Ian Hacking’s critique of the Theory-of-Mind-deficit theory of autism

[This post is the eleventh in our series of Thinking in Action posts, the series being devoted initially at least to discussion of talks at the Cognitive Disability conference in NYC in September. You can go to the Thinking in Action 10 pack, which links to the first 10 posts in the series; and the posts run Tuesdays and Fridays, for the most part. The post below concerns talks by Ian Hacking and Victoria McGeer on theory of mind and autism at the conference.]

Background:

Theory of Mind and its deficit:

“Theory of Mind” (ToM) is a philosophical interpretation of a certain kind of cognitive psychology. The idea is based on what has been called folk psychology. This describes our ordinary understanding of each others’ behavior as analogous to a scientific psychological theory. Each individual’s own folk ToM hypothesizes that other people have unobservable (to the observer) intentions, beliefs, and desires. These hypothesized mental states are seen as analogous to theoretical conjectures in science. On this notion, we begin in childhood to construct a theory of mind about other people, and we elaborate that theory as we develop and mature. An underlying assumption is the double-edged notion that A) human behavior is based on (perhaps caused by) internal, language-like inferential structures in the brain (e.g. beliefs and desires), and B) we hypothesize (in our ToM) that other humans have the same kind of language-like structures that we ourselves use in reasoning about the world.

Beginning of a cartoon illustration of Theory of Mind; you hypothesize the cartoon character's innards

[To observe ToM for the above cartoon click here. Watch right away — on some browsers it only runs once. Your “theory” is about why the cartoon character is going downtown.]

ToM-deficit as a theory of autism:

Psychologists Simon Baron-Cohen, Uta Frith and others built a theory of autism based on the ToM notion of cognitive psychology. People on the autistic spectrum often have difficulties interpreting the behavior of neurotypicals. For example, autistic children fail at certain “pretend” tasks at a later age than neurotypical children (especially tasks involving deception). These difficulties are said to be caused by a failure in the autistic children’s ToM process, which autistic children learn at a later stage than neurotypical people, and possibly never learn at all.

Hacking’s alternative:

Hacking rejects ToM in general, not only in the ToM-deficit theory of autism. He replaces it with a Wittgensteinian Form-of-Life (FoL) theory of language and social knowledge. On this view, language and social interaction is a norm-based practice, and such practices cannot be analyzed in terms of internal, language-like “theories” about the domain governed by the norms. Practices cannot be reduced to theories; you cannot learn to rollerskate by reading a book. The ToM notion that we infer people’s intentions based their behavior is a mistake (says Hacking); we intuitively and directly see people’s intentions. He callse these intuitive “seeings” of mentality are “Köhler phenomena” (after the Gestalt psychologist who, Hacking says, inspired Wittgenstein). The intuitive skills of neurotypicals are falsely described by ToM, and so autistics are falsely described as having a deficit of ToM.

Autistic Narratives:

Hacking proposes that the autistic narratives may actually contribute be constituting (rather than merely describing) the nature of autistic experience. This is especially true of reports of pre-linguistic experience that many autism narratives report — experience that which occurred before the autistic individuals (who wrote the narratives) had achieved linguistic communication. This final claim relates to Hacking’s earlier studies of fugue states and multiple personality conditions. These psychological conditions were, in part, constituted by the ways in which people decided to describe them. This is Hacking’s version of social constructionism, which avoids some of the epistemological relativism that accompanies other versions of constructionism.

McGeer’s support of Hacking:

McGeer argues that Hacking’s FoL theory is superior to the ToM-deficit theory in its treatment of autism on two grounds. One is the fact that the ToM-deficit theory cannot account for the symmetry of the two groups’ of people failing to understand each other — ToM-deficit would presumably predict that neurotypicals would understand autistics, but not vice versa, whereas vice versa appears to be the fact. I will not comment on this defense. The second defense is that ToM-deficit theory offers no grounds to take autistic narratives seriously. It seems to imply that autistic self-narrative (which often describes the subjective experiences of the author) is unreliable because the authors, lacking a ToM, are assumed not to be able to understand what subjectivity even was. In contrast, the FoL view would encourage taking autistic narrative seriously, as potentially constituting an alternative “Form of Life” distinct from that of neurotypicals. McGeer points out a puzzle in this extension of FoL, however. FoL are necessarily social and “co-minded,” having been constructed by a linguistic community. Individual autists (or individual neurotypicals for that matter) cannot be said to be living their own Form of Life because FoLs are necessarily not individualistic. So perhaps a truly autistic FoL is only recently in the process of emerging, of being constructed, as autists are beginning to form social relationships with other autists.

Commentary:

I find Hacking’s approach to social construction very attractive. He avoids the aspects of social constructionism that make it most vulnerable to ridicule. However, I was surprised to recognize the degree to which Hacking and McGeer’s presentations challenged (on Wittgensteinian grounds) any version of cognitive psychology, not merely a ToM version. McGeer mistakenly defines cognitive psychology as co-extensive with ToM psychology. In fact cognitive psychology need not be committed to a ToM version of language or social understanding. Cognitivism can be defined as a form of psychology that depicts psychological abilities as quasi-computational abilities that can be described in terms of the internal manipulation of representations of the objective world (including the psychological subjects themselves). Such a theory might or might not depict these internal representations as similar to a “theory” as theories exist in science. Some cognitive psychological theories will support a ToM approach to human social interaction, but others will not. Hacking’s FoL approach to mentality is inconsistent with all cognitive theories, not just with ToM theories.

Hacking’s Wittgensteinian Form of Life theory differs from cognitivism (and therefore from ToM) in the nature of what is acquired. Cognitivism says that it’s a representation of the world. Wittgenstein (and Ian and Tori) says that it is a practice, and that practices are irreducible to representations. All three of these philosophers — Ian, Victoria, and Ludwig — treat the achievement of linguistic/cultural competence as ineffable from the standpoint of the theoretical resources of cognitive psychology — it cannot be theorized and must simply be practiced. I will set this topic aside for another day (but with one last swipe — it seems to me that Wittgenstein’s own opponents were not cognitivists, but introspective structuralists, a species entirely extinct today).

It is possible to strongly critique the ToM-deficit theory of autism without delving into the fundamental roots of philosophy of psychology; we need not decide whether cognitivism or Wittgenstein were correct (if either). The alternative I will offer is consistent with either theory, and many others. It has the additional virtue that it is endemic to the autism spectrum. My awareness of this alternative came from participation in a majority-autistic internet list, and especially the posts of Jim Sinclair and Cal Montgomery, two philosophically sophisticated autistic authors. I would never have come up with it myself. I will term it the Recognition of Difference (RoD) Theory: Autistic children learn from an early age that their experiences of the world are very different from those of other people (children or adults) and that adults have vastly greater sources of knowledge than they do. For this reason, they do not assume that other children will have the same experiences as they do when placed in the same circumstances.

It is impressive how many of the results of the experiments that supposedly support the ToM-deficit theory are adequately explained by the RoD theory, without dredging down to the philosophical basis of cognitive psychology. All that needs to be done is consider how different are the subjective worlds of people on various positions of the spectrum of neurodiversity. This is a very hard thing to do of course. But not impossible. I will attempt in a follow-up blog post.