mind of the painter … [but] a painter who finds himself practically engaged with the creation of his picture. He has, then, a practical intention, which is not at all his conscious, premeditated intention” (49-50). No wonder the student was perplexed.

The Manet that appears in Bourdieu’s lectures is not a great conversationalist and really has no unique insight into the nature of his craft. Truly as dumb as a painter, as the old saying goes. If, on a given day, you were to ask Manet “what are you doing?” he would very likely give a most honest reply: “I’m painting a picture.” As Bourdieu continues, “He might even have said a little more, for example: ‘I want to paint something that’ll be a work of art. I want to show that you can make something modern out of a classical model.’” Bourdieu believes that a similar boring answer would be given by Zola or [fill in the blank] writer of renown, or really any one of us as we engage in any activity (I am writing a blog post. I’m not quite sure what it is about. I believe something about Bourdieu). “But that does not make [Manet or Zola] or us either totally unconscious automata, or perfectly lucid subjects” (2017: 49).

What is so interesting about these lectures, and other of Bourdieu’s late work on art (1996), is that here we see the application of practice for the purposes of phenomenology, seeking a reconstruction of direct experience, but which I want to claim fits more the profile of a true heterophenomenology (Dennett 1991; 2003), which I’ll explain. Bourdieu wants to put himself and his readers in the artists’ shoes. He wants to know what it was like to be in the world from the “artist’s point of view,” confront the pragmata they confronted, their problems and their solutions for those problems, which he ultimately claims is the legacy they leave behind for other practitioners in a field to recapitulate (e.g. “Beethoven’s solution” or “Manet’s solution”). An opus infinitum, as Bourdieu confesses, and on several occasions in the lectures he admits his embarrassment at what does not know and never will. Yet there is merit in the task, as he explains, especially when the topic is art, or “pure practice without theory” (as Durkheim said).

Heterophenomenology is a coinage of the philosopher Daniel Dennett and my argument is that it is particularly useful for understanding exactly what Bourdieu is up to in this work, as a particularly relevant application of practice. As Dennett claims, heterophenomenology is, quite basically, the “phenomenology of another not oneself” (2003: 19). It is therefore a third-person accounting scheme (e.g. “Manet does”), but one that takes the “first person point of view [e.g. “I did”] as seriously as it can be taken.” What this means is that a heterophenomenologist (mouthful, hereafter HPist) understands that answers to to questions that, seemingly, give access to dimensions first-person experience (“What were you thinking when you painted X?”) are actually third-person investigations that invoke a person’s verbal capacity to link an internal state to a proposition (“I was thinking that …”). An HPist does not dismiss that kind of data, but neither does she take it at face value.

Rather, the job of the HPist is comprehensive. It involves generating a “fictional world … populated with all the images, events, sounds, smells, hunches, presentiments and feelings that the subject sincerely believes to exist in his or her (or its) stream of consciousness. Maximally extended, it is a neutral portrayal of exactly what it is like to be that subject” (Dennett 1991: 99).

The catch is that even though the HPist trusts the person, and “maintains a constructive and sympathetic neutrality” toward what people say about their experience and their action, this does not make those people the authority on the truth of the experience. What the world is like to them is at best an “uncertain guide” to what is going on for them as they experience and act. But neither does this mean that they are zombies (Dennett) or unconscious automata (Bourdieu). The point is that the HPist tries to compile a definitive description of the world according to subjects, but uses everything in that description as primary data, part of a “fictional world,” that is just the beginning of the analysis. It is what must be explained.

Dennett (2003) demonstrates what this using the following example from a famous experiment by Roger Shepard and Jacqueline Melzer (1971). Shepard and Melzer asked subjects whether the two drawings below (Figure 1) were of different objects or just different views of the same object. Nearly every subject reported that they solved the puzzle by rotating one of the figures in their “mind’s eye” to make a comparison. But were they really doing this mental rotation? If we were to ask a hundred people (including me!), they would probably all say that, yes, I mentally rotate the object on the right in order to answer that, indeed, these are two different views of the same object. The phenomenologist would be satisfied with this. The HPist, however, is more intrigued by “mental rotation” as a belief about conscious experience. She willfully submits that what this experience is like for subjects could have nothing to do with what is going on for subjects as they have this experience. But we first need to know that “mental rotation” is what the experience is like for them (see Pylyshyn 2002; Foglia and O’Regan 2016 for literature on mental imagery).

Figure 1

Dennett’s HPist strategy bears a none too faint resemblance to the method Bourdieu uses in answering the art student and in his remarkable analysis (68-72) of Manet’s painting Luncheon on the Grass (1863). Here, he attempts to put himself and his audience in Manet’s “historical place … a kind of imaginary reconstruction [of] Manet’s point of view as he was engaged in producing the Luncheon” (2017: 69). Thus, what Manet does here is basically create a tableau vivant, those still life paintings of groups of people arranged in some kind of scene that were so fashionable under Napoleon III. But he does something more than this. Bourdieu’s analysis is too detailed (even though he declares his shame about lacking the “necessary competence to do it”) to reproduce in its entirety. I’ll just point to two examples that resemble a heterophenomenology, ones that relate specifically to mental imagery.

First, Manet’s “symbolic revolution” is found in microcosm in the Luncheon because he had placed himself in an impossible situation in relation to his models: “he got [them] to adopt a classical pose, but through [their] clothing and his pictorial manner, he gave [them] modern connotations” (2017: 70). Did he intend to do this? Bourdieu says no. And to build that case he says that Manet effectively “pictured in his mind’s eye” a Venetian painting that he had seen at the Louvre (Marcantino Raimondi The Judgment of Paris … people in the bottom right corner).

Second, the woman in the background is a mystery. The kind of wallpaper effect of the woman suggests Japanese art as a model. But Bourdieu argues that Manet likely had a work of “[Jean-Antoine] Watteau in mind” (maybe the Feast of Love, that woman in the righthand background) someone he admired a great deal. He puts this woman in as a “nod to Watteau … [that] closes the triangle of the three figures … She is treated differently: the brushstrokes are light, as in the style of Watteau” (71).

