What is With the "No Theory of Mind" Thing?!

Intro and Background:
Amazingly, 100% of the ASD professionals that I have worked with over the course of the past 16 months have either based their entire bodies of work on autistic theory of mind (ToM)deficits or have used the term quite frequently to refer to things that have nothing to do with ToM. In fact, many times it has been used in reference to me as an autistic adult to explain why the ASD professional and I had a misunderstanding, without any further attempt to understand my perspective. The whole thing is simply “dismissed” (as am I and my views) with a flippant comment that I am lacking ToM. This boggles my mind and also concerns me, as it is so far from the truth.

Theory of mind is often discussed in autism circles in relation to an early work done by Simon Baron-Cohen and Uta Frith and later expanded upon by many other works. In this early pivatol 1985 work, researchers show that some children with autism have delaysin the development of ToM based upon their answers to a first-order false belief test. Since this first work (sprouting many subsequent works on the topic as applied to the autistic community), the ToM term has permeated most every single curriculum, paper, article and theory on autism. This makes me very worried, as it takes only a bit of reading into the more current literature and history of ToM to understand how the entire concept is being misused and overused today by the ASD professional community and caregivers as a whole. I can only conclude that all these professionals are too overwhelmed by their day jobs to stay properly informed or to discern fact versus stigma. Hopefully this article will help summarize some of this data for them.

Problem Statement/Myths:
One of the dogmas about ToM and autism is that all autistic people lack ToM. To make matters worse, ToM is associated with being human in many referenced works and also with empathy. The autistic community (and advocates even) tend to make these statements about autism and lack of ToM black/white (IF autistic THEN no ToM, empathy, human can exist). When I dig into the actual research, though, I find ToM and the statements/tests/research made about it are in various (even extreme) shades of gray. In fact, 20% of the autistic children tested in this most-referenced 1985 study passed the test and therefore had ToM! But that little detail somehow doesn't make the "headlines," and instead it is assumed that since I am autistic I must lack it entirely. Since the ToM deficit is most often applied to all autistics as a meme, and is also tied to empathy and being human, many advocates take grave exception to the stigmas associated with this perception. The result is a flood of articles and research/logic that pick apart these papers on the subject due to inaccurate science and also lack of respect for autistic people as a culture. These works are readily available to anyone who can search the internet.

Rachel Cohen-Rottenberg is perhaps the most interesting of these available resources. She dedicates her entire blog, Autism and Empathy, to dispelling the myths surrounding autism and ToM deficits. Every ASD professional should be required to spend a few hours on her page to really grasp how the words of the papers being written by NT professionals can actually hurt an entire people. Over the years, her work has matured quite nicely as she gets deeper and deeper into the real meat behind these works. She succinctly and sometimes humorously breaks down how generalizations and “NT is superior” statements stigmatize the autistic culture and have far-reaching negative unintended effects. She articulates quite well why it makes me cringe to hear ASD professionals throw this term around as if it is something that is defined and/or applicable to all autistic people for any situation that they do not otherwise understand.

Questioning the Dogma:
There are actually a number of works on ToM completely outside of the autism professional community, and I believe that one should cast an eye towards some of these to understand the foundation of this entire conversation. In philosophy, ToM is often called “folk psychology.” In this high-level article,“Folk Psychology as a Theory,” ToM is hashed out from multiple more objective perspectives. It cites the earliest works on the concept being completed in the 1950s, long before ASD professionals started to delve into this process of testing autistic children for first-order false belief failures. In reading this and other early papers, one can get a very good explanation of why this term is not ideal to adopt as a meme in the autism community for an entire group of people, as well as a glimpse into why first-order belief testing may not be all that it is cracked up to be.

The first thing that should stick out to the reader when digging into this ToM idea is that the term itself cannot really be defined. Through the decades, the definition itself is simply not solid or agreed upon. This causes all manners of confusion in the industry. In fact, the concept is so vague that two people (unless generalizing to a high level) speaking of ToM may be talking about entirely differing things. This sort of concept automatically creates the need to generalize versus be precise. But beyond that, and perhaps much more important, is the sheer volume of conflicting research, theories and ideas that surround the concept as a whole. For every key assertion made about ToM, there is at least one conflicting and equally valid one to be argued. After poring through a great deal of viewpoints and research on this, I have concluded that philosophy is a good place for a concept like ToM to reside.

False Belief as a Marker:
ASD professionals use a first-order false belief test on autistic children to determine that they have deficits in ToM compared to NT peers. The 1985 research done by Uta Frith and Simon Baron-Cohen seems very valid at first blush, but if you look beyond the SBC followers (Simon Baron-Cohen is "known" in the community as an authority) and into other related and non-related research on “false belief testing,” it soon becomes clear that not everyone in the field accepts that these tests reveal a conceptual deficit. Some argue that the tasks reveal performance deficit rather than ToM incompetence (Goldman 2006, section 4.3). More interestingly, though, is that recent research in this area shows that the ability to pass ToM tests is actually closely coupled with verbal IQ. Heck, even Simon Baron-Cohen suggests, in this work, that ToM develops quite regularly in children with Asperger's by the age of 9 (about 4.5 years later than NT kids), though he fails to make the verbal IQ leap/correlation at the time of this writing.

