Deconstructing an interview with Vikram Jaswal

In this week’s post I focus on an interview published two weeks ago by Shannon Rosa, of the Thinking Person’s Guide to Autism, with Vikram Jaswal, a University of Virginia psychology professor. In the world of facilitated communication (FC), Jaswal is best known for two things: (1) his collaboration with Elizabeth Vosseller, credited with inventing of a variant of FC called “Spelling to Communicate” (S2C, akin to the Rapid Prompting Method, aka RPM), and (2) his publication in Scientific Reports of a study that used eye-tracking equipment to detect evidence that purportedly showed that S2C is a legitimate means of communication. (A critique of that study is here).

In his interview with Rosa, Jaswal shares his thoughts on autism, FC, and FC skeptics. As an FC skeptic who has her own thoughts on autism and FC, I regret not having been there to interject a few thoughts. So I’ll do that here instead, inserting myself into the parts of the interview I found most provocative.

VJ: I was trained as a developmental psychologist, not an autism scientist.

KB: Thank you for acknowledging this.

VJ: …[A]bout seven or eight years ago… I had the chance to spend a week with a group of nonspeaking autistic young adults. One way that members of this group communicate is by pointing to letters on a letterboard held by a trained communication and regulation partner… I’d never spent time with a group of people who communicate in this way.

KB: Did you investigate the concerns raised by researchers who are familiar with this method (S2C, aka RPM) about the likelihood that the facilitators are the ones directing the messages? It sounds like you assumed from the get-go that what you observed was a legitimate form of assisted communication.

VJ: It’s difficult now to think about the ableist assumptions I went into that week with, but honestly, I was bowled over by the group’s curiosity, humor, and thoughtfulness. I was also moved by the friendships I saw and the care members of the group showed for one another. None of this fit the textbook narrative of autism that gets fed to undergraduates or the deficit-focused research that is so pervasive in autism science.

KB: Given (1) the concerns raised by researchers familiar with S2C/RPM, and (2) how S2C/RPM-generated messages contradict the established research about autism, why did you decide that it was only your prior assumptions, and not S2C and RPM, that were questionable? Also, it sounds like you’re assuming that deficit-focused research is necessarily “ableist.”

VJ: One of the most famous experimental studies in all of autism science (and arguably, one of the most damaging to autistic people) claimed that autistic children lack a theory of mind.

KB: What the research consistently shows is that autistic individuals have difficulty with three elements of so-called “theory of mind”: reading body language, especially facial expressions involving the region around the eyes; automatic, intuitive perspective taking; and understanding social emotions. These difficulties hold even if you control for language level, and first person accounts by individuals with high-functioning autism affirm this. Take a look at Temple Grandin’s Thinking in Pictures and John Elder Robison’s Look Me in the Eye.

VJ: I asked one nonspeaker to generate an idea for an experiment that could tap into whether autistic people understand that behaviors are motivated by intentions. He quickly spelled something like, “I’d make a movie of someone about to go into a bathroom. I’d stop the movie and ask the subject what the person was about to do.” Impressive! What a simple and elegant way to get at the understanding of intention, I thought.

KB: Again, it sounds like you assumed from the get-go that what you observed was a legitimate form of assisted communication. Do you not see a problem with treating as a reliable source of data about autism communications that are generated by a methodology that is as susceptible to facilitator influence as S2C is?

VJ: ..[A] small, loud, and strangely influential group of individuals has dismissed the possibility that people who communicate with assistance from someone else are conveying their own thoughts.

KB: No, we have questioned the certainty that people who communicate this way are conveying their own thoughts: a certainty that has led people like you to assume, a priori, that what you’re seeing is legitimate communication.

VJ: They insist that all users of these methods are actually being directed to point to particular letters by subtle cues that the assistant provides.

KB: No, we suspect that most (and perhaps all) users of these methods are actually being so directed. This suspicion could be set to rest by the kinds of authorship testing (blinding the facilitators) that you and other RPM/S2C supporters unanimously refuse to conduct.

VJ: This group of skeptics has made it impossible for nonspeaking autistic people who communicate with assistance to get the support they need in schools.

KB: Our goal is to ensure that autistic people who cannot communicate without assistance get evidence-based support in schools. We feel that taxpayer dollars should not support untested methodologies, particularly those methodologies that the research suggests are overriding and suppressing the communicative behaviors and educational needs of vulnerable individuals.

VJ: Some of these skeptics regularly ridicule and denigrate users of these methods and the people who support them.

KB: We critique arguments, not people. When we call out people, we call them out for refusing to subject FC and its variants to rigorous testing, for publishing FC propaganda (especially propaganda directed at desperate parents), and for attempting to fund FC through taxpayer dollars. As for the “users” of these methods, we do not denigrate them; we see them as victims, and we do not blame victims. Instead, we state our concerns about how their typing is being influenced, how their communication rights are being suppressed, and how their authentic bids at communication are being ignored.