Luncheon on the Grass, Manet, 1863

The Judgment of Paris, Raimondi ca. 1510-1520

The Feast (or Festival) of Love, Watteau 1718-19

As Bourdieu finishes this “very strange exercise” he feels ashamed doing it, because he does not have the necessary competence. The problem is that those who do have the competence (art historians) don’t think of doing it and remain instead in the role of “interpreter, observer and analyst.” The question here is not a matter of “influence.” It is rather the HPist’s question that what painting the Luncheon was like for Manet gives no special insight into what was going on for Manet as he painted.

According to Bourdieu, what was going on for Manet was practice. This is particularly true because Manet was really never one to serve as Cartesian observer of his own interior process and usually gave boring, trite, cliche answers about his revolutionary work, including to Zola when Manet was doing his portrait. He seemed to be a zombie when he spoke about painting, but of course he was anything but a zombie. That in itself is data to the HPist, as it suggests a dimension of experience running so far past what we can access simply by recording verbalized beliefs about that experience.

At the end of this long examination, motivated by the art student who is despairing at the thought that his first-person experience may not have complete authority, Bourdieu finally gets around to answering him in the best way he can, with a kind of practical charity principle:

[Manet] was not entirely sure of what we wanted to do: he had a vague plan … And it was in the process of doing it that he found out what he wanted to do, as we do ourselves when we do something … To reply once again to the question put to me, which has caused me to slow down and fill out my argument: one can be very intelligent with one’s body, without using language, of course, when one is a dancer or pianist. Having said this, there is the particular problem of the executant as opposed to the person who produces his own text, or his own work of art. So there is an intelligence of the body … Painters know how to adopt towards painting a viewpoint that is a practical understanding; they have a practical perception which is based on know-how … they understand a savoir-faire as savoir-faire, and don’t write lectures on it (73)

References

Bourdieu, Pierre. (2017). Manet: A Symbolic Revolution. London: Polity.

Bourdieu, Pierre. (1996). The Rules of Art. Stanford: Stanford UP.

Dennett, Daniel. (1991). Consciousness Explained. Boston: Back Bay

Dennett, Daniel. (2003). “Who’s on first? Heterophenomenology explained.” Journal of Consciousness Studies 10: 19-30

Foglia, Lucia and Kevin O’Regan. (2016). “A New Imagery Debate: Enactive and Sensorimotor Accounts.” Review of Philosophy and Psychology 7: 181-196.

Pylyshyn, ZW. (2002). “Mental Imagery: In Search of a Theory.” Behavioral and Brain Sciences 25: 157-182.

Shepard, Roger and Jacqueline Melzer. (1971., “Mental rotation of three-dimensional objects” Science 171: 701–3.

The Decision to Believe

As noted in a previous post, there are analytic advantages with reconceptualizing the traditional denizens of the folk-psychological vocabulary from the point of view of habit theory. So far, however, the argument has been negative and high-level; thinking of belief as habit, for instance, allows us to sidestep a bunch of antinomies and contradictions brought about by the picture theory. In this post, I would like to outline some positive implications of recasting beliefs as a species of habit. However, I will begin by discussing other overlooked implications of the picture theory and then (promise) move on to some clear substantive implications of the habit conception.

As noted before, the picture theory of belief is part of a more general set of folk (and even technical) conceptions of how beliefs work. I have already noted one of them and that is the postulate of incorrigibility: If somebody assents to believing p, then we presume that they have privileged first-person knowledge as to this. It would be nonsensical (and socially uncouth) for a second person to say to them “I know better than you on this one; I don’t think you believe p.” Folk Cartesianism thus operates as a philosophical set of tenets (e.g. the idea we have privileged introspective and maybe even non-inferential access to personal beliefs), and as a set of ethnomethods to coordinate social interaction (accepting people’s claims they believe something when they tell us so without raising a fuss).

I want to point to another, less obvious premise of both folk and technical Cartesianism. This is the notion (which became historically decisive in the Christian West after the Protestant Reformation) that you get to choose what you believe. Just like before, this doubles as a philosophical precept and as an ethnomethod used to organize social relations in doxa-centric societies (Mahmood 2011). If you get to choose what you believe, and if your belief is obnoxious or harmful, then you are responsible for your belief and can be blamed, punished, burned at the stake and so on. As the sociologist David Smilde has also noted, there is a positive version of this implication of folk Cartesianism: if the belief is good for you (e.g. brings with it new friends, behaviors, resources) then we should expect you (under the auspices of charitable ascription) to choose to believe it. However, the weird prospect of people believing something not because they find its truth or validity compelling but because of instrumental reasons raises its ugly head in this case (Smilde 2007, 3ff; 100ff).

The idea of choosing to believe is not as crazy as it sounds. At least the negative of it, the idea we could bring up a consideration (let’s say a standard proposition) and withhold belief from it until we had scrutinized its validity was central to the technical Cartesian method of doubt. Obviously, this requires that we have some reflective control over our decision to believe in something or not while we consider it, so in this respect technical and folk Cartesianism coincide.

As Mike and I discuss in the 2015 paper, rejecting the picture theory (and associated technical/folk Cartesianism) of belief makes hash of the notion of “choosing to believe” as a plausible belief-formation story. Here the strict analogy to prototypical habits helps. Consider a well-honed habit; when exactly did you choose to acquire it? Now even if you made a “decision” to start a new training regimen (e.g. Yoga) at what point did it go from a decision to a habit? Did that involve an act of assent on your part? Now consider a traditional belief stated as an explicit linguistic proposition you claim to believe (e.g. “The U.S. is the land of opportunity”). When did you choose to believe that? We suggest, that even a fairly informal bit of phenomenology will lead to the conclusion that you do not have credible autobiographical memories of having “chosen” any of the things you claim to believe. It’s as if, as Smilde points out, the original memory of decision is “erased” once the conviction to believe takes hold.