Besides the obvious conclusion that delayed language (verbal IQ) by definition of their diagnosis is a possible reason for autistic kids' test failures, one might also leap to the conclusion that most autistic adults probably can pass the first-order ToM tests. In other words, I know that you have thoughts, feelings, beliefs and experiences different than mine, so please stop using ToM as a reason why we do not understand one another.

Debunking the Sally Ann Test:
If one is to properly argue against the assertion that autistic people lack ToM, they must first look to the tests that keep saying otherwise in the field. The most common form of ToM test (for first-order testing) is called the Sally Ann Test.

In this video of the Sally Ann Test, you can see how the test is run and witness a young autistic boy failing the test.

If it is true that ToM exists in autistic people, then why do so many of them fail the Sally Ann Tests in the field? What could possible explanations be?

Reason #1. Sensory Processing Issues: One doesn’t have to search too hard to find blogs from autistic adults who theorize why this is happening. I may be biased, but if I wanted to understand the perspective of an autistic person, I would look to ask one. So that is where I started on my search for information to write this paper. The number one postulation put forth by autistic adults seems to be in the way an autistic person processes sensory input. In all of these blogs, one of the primary reasons listed as reasons why an autistic person may fail Sally Ann is sensory issues or a difference in the way we process sensory input.

Critique of the ToM Test (Rachel Cohen-Rottenberg)
Deconstructing the Sally Ann (unknown author)
ToM an Autistic Perspective (Bob Castleman)

I give much credence to this possible explanation but feel it is perhaps too simplified in light of the extensive data on this topic. When looking at the various ways this test is run and the subsequent results, the facts are that over and over again, autistic children do more poorly on this test than NT children even when matched at about the same verbal age.

This leads me to keep digging and to put forth other possible reasons for this disconnect.

Reason #2. Language Processing: The most glaringly obvious thing that jumps out to me when looking at the critiques of Theory of Mind testing is that this testing is less about understanding that others have feelings and much more about processing language and context.

In this this paper (Does the Autistic Brain Lack Core Modules? Gernsbacer and Frymiare; 2005) the authors note that all children with language difficulties (not just autistic children) tend to fail the Sally Ann test. The authors then, very astutely, point to the (now obvious) fact that the key question asked during this test is one of the most complex constructions of language. The question, “Where will Sally look for her Marble?” consists of a complement clause embedded in the matrix clause. But even more fascinating they unearth research that proves emphatically that there is more than one way to “skin the cat” (or pass the Sally Ann test)…

Although training in mind reading has become a popular intervention for autism
(Howlin, Baron-Cohen, & Hadwin, 1998), Tager-Flusberg and her colleagues (Hale
& Tager-Flusberg, 2003) have demonstrated that grammatical training on sentential
complement structures—low-frequency grammatical constructions such as “what will
Jesse think is inside the box before I open it”—improves performance of false belief
tasks as successfully as training on only false belief tasks. And training on sentential
complement constructions has an added benefit: Improving the understanding of
other sentential complement sentences.


Another read that stands out as noteworthy to me is Tager-Fusberg’s longitunal study (Steele, Joseph and Tager-Flusberg). In this one, the researchers looked at 57 autistic children between the age of 4 and 14 years over a long period of time. They concluded that vast improvements of theory of mind correlated to language acquisition.

Since a primary diagnostic criterion for autism is “a qualitative impairment in communication manifested by a delay in language” it is sort of a no-brainer that autistic children will fail a test that relies heavily on Language processing to pass until they gain the appropriate language skills to pass the test. It is actually quite possible that the only thing that Uta Frith and Simon Baron-Cohen demonstrated in that infamous 1985 research work is that autistic children have language delays.

I have it on my list of things to do (or possible PhD thesis for someone) to reconstruct the Sally Ann test in a way that allows for people without language and EF issues to also pass but still provide evidence that they can process other people’s thoughts. One of the ideas is to have Sally come in and look in the basket for her marble and not find it there. Then ask the child why she looked in the basket instead of the box. If the child can articulate that she looked there because that is where she put it before she left, that may be enough. (more on this as time allows...)

Of course, in order for this to work, there will have to be accommodations for lack of EF skills and correct context too (see reason other reasons).

Reason #3 Executive Functioning Issues:It is a well accepted paradigm that with any neurological difference comes also a vast array of Executive Functioning (EF) issues. Autistic people typically experience the issue as secondary to autism and the deficit can be mild or extreme but most exhibit this deficit in some form. Executive functioning (very simply put) is the ability to plan, solve problems, stop impulses, use foresight, hindsight and a sense of time, etc. Just by reading this list, it should be starting some thought process about how this might play in the Sally Ann test results.