VJ: They don’t seem to understand that the assistance is meant to help individuals gain the skills to be able to communicate independently.

KB: Skeptics have long understood that most facilitators have the best of intentions, assume that they’re helping their clients towards independent communication, and don’t realize that they’re the ones directing the messages.

VJ: …or that some individuals who used to require assistance now type independently.

KB: Cues evolve towards increasing subtlety—a subtlety that may eventually be undetectable to eye-witness observers. The occasional demonstration of someone appearing to type independently does not count as evidence of legitimacy—any more than a magic show demonstration counts as evidence of the ability of doves and rabbits to appear out of nowhere.

VJ: They don’t seem to understand the significant sensory-motor and regulatory challenges experienced by most nonspeaking autistic people which the trained assistant helps to mitigate.

KB: The sensory-motor and regulatory challenges of autism are well understood. These challenges do not explain why non-speaking autistic people need someone to hold up a letterboard to their index fingers in order for them to start typing out sophisticated sentences.

VJ: And they don’t seem to be aware of research showing that some communication interventions they believe are evidence-based (e.g., pointing to pictures, verbal training) are not effective for most nonspeaking people. 

KB: “Evidence-based” isn’t the same as “effective for everyone”. It has long been recognized that some established methods—including legitimate Augmentative and Alternative Communication (AAC)—have been only minimally helpful to some non-speakers. But many non-speakers are able to communicate basic wants and needs through AAC. For those who aren’t, untested methods that promise miraculous linguistic breakthroughs, while tempting, are not the answer.

At this point, the interview turns to Jaswal’s eye-tracking study.

VJ: If spellers were responding to cues from the assistant, we know from research in cognitive psychology that this would take time and be subject to error—imagine having to isolate one target letter out of 26 on the basis of some kind of subtle cue the assistant provided, and then having to do so for each letter in each word in a response that is several words long.

KB: None of the cues routinely seen in S2C—which include the facilitator (unwittingly) moving the targeted letter closer to the extended index finger, calling out the letter before is it selected, and using her free hand to indicate the target letter or to exert carefully timed tactile pressure on the typist—require a time-consuming, deliberative isolation of target letters. It is true, however, that S2C/RPM typing is several orders magnitude slower than ten-finger typing, with many milliseconds consumed by index fingers wandering around the letterboard rather than moving swiftly, deliberately, and directly to the next letter (particularly where less common letter sequences are concerned).

VJ: The participants in our study pointed to about one letter per second, rarely made spelling errors, and looked at the next letter in a response about half a second before pointing to it.

KB: You’re making it sound like your participants always looked ahead at the next letter. But, as you reported in your study, your participants looked at ahead at the correct letters an average of 71% of the time, ranging from 32% to 99%. In addition, they sometimes fixated on the wrong letter before fixating on the right one. And since your study had no control group, we have no idea how this behavior and level of accuracy compares to what would be found in participants whose intentional communication is not in question.

VJ: Additionally, they showed a couple of patterns in their response times that, when observed in non-autistic typists, have been attributed to cognitive processes underlying fluent spelling: They were faster to look at and point to letters within words than letters between words, and they were faster to point to the second letter in more common letter pairs than less common letter pairs (e.g., “e” following “h” than “e” following “j”). 

KB: There’s a huge difference between subconsciously habituating to common letter sequences and deliberately spelling out words as part of an intended message. Pauses between words may reflect facilitator cuing patterns rather than comprehension by typists of words and word boundaries.

VJ: I think the study is important because it suggests that the blanket dismissal of assisted autistic communication is unwarranted.

KB: No one is dismissing legitimate AAC. As for “blanket dismissal”, I’ll repeat what I said earlier: skeptics have questioned the blanket certainty that people who communicate this way are conveying their own thoughts.

VJ: Our study shows that some nonspeaking autistic people who communicate with assistance can convey their own thoughts.

KB: No, it shows that nine participants selected according to undisclosed criteria (your study leaves out your inclusion and exclusion criteria) fixated on the correct letters an average of 71% of the time, ranging from 32% to 99%; that they sometimes fixated on the wrong letter before fixating on the right one; and that they did so in a context that allowed multiple possibilities for facilitator cues (some of which were observable in the three videos included in the study’s supplementary materials—videos that, in turn, showcased a tiny fraction of the study’s spelling episodes and whose criteria for inclusion in the appendix, in contrast to the vast majority of video data that was not included in the appendix, were unspecified).

VJ: Of course, there is much more research to be done.

KB: I appreciate that you’ve acknowledged this. Are there any pre-registered studies that will test whether your results can be replicated?