We suggest that the apparatus of erased memories and decisions that may or may not have taken place is an unnecessary outgrowth of the picture theory. Just like habits, beliefs are acquired gradually. The problem is that we take trivial (in the strict sense of trivia) encyclopedic statements (e.g. Bahrain is a country in the middle east) as prototypical cases of belief. Because these could be acquired via fast memory binding after a single exposure they seem to be the opposite of the way habits are acquired. However, these linguistic-assent to trivia beliefs are analytically worthless because it is unlikely that if there’s anything like belief that plays a role in action, it takes the form of linguistic-trivia beliefs. That we believe (no pun intended) that these types of propositions are “in control” of action is itself also an unnecessary analytic burden produced by the picture theory.

Instead, as noted before, a lot of our action-implicated beliefs are clusters of dispositions and not passive acts of private assent to linguistic statements. However, trivia-style beliefs capable of being acquired via a single exposure are the main stock in trade of both the folk idea of belief and the intellectualist strand of philosophical discussion on the topic. Thus, they are important to deal with conceptually, even if, from the point of view of the habit theory they represent a degenerate case since from this perspective, repetition, habituation, and perseverance is the hallmark of belief (Smith and Thelen 2003).

That said, what if I told you that the folk-cartesian notion of deciding to believe is inapplicable even in the case, of trivia-style one shot belief? This is the key conclusion of what is now the most empirically successful program on belief formation in cognitive psychology. The classic paper here is Gilbert (1991), who traces the idea back to Spinoza, although the subject has been revived in the recent efflorescence of work in the philosophy of belief. See in particular Mandelbaum (2014) and Rott (2017). This last notes that this was also a central part of the habit-theoretic notion of belief shared by the American pragmatists.

When it comes to one shot propositions, people are natural born believers. In contrast to the idea that conceptions are first considered while withholding belief (as in the Cartesian model) what the evidence shows is that mere exposure or consideration of a proposition leads people to treat as a standing belief in future action and thinking. Thus, people seem incapable of not believing what they bring to mind. While this may seem like a “bug” rather than a feature of a cognitive architecture, it is perfectly compatible with both a habit-theoretic notion of belief, and a wider pragmatist conception of mentality, of the sort championed by James, Dewey, and in particular the avowed anti-Cartesian C. S. Peirce. Just in the same way that every action could be the first in a long line that will fix a belief or a habit, the very act of considering something makes it relevant for us without the intervention of some effortful mental act of acceptance.

So just like you don’t know where your habits come from, you don’t know where your “beliefs” (in the one-shot trivia sense) come from either. The reason for this is that they got in there without having to get an invitation from you. In the same way, an implication of the Spinozist belief-formation process is that the thing that requires effort and controlled intervention is the withdrawal of belief (which is difficult and resource demanding). This links up the Spinozist belief-formation story with dual process models of thinking and action (Lizardo et al. 2016).

This is also in strict analogy with habit: While lots of habits are relatively easy to form (whether or not desirable) kicking a habit is hard. Even the habits that seem to us “hard” to form (e.g. going to the gym regularly) are not hard to form because they are habits; they are hard to form because they have to contend with the existence of even stronger competing habits (lounging at home) that will not go away without putting up a fight. It is the dissolution of the old habit and not the making of the new one that’s difficult.

So with belief. Beliefs are hard to undo. Once again, because we mistakenly take the trivia one-shot version of belief as the prototype this seems like an exaggeration. So if you believed “Bahrain was a country in Africa” and somebody told you “no, actually it’s in the Persian Gulf” it would take some mental energy to give up the old belief and form the new one, but not that much; most people would be successful.

But as noted in a previous entry, most beliefs are clusters of habitual dispositions, not singleton spectatorial propositions toward which we go yea or nay. So (easily!) developing these dispositional complexes in the context of, let’s say, a misogynistic society like the United States, would mean that “unbelieving” the dispositional cluster glossed by the sentential proposition “women can’t make as good as leaders as men” is not a trivial matter. For some, to completely unbelieve this may be close to impossible. This is something that our best social-scientific theories (whether “critical” or not) have yet to handle properly because their conception of “ideology” is still trapped in the picture theory (this is a matter for future posts).

Beliefs, as Mike and I noted in a companion paper (Strand and Lizardo 2017), have an inertia (which Bourdieu referred to as “hysteresis”) that makes them hang around even after a third person observer can diagnose them as “out of phase,” or “outmoded.” This is the double-edged nature of their status as habits; easy to form (when no competing beliefs are around) and easy to use (once fixed via repetition), but hard to drop.

References

Gilbert, Daniel T. 1991. “How Mental Systems Believe.” The American Psychologist 46 (2). American Psychological Association: 107.

Lizardo, Omar, Robert Mowry, Brandon Sepulvado, Dustin S. Stoltz, Marshall A. Taylor, Justin Van Ness, and Michael Wood. 2016. “What Are Dual Process Models? Implications for Cultural Analysis in Sociology.” Sociological Theory 34 (4). journals.sagepub.com: 287–310.

Mahmood, Saba. 2011. Politics of Piety: The Islamic Revival and the Feminist Subject. Princeton University Press.

Mandelbaum, Eric. 2014. “Thinking Is Believing.” Inquiry: A Journal of Medical Care Organization, Provision and Financing 57 (1). Routledge: 55–96.

Rott, Hans. 2017. “Negative Doxastic Voluntarism and the Concept of Belief.” Synthese 194 (8): 2695–2720.

Smilde, D. 2007. Reason to Believe: Cultural Agency in Latin American Evangelicalism. The Anthropology of Christianity. University of California Press.

Smith, Linda B., and Esther Thelen. 2003. “Development as a Dynamic System.” Trends in Cognitive Sciences 7 (8): 343–48.