One striking statement made by Temple Grandin in a very recent article is that she cannot remember more than 3 steps which made her have issues with word problems in Math. I know in my own case that my attention also dissipates long before the “key” question is asked when given this test. In fact if you begin to break it down in any sort of detail, you could pretty reasonably hit 15 or more key points to consume before the “key” question is asked…

1. This is Sally
2. This is Ann
3. Sally has a basket
4. Ann has a box
5. Sally has a marble
6. etc...

Recently, I took this test out to the autistic adult community and administered it to adults that I know have some ToM skills. Not surprisingly several of them were unable to answer the key question. These same people were confused as to what happened during the test or confused about the context of the question.

Reason #4 Context differences: The more I interface with the NT world as an Autistic person, the more I understand how we simply do not speak the same language from the same context or starting point. My words are very, very specific (usually to one meaning) and NT words can be abstracted, inferred and generalized in ways that completely confuse me. Additionally, I will focus not on people or on social stories but on objects, patterns and Math in scenes, movies, books, etc. The end result is that we do not see the same picture at the end of the same story most of the time.

In my personal case, with respect to this test I care about neither Ann nor Sally when I am being told this story. I instead focus about the marble and what happens to it. I am hyper-focused on the ball, its texture, colour, size, etc.. and disregarding the people. This does not equate to a lack of ability on my part to understand that people have different thoughts, feelings and beliefs than I. Nor is it the same as lacking empathy. It means I have a different set of data with which to take the test.

This matches exactly with how I tend to watch movies and/or TV as well. I rarely know anything at all about the people and often even the storyline fades away for me. I can tell you in much detail about the settings or the objects in the scenes and sometimes I even see a unique storyline to what other see. I realize some autistic people care for and seek out people more than objects, but I am an object person first. I will typically be more interested in the things in the room at a party than who is in the room at a party. Why would that not also hold true for this test? You may be telling this story about Sally and Ann, and even believe it is a story about these people, but what I may hear is,

  1. Marble (note all the details about it now)
  2. Marble in basket (note all the details about basket)
  3. Marble in box (note all the details of box)
  4. Where will xx look for marble?

At this point the logical answer is for her to look for it in the box where it is, not where she left it before she left the room because her leaving the room is not part of the storyline from my perspective. Other autistic people I spoke with cite similar context differences including not fully understanding social constructs of the story and feeling the need to show Sally where Ann put her marble instead of answer where Sally will look for the marble.

Conclusion:All this “debunking” aside, I strongly believe that there is “something” to be said for the work done thus far on ToM and autistic people. There really are core differences in how I think of others versus how NTs tell me they think of others. When an NT enters a conversation with other human beings, they have at the front of their minds a sort of paranoia about what the other person is thinking of them. They seem to care about this sometimes to the exclusion of the reason they are speaking to that person in the first place. I can honestly say that I have never thought about what another person thinks of me when I enter a conversation. This is not because I do not know that they have thoughts independent of me, but rather because I do not care what they think about me. I do not enter a conversation to impress upon anyone that I am clever, or kind or any other thing I may think about me. I enter a conversation in order to learn and discuss matters of importance. What a person thinks about me is not important so long as I am not hurting or impeding them in any way. What they think of my work, however, is very important. I want my work to be valued so that is where my focus is in conversation. I could care less if they value me. My work holds me up in this world, not my social skills. So there is most certainly a difference in how NTs see the world from a “thinking about you thinking about me” point of view. I do not do this thing and find it rather pointless. Small talk is not in my abilities or repertoire. Some autistic people actually do this “worry about how people perceive me” thing, and even learn to script small talk. They typcially develop social anxiety over the whole thing even. But this is not how I work at all. It just is not there… So in this way I have less concern about what you are thinking about me which may be perceived as a ToM difference.

Also in order to have fully developed ToM skills, one must first have developed “self” awareness skills. Here again, I am significantly different than my NT peers. I have a far simpler understanding of emotions and also awareness of them. I tend to run exceedingly flat emotionally compared to an NT. Again there are many autistic people who tend to run exceedingly “hot” emotionally and have extreme emotional experiences. I marvel when I read their blogs at how rich their experiences are emotionally, but I am not like that even a tiny bit. So here again a ToM difference may be cited as there is an inability to fully understand/appreciate the emotional experiences of people in general for someone like me. Much of this, I will never really get on anything other than some cognitive level due to my impaired abstraction abilities but that does not mean I have no empathy, or emotion. It does not mean that I do not appreciate that you have these feelings more deeply than I. It only means that emotions do not dictate my life. Logic tends to prevail in all my major (and minor) life decisions and emotions are a secondary consideration.

But do these differences really equate to this general claim of lacking Theory of Mind? I believe it clearly is not the case and I hope this document gives enough evidence and pointers to agree. If not, perhaps when the APA has settled the current big DSM V debate, they can all agree on what Theory of Mind really is so that we can answer the question about my differences definitively…

....Perhaps my neighbor’s pigs will fly.

Relevant links:
Not Mind Blind but Context (Culture) Blind
Deconstructing The Sally Ann Test
Why Sally Ann Tests fail autistic people
Critical reexamination of learning models in autistic children
Sally Ann Tests failure due to sensory?(Rachel Cohen-r blog)
http://www.huffingtonpost.com/ariane-zurcher/autism-theory_b_1594706.html (Huff post article debunking theory of mind)