VJ: But access to communication is a fundamental human right affecting everything from educational opportunities to the ability to participate in medical decision making. I believe we should be supporting folks to learn to communicate in as many ways as they can.

KB: Yes, indeed it is, indeed it does, and indeed we should.

VJ: I hope that our study helps open the door for more research on assisted autistic communication. I hope it also gives pause to those who seek to deny access to a method of communication that is clearly effective for some nonspeaking autistic people.

KB: See above.

The interview then turns to whether autism is a social disorder and what accounts for speech limitations.

VJ: [The] assumption that autistic children and adults lack social motivation [is] an assumption for which the evidence is problematic [he cites one of his papers] and which is contested by many autistic people…

KB: The claim you cite about a near-absence of social motivation has been made only about those individuals who are non-speaking/non-verbal. Your paper contests this claim in part by quoting high functioning individuals, who, as is well known, often have strong social motivation. You also quote FCed individuals, whose words, you assume, are their own words and not the words of their facilitators.

KB: In general, your paper amounts to a combination of skewed testimonials (testimonials largely generated through FC), uncontested statements, and speculation that autism is a motor disorder rather than a social disorder.

KB: You propose that the reason autistic individuals don’t look at other people’s eyes is that they find it aversive: no one disputes this. But if you don’t look at someone’s eyes long enough to follow their eye gaze over to what they’re looking at (Joint Attention), then you miss out on incidental language learning opportunities. The more diminished Joint Attention is, the more diminished language is.

KB: You speculate that the reason autistic individuals don’t point to things in order to share their interests with other people (another form of Joint Attention) as often as non-autistic children do is that they find it really difficult to point to things. You suggest, without evidence, that pointing taps into known autistic weaknesses in coordination and motor planning. You do not explain why, if pointing is so challenging for them, autistic individuals nonetheless manage to communicate lengthy, witty, eloquent messages via S2C/RPM.

KB: You then cite three first-hand testimonials about motor control. One is from Temple Grandin, who, whatever she might say about coordination and motor planning, has had no trouble pointing to things. And two testimonials are from facilitated individuals—Ido Kedar and Tito Mukhopadhyay.

KB: You also speculate that the reason autistic children don’t point to things as often as their non-autistic counterparts do is that the things they find interesting are different from the things that non-autistic people find interesting. But does this necessarily reduce someone’s motivation to point?

KB: Finally, you speculate that autistic individuals use modalities other than looking and pointing when they try to initiate Joint Attention with others—namely, proximity and touch. That, of course, is sheer speculation.

Back to Jaswal, who now turns to speech limitations.

VJ:  [W]e don’t yet know why about 30% of autistic people cannot communicate effectively using speech. But speech is the finest of fine motor skills, and one promising hypothesis is that the well-documented motor challenges many autistic people face play an important role.

KB: There are more plausible explanations for lack of speech in autism than motor-related speech challenges (aka apraxia of speech).

VJ: In a remarkable study published in 2008, Morton Gernsbacher and colleagues showed that “minimally fluent” autistic two- to eighteen-year-olds could be distinguished from “highly fluent” autistic children on the basis of their motor abilities as infants and toddlers. Gernsbacher and her colleagues also pointed out that other scientists have shown that motor skills can be better predictors of preschool speech fluency than the usual social suspects, like joint attention.

KB: There are a couple of confounding factors here. First, both poor motor abilities and poor language abilities are associated with a third factor: intellectual disability. Second, the only specific aspect of language that Gernsbacher et al looked at, beyond overall language level, was speech intelligibility. Speech intelligibility—i.e., pronunciation—is much more of a motor issue than, say, vocabulary. Multiple studies find strong causal links between Joint Attention and subsequent vocabulary. I’m aware of none that find causal links between motor skills and subsequent vocabulary.

VJ: I don’t work in this particular area, but I think the role of motor differences in autism is one of the most exciting and promising directions for autism science (for examples, see here and here).

KB: Your first reference is Elizabeth Torres’ interview at Thinking Person’s Guide to Autism. Here, Torres makes no statements about specific motor differences in autism. She merely speculates that, if there are motor problems, they would have downstream effects. Your second reference is equally speculative, proposing various ways in which motor challenges could potentially impede opportunities for language learning.

VJ: There are some fascinating recent efforts that may help to shed light on issues related to literacy and receptive language in nonspeaking people. Alexandra Woolgar, a neuroscientist at the University of Cambridge, is investigating whether nonspeaking autistic people show a signature brain wave pattern that is associated with language comprehension.

KB: Woolgar is routinely cited by FC-proponents, but her work is extremely preliminary and mostly merely promissory.

VJ: And Yoram Bonneh, a vision scientist at Bar-Ilan University, and his group have reported preliminary findings that some nonspeaking people show signs of literacy when they are tested using eyetracking methods, but not when they are required to point at pictures. 