Strand, Michael, and Omar Lizardo. 2017. “The Hysteresis Effect: Theorizing Mismatch in Action.” Journal for the Theory of Social Behaviour 47 (2): 164–94.

Are the Folk Natural Ryleans?

Folk psychology and the belief-desire accounting system has been formative in cognitive science because of the claim, mainly put forth by philosophers, that it forms the fundamental framework via which everybody (philosopher and non-philosopher alike) understands human action as meaningful. Both proponents of some version of the argument for the ineliminable character of the folk psychological vocabulary (Davidson, 1963; Fodor, 1987), and critics that cannot wait for its elimination by a mature neuroscience as an outmoded theory (Churchland, 1981) accept the basic premise; namely, that when it comes to action understanding, folk psychology is preferred by the folk. The job of philosophy is to systematize and lay bare the “theoretical” structure of the folk system (to save it or disparage it).

In a fascinating new article forthcoming in Philosophical Psychology, Devin Sanchez Curry, tries to challenge this crucial bit of philosophical common wisdom, which he refers to as “Davidson’s Dogma” (Sanchez Curry acknowledges that this might not be exegetically strictly true of Davidson’s writings, although it is true in terms of third-party reception and influence). In particular, Sanchez Curry hones in on the claim that the folk use a “theory” of causation to account for action using beliefs: Essentially the idea that beliefs are inner causes (the cogs in the internal machinery) that produce action when they interact with other beliefs and desires. This is the subject of a previous post.

Sanchez Curry, rather than staying at the purely exegetical or conceptual analysis level, turns to the empirical literature in psychology on lay belief attribution to shed light on this issue. There he notes something surprising. There’s little empirical evidence that the folk resort to a belief-desire vocabulary or to a theory of these as inner causes (cogs and wheels in the internal machinery) of action. Going through the literature on the development and functioning of “mindreading” abilities, Sanchez Curry shows that the primary conclusion of this line of work is that the explicit attribution of representational (e.g. “pictures in the head”) versions of belief is the exception, not the rule.

Instead, the literature has converged (like many other subfields in social and cognitive psychology) on a dual systems/process view, in which the bulk of everyday mindreading is done by high capacity, high-efficiency automatic systems that do not traffic in the explicit language of representations. Instead, these systems are attuned to routine behavioral dispositions of others and engage in the job of inference and filling-in of other people’s behavior patterns by drawing on well-honed schemata trained by the pervasive experience of watching conspecifics make their way through the world. Explicit representational belief attribution practices emerge when the routine System I process encounter trouble and require either observers or other people to “justify” what they have done using a more explicit accounting.

As Sanchez Curry notes, the evidence here is consistent with the idea (which I alluded to in a previous post) that persons may be “natural Ryleans” but that the Rylean (dispositional) action-accounting system is so routinized as to not have the flashy linguistic bells and whistles of the folk psychological one. This creates the illusion that there’s only one accounting system (the belief-desire one), when in fact there are two, it is just that the one that does most of the work is nondeclarative (Lizardo, 2017), while the declarative one gets most of the attention, even though it’s actually the “emergency” action-accounting system, not the everyday workhorse.

As Sanchez Curry also notes, evidence provided by “new wave” (post-Heider) attribution theorists show that the explicit (and actual) folk psychological accounting system even when activated, seldom posits beliefs as “inner causes” of behavior. Instead, when people enter the folk-psychological mode to explain puzzling behavior that cannot be handled by System I practical mindreading, they look for reasons, not causes. These reasons are holistic, situational, and even “institutional” (in the sociological sense). There are “justifications” that will make the action meaningful while saving the rationality of the actor, given the context. They seldom refer to internal machineries or producing causes. We look for justifications to establish blame, to “make sense” (e.g. “explain”) or “save face” not to establish the inner wellsprings of action. So even in this case the folk are natural Ryleans and focus on the observables of the situation and not the inner wellsprings. This means, that the “theory” of folk psychology is a purely iatrogenic construction of a philosophical discourse on action that plays little role in the actual attributional practices of the folk: Folk psychology in the Davidsonian/Fodorian sense turns out to be the specialized construction of an expert community.

One advantage of this account is that it solves what I previously referred to as the “frame problem” faced by all “pictures in the head” as causal drivers of action. The problem is that the observer has to pick one of a myriad of possible pictures as the “primary” cause for the action. But there is no way to make this selection in a non-arbitrary way if we are stuck with the “inner cause” conception. In the Rylean conception, the “reason” we attribute will depend on the pragmatics and goals of the reason request. Are we seeking to establish blame? Make sense of a puzzle? Save the agent’s face? Make it seem like they are devious?

These arguments have several important implications. The most important one is that mostly, nobody is imputing little world pictures to other people to explain their action, empathize, or even predict or make inferences as to what they will do next. Dedicated, highly trained automatic systems do the job when people are behaving in “predictable” ways. No representations required there (Hutto, 2004). When this action-tracking system fails, we resort to more explicit action accountings, but more accurately we resort to the placing of strange or puzzling action in a less puzzling context. Even here, this is less about getting at occult or inner well-springs than of trying to construct a “reason” why somebody might have acted this way that makes the action less puzzling.

References

Churchland, P. M. (1981). Eliminative Materialism and the Propositional Attitudes. The Journal of Philosophy, 78(2), 67–90.

Davidson, D. (1963). Actions, Reasons, and Causes. The Journal of Philosophy, 60(23), 685–700.

Fodor, J. A. (1987). Psychosemantics: The Problem of Meaning in the Philosophy of Mind. MIT Press.

Hutto, D. D. (2004). The Limits of Spectatorial Folk Psychology. Mind and Language, 19(5), 548–573.

Lizardo, O. (2017). Improving Cultural Analysis: Considering Personal Culture in its Declarative and Nondeclarative Modes. American Sociological Review, 0003122416675175.

The Ascription of Dispositions

“It’s what you do” is the title of a wildly successful advertising campaign by the American insurance company GEICO. In each spot, we see either a “type” (people in a horror movie, a camel, a fisherman, a cat, a Mom, a golf commentator) or people familiar enough to the intended middle-aged audience of insurance buyers to be considered types (mainly 80s and 90s musical acts like Europe, Boyz to Men, or Salt-N-Pepa) doing things they “typically” do. These things are either out of place, annoying, rude, or irrational and thus funny within the context of the “frame” (an office, a restaurant, etc.) in which they are presented.

For instance, in a viral spot, Peter Pan shows up at the 50th-anniversary reunion to remind everybody else of how young he is (and how old they are). The voiceover reads: “If you are Peter Pan you stay young forever. It’s what you do.” In another one, a poor guy slowly sinks to his death in quicksand, while imploring a nearby cat to get help. The cat of course just licks her paws without looking at him: “If you are a cat you ignore people. It’s what you do.”

The commercials are of course funny due to the specificity of each setup. I want to suggest, however, that they may carry a more general lesson. Perhaps they strike us as noticeable (and thus humorous) because they use an action accounting system that is inveterately familiar but that we usually keep in abeyance. In fact, it is so familiar that it requires the odd situations in the GEICO commercials to make it stand out. This action accounting system, rather than relying on “belief-desire” ascriptions, points to typicalities in behavior patterns as their own justification. Thus the template “If are you X, you do Y, it’s what you do” may hold the key for prying ourselves loose of belief-desire talk.

In a previous post, I argued that the belief-desire accounting system commits us to a model in which action is driven by “little pictures in the head.” An entire tradition of explaining action by making recourse to the “ideas” that “drive” it is based on such a strategy (Parsons, 1938). This is not as innocent of a move as it may seem. Pictures in the head are entities assumed to have specific properties (e.g. representational, content-ful, and casually power-ful) that ultimately need to be cashed in in any scientific account of action. This may not be possible (Hutto & Myin, 2013).

In a follow-up post, I noted that, even if taking an ontology-neutral stance (Dennett, 1989), the ascription of belief from a third-person perspective is not an unproblematic practice either. Sometimes, different pieces of evidence (e.g. what people claim to believe) clashes with other pieces of evidence (what people do) to make belief ascription a problematic affair. The point there was that sometimes, even in our routine ascription behavior, we don’t treat beliefs as purely pictures. Actions matter too and sometimes we may conclude that what people really believe has nothing to do with the pictures (e.g. propositions) that they claim to have in their head.

So maybe our ascription practices and our action accounting systems can go beyond the usual belief-desire combo of folk psychology. This is important because one of the reasons why the claim that belief is a kind of habit might be problematic to some is that it doesn’t seem to fit any intuitive picture of the way we keep track and explain other people’s action (or our own). Here I will build some intuition for the claim that there are other ways of “explaining” action that doesn’t require the ascription of picture-like constructs that drive action. These are also compatible with the idea that beliefs are a kind of habit. Moreover, these are already ascription practices that we follow in our everyday accountings; it’s just that they are too boring to be noticeable.

The most obvious way in which we sometimes explain action without using the language of belief is to talk about somebody’s tendencies, propensities, inclinations, etc. Just like in the GEICO commercials, instead of ascribing beliefs and desires we simply point to the action as being “typical” of that doer. In the philosophy of action, at least since Ryle (2002), this is usually referred to as using a “dispositional” language. Just like ideas, dispositions are sufficient “causes” of the action they help account for. So going back to the example of Sam the fridge opener: Instead of saying that Sam opened the fridge because they believed there was sandwich inside, we can say: “Sam tends to open the fridge when they are hungry. It’s what they do.” This is a way of accounting for the action that does not resort to the ascription of world pictures. Instead, it points to a regularity or a tendency in Sam’s action that is noted to occur under certain (usually typical) conditions.

These kind of dispositional ascriptions are fairly common. In fact, they are so common they are kind of boring. Maybe they stand out less than the usual belief-desire combo of folk psychology because they are seldom used for action justification, rationality ascription, or storytelling. A serial killer who attempted to mount a defense based on the claim that “killing is just what I do” would be the subject of a short trial. In this sense, dispositional ascriptions are gray and drab (in spite of their strict accuracy) while the trafficking in (and sometimes the clash between) beliefs and desires just tell a more interesting story (in spite of their inherently speculative nature). But the pragmatics of belief-desire language use or their mnemonic advantage should not dictate their use in social-scientific explanatory projects. Dispositions have an advantage here because they commit us to a less inflationary ontology compatible with the naturalistic commitments of cognitive neuroscience.

As Schwitzgebel (2010) has argued, the dispositional approach can be extended to account for our ascription of the usual “attitudes” whether propositional (like beliefs and desires) or not. This also points to a solution to the ascription problems that arise when sayings (or phenomenological experience) does not match up with action. In contrast to pro-judgment (which favor subjective certainties and verbal reports) or anti-judgment (which favors action) views, the idea is to think of the global entity (e.g. the “belief” or the “desire”) as a cluster of dispositions. So rather than any one member (the saying or the doing) being decisive in our ascription, they all count (although we may weigh some more than others). This means that sometimes, the matter of whether somebody “believes” P will be undecidable (the cases of implicit/explicit dissociation) because different dispositions point in different directions.

The bigger point, however, is that all dispositional ascriptions have the structure of “habituals” (Fara, 2005). So when we say Sam “believes” P, what we are really saying is that Sam is predisposed to agree that P under a certain broad range of circumstances. But we also say that Sam is likely to act as if P is true, to have certain subjective experiences consistent with the truth of P and so on. In this respect, the “belief” that P is just a cluster of cognitive, phenomenological, verbal, and behavioral dispositions. This cashes in on the insight that “habit” (or disposition) is the superordinate category in mental life and that the other terms of the mental vocabulary fall of as special cases. This also reinforces the point which Mike and I made in the original paper (see in particular 56-57), that the issue is not the elimination of the language of belief and desire (or the other folk mental concepts), but their proper re-specification within a habit-theoretic framework.