KB: Bonneh’s work, not a published paper but a poster presentation at an autism conference (INSAR), found that 9 out of 12 minimally verbal participants showed evidence of looking, but not pointing, at pictures of basic concepts like bear and airplane when the corresponding words (“bear” and “airplane”) were presented in writing or speech.

The interview then turns to the question of facilitator influence.

VJ: Can this happen when someone is assisting a speller to communicate? Absolutely. Do I think it’s unique to this population? No, people routinely say things that they think others want to hear, or are pressured into making claims they would not make on their own.

KB: The kinds of facilitator influence that rigorous studies have detected (studies that blindfold the facilitator) show that the facilitator is directing the messages. Typing out someone else’s messages is qualitatively different from typing out or saying something you think others want to hear.  

VJ: Still, I think it’s important to consider the possibility that nonspeaking people are especially susceptible to influence... By virtue of being autistic and not having a language-based way to communicate, nonspeaking autistic people are not given nearly the level of autonomy (or opportunities to develop autonomy) that most non-autistic people are. They are told what to do, when to do it, and how to do it. They are praised for doing things people tell them to do and reprimanded for deviating. When they use the means they have available to object to something… their objections may be ignored (see here for a truly awful example).

KB: And see here for another.

VJ: It’s not difficult for me to see how this kind of experience, day-in and day-out, could lead someone to become compliant and highly attuned to what other people want them to do.

KB: Indeed. Diane Twachtman-Cullen writes about this in her up-close-and-personal investigation of FC in action. Here we see facilitated individuals becoming deeply distressed when they realize they’re not hitting the letters that their facilitators want them to.

VJ: In terms of ways to mitigate the risk of undue influence in assisted autistic communication, practitioners and spellers have shared several strategies.

KB: How do you know that the strategies shared by the spellers weren’t unduly influenced by the facilitators?

VJ: …[O]ne long-term strategy would be cultivating self-advocacy skills throughout development. This would entail, for example, respecting an individual’s protests where possible rather than insisting on compliance.

KB: Do the woman’s oral protests of “No more! No more!” here count, or should we believe the movie’s voiceover (“Please don’t assume that every word I speak is what I intend to say. Making words with your mouth isn’t the same as communication.”)?

In answer to a question along these lines from Shannon Rosa, Jaswal gives an example of an autistic individual who perseverated on the word “yes” even when he clearly didn’t mean it.

VJ:  So although one might have the intuition that speech is the most accurate indication of someone’s intent, clearly it is not. When there is a conflict between what someone says and what they write, I think asking for clarification, using context, history, and other information will be necessary. 

KB: Presumably the clarification is facilitated through typing, not through speech. So much for “No more! No more!”

The interview then turns to the possibility of false allegations of abuse via FC.

VJ: Disabled people are much more likely to suffer abuse than non-disabled people, and nonspeaking people are especially vulnerable to abuse because most of them have no way to communicate about it. Those who would deny people access to their most effective method of communication because of concerns about the potential for false accusations should, as Rua Williams recently wrote, “ask [themselves] why a false accusation is more harmful than the ability to accuse.”

KB: What Rua Williams actually said was “why a false accusation is more harmful than the inability to accuse.” But a more accurately phrased question would be “why a false accusation is more harmful than ensuring that facilitators are unable to attribute their own accusatory words to their clients.” The answer, of course, is obvious.

The interview then turns to the question of backwards pressure in (traditional, contact-based) facilitated communication.

VJ: I don’t know about experimental research on backward pressure and facilitated communication. I will say that nonspeaking (and speaking) autistic people have described difficulties feeling their bodies in space and controlling their movements. They have explained that pressure or resistance can help. We should take their testimony seriously.

KB: Your sources again amount mostly to testimonies elicited through FC. And you’ve just outed yourself as a likely supporter, not just of S2C/RPM, but also of the old school FC that, thoroughly debunked as it has been by empirical research, many S2C/RPM supporters prefer to distance themselves from.

Rosa then asks Jaswal how eye gaze is evidence of authorship rather than merely of attention.

VJ: …[V]isual attention toward a particular object can… serve as an indication of someone’s intention toward that object. For example, sighted people tend to look at things before pointing to them, picking them up, or interacting with them. The reason is because, as cognitive scientists have shown, visual information helps guide their movements toward them. 

KB: And if you’re being cued by someone else to point to a letter, you’ll also tend to look at it as you move your finger towards it—for the same reason. Eye gaze may be evidence of intentional pointing, but intentional pointing includes cases where someone deliberately points in response to someone else’s cues.

Neither eye gaze nor intentional pointing is evidence of authorship. Message authorship is more than intentionally pointing to one letter after another. Authoring a message entails intentionally spelling out meaningful words and intentionally sequencing these words into a message that one both comprehends and intends to communicate.