Another nice feature of the dispositional ascription approach is that when we ascribe a belief, we no longer have to commit ourselves to the existence or causal efficacy of problematic entities (e.g. world pictures) but point to the usual set of things clear in experience: Actions, linguistic declarations, comportments, moods, etc.). Usually, these hang together and point in the same direction, sometimes they do not. However, whether this hanging together no longer has to result in a contest between heterogeneous entities (e.g. sayings versus doings) but between different species of the same dispositional genus.

Note, however, that picking one disposition in the cluster as the decisive element in an act of ascription is a conclusion that cannot be reached by virtue of a priori methodological policy (such as those privileging doings over sayings or vice-versa). Instead, we need to commit ourselves to an ascription standard combining inference to the best explanation with a coherentist approach: Attitude ascriptions should maximize harmony across the entire dispositional profile. So it would be a mistake, for instance, to select a single disposition (or phenomenal experience, or verbal report) as the criterion for attitude ascription, when there’s an entire panoply of other dispositions pointing in a different direction.

So the issue is not whether there’s a contest between “sayings” and “doings” (Jerolmack & Khan, 2014). Rather, the best tack is taking a tally of the entire dispositional panoply, which may involve lots of tendencies to say, do, and experience into account. Here some sayings might clash against some sayings and some doings against other doings. Whether people strive for consistency across their dispositional profile may be as much of a sociocultural matter (as argued by Max Weber) than an a priori analytic issue. In all, however, what we are confronting are dispositions clashing (or harmonizing with) other dispositions, so in this sense, the analytical task becomes tractable from within a single action vocabulary.

References

Dennett, D. C. (1989). The Intentional Stance. MIT Press.

Fara, M. (2005). Dispositions and Habituals. Nous , 39(1), 43–82.

Hutto, D. D., & Myin, E. (2013). Radicalizing Enactivism: Basic Minds Without Content. MIT Press.

Jerolmack, C., & Khan, S. (2014). Talk Is Cheap: Ethnography and the Attitudinal Fallacy. Sociological Methods & Research. https://doi.org/10.1177/0049124114523396

Parsons, T. (1938). The Role of Ideas in Social Action. American Sociological Review, 3(5), 652–664.

Ryle, G. (2002). [1949], The Concept of Mind. Chicago: The University of Chicago Press,. With an lntroduction by Daniel C. Dennett.

Schwitzgebel, E. (2010). Acting contrary to our professed beliefs or the gulf between occurrent judgment and dispositional belief. Pacific Philosophical Quarterly, 91(4), 531–553.

Ascription Practices: The Very Idea

How do we know what others believe? The answer to this question may seem clear but as we will see it has some interesting hidden complexities. Some of these bear directly on some established policies in social-scientific method.

One obvious answer is that if we want to know what others believe, and these others are language using creatures like ourselves is that we ask them: “Do you believe P?” If the person verbally reports believing P, then we are on safe grounds in ascribing belief P.

So far so good. But let us say a person assents to P but acts in other ways that seem counter to the content of that belief. What to do then? Cases of this sort have become popular grist for reflection in recent work in the philosophy of belief. Spurred by a series of papers by Tamar Gendler (Gendler, 2008a, 2008b), a lively literature has developed on what to ascribe when belief “sayings” (usually referred to as “judgments”) come apart from “doings” (for a sampling see Albahari, 2014; Kriegel, 2012; Mandelbaum, 2013; Schwitzgebel, 2010; Zimmerman, 2007). Some cases are stock in trade and involve people who verbally commit to a belief or attitude but act in contrary ways.

Of most interest for social and behavioral scientists are cases of what are called dissociations between “explicit” or more accurately, direct, measures of a construct (such as a belief or an attitude), which usually rely on self report, and so-called “implicit,” or more accurately, indirect measures of the same construct (Gawronski, Peters, & LeBel, 2008). While the so-called “implicit attitude test” is the most familiar indirect measure, there is an entire family composed of dozens of distinct indirect measurement strategies (Nosek, Hawkins, & Frazier, 2011). The key point is, however, that indirect measures usually rely not on verbal reports but on observations of rapid-fire action or behavioral responses that are assumed not to be under voluntary control.

Dissociations between direct and indirect measures are usually good examples of “sayings” and “doings” coming apart. If these are as common as the literature suggests, then belief (or attitude) ascription problems are also more common than we realize.

So let’s say we observe someone who fits the profile of “Chris the implicit racist.” Chris is a white high school teacher who professes the belief(s) that black people in the United States are no more or less intelligent, violent, or hardworking than white people. Yet systematically in their unguarded behavior (observed via ethnographic classroom observation or via indirect measures of implicit bias collected in the lab) Chris shows a preference for white people (e.g. they discipline black students more harshly for similar offenses; they are more likely to call on white students, etc.) and implicitly associates people with dark skin with a host of negative concepts, including lack of intelligence, proneness to violence, and a weaker work ethic.

What does Chris believe? Considerations of dissociative cases like this lead in two interesting directions. The first is that actions, practices, and behaviors have a non-negligible weight in our belief ascription practices. So judgments aren’t everything. This is important because, in polite company, our everyday practices of belief ascription follow folk cartesianism: That is, belief ascription is fixed by explicit reports of what people say they believe and people have incorrigible knowledge about those personal judgments. We cannot ascribe a belief to a person they profess not to have (alternatively people have veto power over second-person ascription practices).

In philosophy, this is called the “pro-judgment” view on ascription. This stance holds we can only ascribe the belief that P if a person reports they believe P. The actions inconsistent with P (e.g. for Chris, the automatic association of Black people with violence, or their penchant to send black students to detention for minor offenses) are acknowledged to exist but they are just not a kind of belief. They are something else. Gendler (2008a) has proposed that given the recalcitrant existence of these types of counter-doxastic behaviors, we should add a new mental category to our lexicon: “aliefs.” So we can say Chris “believes” blacks are no more violent than whites (as given by their self-report), but “alieves” they are more violent (as given by their performance on indirect attitude measures and their avoiding walking in certain predominantly black neighborhoods).

An alternative view, called the “anti-judgment” view says actions speak louder than words. Or, as indicated by the title of a recent entry in a book review symposium dedicated to Arlie Hochschild’s Strangers in Their Own Land, anti-judgment people say “who cares what they think?” (Shapira, 2017). If Chris walks like P, and quacks like P, then Chris believes P.

I should note that this belief ascription strategy is not that bizarre and that it exists as a “second option” in our commonsense arsenal, even if our first option is folk cartesianism. This should alert you that ascription practices may be very much a matter of socio-cultural tradition and regulation as they are a matter of the usual canons of rationality.

This matters if these same ascription practices are followed mindlessly in social science research. This would be a case of folk practices dictating what should be a matter of social scientific consideration. In this sense, social scientists should care very much if they are pro-judgment folk cartesians, anti-judgment, or something else, as this will bear directly on their conclusions. But we are getting ahead of ourselves. The point is that “trumping” a person’s cartesian incorrigibility by pointing to their inconsistent actions is an available move (both in social science and everyday life) but it is also one that should seldom be undertaken lightly.

Note that the pro-judgment and anti-judgment views are not the only available ones. To fill out the space of options: We may ascribe both P and ~P beliefs (the contradictory belief view). Or we may ascribe a mixture of the inconsistent beliefs (the “in-between” belief view). Or we may say Chris believes neither P nor ~P, but that they vacillate inconsistently between the two depending on circumstances (the shifting view) (Albahari, 2014).

And this is the second thing that inconsistency between sayings and doings highlight. Rather than focusing on goings-on trapped within each person’s cartesian theater, we can now see that beliefs are very much a matter of both actions and practices, both in terms of the believer and the ascriber. In this respect, two consideraations come to the fore.

First, there are the “belief proclamation” practices we are used to (e.g. people verbally saying they believe thus and so) but also the myriad of behaviors and actions that other people monitor and that they use to ascribe beliefs to others. It is these belief ascription practices I wanted to highlight in this post. As noted, they are both a matter of everyday interpersonal interaction, and for my purposes, they are a standard, but seldom commented upon, aspect of every social scientific practice. After all, social scientists (especially those who do qualitative work) constantly ask people what they believe about a host of things (e.g. Edin & Kefalas, 2011; Young, 2006), and are thus confronted with self-reports of people claiming to believe things. These same social scientists may also sometimes have the opportunity to observe these people in ecologically natural settings, which allows them to compare self-reports to doings (Jerolmack & Khan, 2014).

In this post, I will not try to adjudicate or argue for what the best ascription strategy is. I will leave that for a future post. Here I note two things. First, it is clear certain ascription practices have elective affinities with certain conceptions of what beliefs are. For instance, “pro-judgment” views have an affinity with the “pictures in the head” conception of belief. As such, pro-judgment views assume all of the things that such a conception assumes, such as representationalism, the relevance of the truth/falsity criterion and so on. “Anti-judgment” views focusing on action, may be said to be more consonant with some forms of practice theory, as can some other (e.g. “in-between” or “contradictory” views).

Second, belief ascription practices have the familiar duality of being both possible “topics” and “resource” for sociological analysis that fascinated early ethnomethodology. We can be “neutral” about their import for sociological research and study belief ascription practices as a topic. We may ask questions such as under what circumstances people default to folk cartesianism, when they prefer anti-judgment views, when they go “in between” or “contradictory” and so on.

Alternatively we may examine the issue by considering the role of belief ascription practices as a resource for sociological explanation. Are pro-judgment views always effective? Should we go anti-judgment and ignore what people say in favor of their behavior? These are some issues I hope to tackle in future posts.

References

Albahari, M. (2014). Alief or belief? A contextual approach to belief ascription. Philosophical Studies, 167(3), 701–720.

Edin, K., & Kefalas, M. (2011). Promises I Can Keep: Why Poor Women Put Motherhood before Marriage. University of California Press.

Gawronski, B., Peters, K. R., & LeBel, E. P. (2008). What Makes Mental Associations Personal or Extra-Personal? Conceptual Issues in the Methodological Debate about Implicit Attitude Measures. Social and Personality Psychology Compass, 2(2), 1002–1023.

Gendler, T. S. (2008a). Alief and Belief. The Journal of Philosophy, 105(10), 634–663.

Gendler, T. S. (2008b). Alief in Action (and Reaction). Mind & Language, 23(5), 552–585.

Jerolmack, C., & Khan, S. (2014). Talk Is Cheap: Ethnography and the Attitudinal Fallacy. Sociological Methods & Research. https://doi.org/10.1177/0049124114523396

Kriegel, U. (2012). Moral Motivation, Moral Phenomenology, And The Alief/Belief Distinction. Australasian Journal of Philosophy, 90(3), 469–486.

Mandelbaum, E. (2013). Against alief. Philosophical Studies, 165(1), 197–211.

Nosek, B. A., Hawkins, C. B., & Frazier, R. S. (2011). Implicit social cognition: from measures to mechanisms. Trends in Cognitive Sciences, 15(4), 152–159.

Schwitzgebel, E. (2010). Acting contrary to our professed beliefs or the gulf between occurrent judgment and dispositional belief. Pacific Philosophical Quarterly, 91(4), 531–553.

Shapira, H. (2017). Who Cares What They Think? Going About the Right the Wrong Way. Contemporary Sociology, 46(5), 512–517.

Young, A. A. (2006). The Minds of Marginalized Black Men: Making Sense of Mobility, Opportunity, and Future Life Chances. Princeton University Press.

Zimmerman, A. (2007). The Nature of Belief. Journal of Consciousness Studies, 14(11), 61–82.

Are Beliefs Pictures in the Head?

In a recently published piece (Strand & Lizardo, 2015) Mike and I argued that the notion of “belief” if it is to do a more adequate job as a category of analysis in social-scientific research, can best be thought of as a species of habit. I refer the interested reader to the paper for the more detail “exegetical” argumentation excavating the origins of this notion in American pragmatism (mostly in the work of Peirce and Dewey) and European practice theory (mostly in the work of Bourdieu). Here I would like to explore some reasons this proposal may seem to be so counterintuitive given our traditional conceptions of belief.

The fear, to some well-founded, is that substituting the usual notion for the habit notion would cause a net loss, and thus an inability to account for things that would like to account for (e.g. patterns of action that are driven by ideas or thoughts in the head) adequately.

What is the standard notion of belief that the “habit” notion displaces (if not replaces)? The easiest way to think of it is as one in which beliefs are thought to be little “pictures” in the head that people carry around. But what are beliefs pictures of? After all, pictures (even in modern art) usually depict something, however faint. The answer is that they are supposed to be pictures of the world that somehow the person uses to get by.

Because the beliefs are “pictures” (in cognitive science sometimes the word representation is used in this context) they have the representational properties usual pictures have. For instance, they portray the world in a certain way (e.g. under a particular description). In addition, because they are pictures, beliefs have content. That is a belief is always about something (in some philosophical segments, the word “intentionality” is usually brought up here (Searle, 1983)). In this way, some beliefs may be directed at the same state of affairs in the world, but “picture it” in different ways (Hutto, 2013). Finally, and building in on this last distinction, just like pictures claim to depict the world as it is (or at least have a resemblance to it), beliefs can be true (if they portray the world as it is) or they can be false (if the description does not match the world). This truth/falsity relation between the pictures in the head and the world turns out to be crucial for their indispensable job in “explaining” action.

For instance, if somebody opens a refrigerator, grabs a sandwich from it, and eats it, and an outside observer can “explain” the pattern by ascribing a belief to the person. So the person opened the fridge door because they thought (believed) there was a sandwich there. We usually complete this belief-based explanation by adding some kind of motive or desire as a jointly sufficient cause (“they believed there was a sandwich in the fridge and they were hungry”).

But suppose we were to see the same person open the fridge, look around and then go back to the couch empty-handed. This is a different behavioral pattern as before. However, note we can also “explain” this behavior using the same “sandwich” belief mechanism as before. The trick is simply to ascribe a false belief to the person: The imputed picture in the head does not match the actual state of the world. So we can now say, “Sam opened the fridge because they believed there was a sandwich in there and they were hungry.” We attach one more disclaimer: “But Sam was wrong, there was no sandwich.”

This flexibility makes belief-based explanations fairly powerful (they can account for a wide range of behavioral patterns). However, flexibility is also a double-edged sword: Become too flexible and you risk vacuity, explaining everything and thus nothing (see Strand & Lizardo, 2015, pp. 47–48).

Because the belief-desire combo is so flexible (and so pervasive even in our “folk” accounting of each other’s action) some people have argued that it is inevitable. So inevitable it may be the only game in town for explaining action. This would make the “pictures in the head” version of the notion of belief essentially a non-negotiable part of our explanatory vocabulary. One of the main goals of our paper was to argue that there are other options even if they seem weird at first sight.

The alternative we championed was to think of belief as a species of habit. This requires both a revision of our implicit classification of mental concepts and a revision of what we mean by “belief.” In terms of the first aspect, the usual way to think of belief and habit is to see them as distinct categories in our mental vocabulary. A habit is a “thoughtless” activity, while an action driven by belief requires “thought” to be involved. So they are two sets of mental categories, but they are as a distinct as a frog is from a zebra (even if both are a species of animal). In our proposal, however, the overarching category in mental life (for both human and nonhuman animals) is habit, and belief is a subcategory of habit. This does violence to the standard classification so it may take time to get used to.

In this respect, note that the “picture” theory of belief seems to be important in how people differentiate belief from habit. Both can be involved in action, but when action is driven by belief, the picture inside the head is on the driver’s seat and is thus an important (but always presumed) component of the action. In fact, the picture is such an important component we attribute causal force to it. Sam got up from the couch and walked to the fridge because they thought there was a sandwich in there (and they were hungry).

One last observation about the pictures in the head account of action. When the observer imputes the belief “sandwich in the fridge” to Sam and selects this belief as the “cause” of the action, by what criteria is this selection made? I bring this up only to note that there are actually a bunch of other “beliefs” that the observer could have imputed to Sam, and which could be argued to be implicated in the action, but somehow didn’t. For instance, the observer could have said one of the beliefs accounting for Sam’s action is that “there was a fridge in the room.” Or that “the fridge was plugged in” or that “the floor could sustain their weight,” and so on.

This is not just a trivial “philosophical” issue. We could impute an infinity of little world pictures in the head to Sam. In fact, as many as there are “states of affairs” about the world that make it possible for Sam to get up and check the fridge, inclusive of purely hypothetical or even “negative” pictures (e.g. the belief that “there’s not a bomb in the fridge which will be triggered to detonate when the door is opened.”). Yet, we do not (sometimes this is referred to as the “frame problem” (Dennett, 2006) in artificial intelligence circles). This means that belief imputation practices following the picture version are necessarily selective, but the criteria for selection remain obscure. This kind of obscurity should be suspicious for those who want to recruit these types of explanations as scientific accounts of action.

But we are getting ahead of ourselves. The main point of this post is simply to warm you up to the intuition that maybe the pictures in the head version of belief is not as intuitive as you may have thought nor as unproblematic or non-negotiable as it is sometimes depicted. In a future post we I will introduce the alternative conception of belief as habit and see whether it is not subject to these issues.