When nothing else works: Blaming the scientific method instead of the pseudoscience
The world of facilitated communication and variants (RPM, S2C, Spellers Method) is heavily fortified against researchers and skeptics. Those fortifications are one reason why, despite all the evidence against it, and besides all the support it provides to desperate parents in impossible situations, FC/RPM/S2C has persisted for so many decades.
Some fortifications derive from social pressure. Using FC typically involves joining a tight-knit community of other FC users that becomes one of your main support networks. Should you ever grow disillusioned with FC, you lose all those friends, if you agree to undergo authorship testing, you incur their hostility; if you speak out publicly, you become their enemy. Most defectors keep quiet (our Janyce is an extraordinarily rare exception).
Other fortifications derive from professional pressure. If you’re a “certified” S2C practitioner and agree to authorship testing, you risk losing your “certification.”
Other fortifications derive from legal pressure. If you live or work in proximity to a facilitator and their clients—whether you are the spouse of a person who facilitates your child, or a teacher/clinician who works with a child who is facilitated by others—and if you express skepticism about, or otherwise try to resist, the use of FC, you risk serious harm to your reputation, your livelihood, and even your basic rights. To be more specific, you risk being accused, via FCed messages controlled (however unwittingly) by those who now see you as their adversary, of abusing the FCed individual. You may lose your job or custody of your child; you may get locked up for months in solitary confinement. (See Stuart Vyse’s highly disturbing piece on one such account).
Fortress of the Îlette of Kermovan, in municiplaity of Le Conquet (Finistère, Britanny, west of France). Wikimedia Commons.
But FC/RPM/S2C has also fortified itself in more theoretical ways. That is, the theories that proponents have advanced in support of FC include claims that attempt to make FC impossible to explore scientifically. One is a claim about minimally and non-speaking autism that, if true, would invalidate any observations or tests of the communicative and cognitive capabilities of minimal and non-speaking autistics—except for those observations and tests based solely on FCed output. Other claims seek to invalidate tests that are based on FCed output—specifically, message-passing tests, or tests that blind the facilitator in order to determine who is authoring the messages. Collectively, these claims, if true, would mean it’s impossible to assess the validity of FC, whether by testing via a standard language assessment if the facilitated person understands the words they spell, or by testing via a facilitator-blinded condition if the facilitated person can spell correct answers that their facilitator doesn’t know and therefore can’t cue.
In the rest of this post, I’ll take a closer look at these claims.
The claim about minimally and non-speaking autism is that the underlying disability isn’t the socio-cognitive challenges that eight decades of clinical observation, diagnostic screening tools, and empirical research have shown—and continue to show—autism to be. Rather, the underlying disability, according to FC proponents, is a mind-body disconnect. Individuals with minimal and non-speaking autism are, allegedly, unable to control their bodies, especially their fine motor movements. This effectively invalidates any attempt to assess the cognitive or linguistic skills of minimal and non-speakers with autism. That’s because all responses to test questions involve bodily responses, particularly fine motor responses: speaking, pointing to pictures, assembling shapes, and using a pencil. Indeed, all inferences based on any behaviors are purportedly invalid. Included among the invalidated inferences are these:
That if someone is always looking away from the people speaking in their presence, they’re probably not paying attention to what’s being said;
That if they don’t respond appropriately to requests, directions, or other communications, it’s probably because they don’t understand you;
That if they don’t speak or type independently, it’s probably because they don’t have expressive language skills;
That if they walk away or try to escape while being facilitated, or say “No more! No more!”, they don’t want to be facilitated.
As Elizabeth Vosseller puts it, “Insides don’t match outsides.” But where does that leave us?
Even eye movements, powerful indicators of what someone is attending to or thinking about, purportedly reveal nothing in minimally and non-speaking autism. Some FCed individuals are said to have “ocular apraxia,” such that they can’t control their eye movements. Some are said to use peripheral vision instead of direct gaze, such that where they’re looking isn’t necessarily what they’re looking at, and such that they can supposedly see the letters they’re pointing to during FC sessions even when they appear to be looking away. Some routinely wear sunglasses, allegedly because of visual sensitivities, making it hard to see what they’re looking at.
But, in fact, the only basis we have for making any judgments about anyone (unless, of course, we’re telepathic and can read their minds directly) are their body movements: their speech, gestures, facial expressions, and actions. The mind-body disconnect theory of autism, therefore, makes autistic individuals essentially impossible to evaluate—and makes all claims about autistic individuals both unverifiable and unfalsifiable.
This, in turn, makes the mind-body disconnect theory not only profoundly wrong (i.e., in conflict with eight decades of clinical observation, diagnostic screening tools, and empirical research, and without any empirical support of its own), but profoundly unscientific. That’s because falsifiability—being susceptible to being disproven—isn’t just the foundation of science, but the largely agreed-upon demarcation between science and pseudoscience that dates back to philosopher of science Karl Popper.
Which brings us to a quick detour into Karl Popper’s insights.
Popper, recognizing that some claims are hard to prove, also recognized that some such claims are nonetheless more scientific than others. In particular, he recognized a distinction between claims like All swans are white and claims like We’re living in a computer simulation (my example; not his). While it’s impossible to prove that all swans are white (because that would involve somehow inspecting every single swan, past, present, and future), All swans are white, at least, has the virtue of being falsifiable. That is, it can be falsified if a single non-white swan is found. And thus, All swans are white counts as a scientific claim (one that may turn out to be false).
We’re living in a computer simulation, however, cannot be falsified. That’s because anything that looks like evidence that we’re not living in a computer simulation could be part of the simulation. Any experiment we try to do to test whether we’re in a computer simulation, including the results of that experiment, could be part of the simulation. Thus, this claim is not a scientific one—it will never turn out to be false, even if it is false.
Curiously, there’s at least one outlier in the philosophy universe who’s proposed a totally different way to demarcate science from pseudoscience. This wasn’t someone I’d ever heard of (he appears to be unaffiliated), but he was recently cited on X, seemingly in support of the paranormal claims made on the Telepathy Tapes podcast about non-speaking individuals with autism who are subjected to S2C.
This self-styled philosopher—I’m not sure what else to call him—states that it isn’t the various empirical claims out there that we should be judging as pseudoscientific, but, rather, the methods we use to judge those claims (Westcombe, 2019). For him, a method is scientific “if it is well suited to establishing the truth or falsehood of a particular empirical claim” and pseudoscientific “if it is not well suited to establishing the truth or falsehood of a particular empirical claim.” Since no method is well suited to establishing the truth or falsehood of an empirical claim that turns out to be unfalsifiable, any method that takes on such a claim is, according to Westcombe’s approach, a pseudoscientific method.
By shifting the stigma of “pseudoscience” from claim to method, this approach effectively shields all claims from all charges of being pseudoscientific. Instead, the only entities that are potentially pseudoscientific are the methods that investigate these claims. Any method that investigates a claim that turns out to be unfalsifiable takes the blame for that unfalsifiability: the method is now pseudoscientific, while the claim itself stays above reproach. Since Westcombe fails to show why this seemingly upside-down philosophy is superior to Popper’s—he doesn’t even mention Popper—it isn’t worth spending more time on it. Except that, as we’ll see below, Westcombe isn’t the only one to suggest that methods, rather than claims, are the ones at fault when methods investigate problematic claims.
Which takes us to the other set of claims that attempt to fortify FC against scientific exploration. These are claims that attempt to discredit message-passing tests—tests that rigorously assess who is authoring the facilitated messages by blinding the facilitator to the answers to questions directed at the person they’re facilitating. Such tests, mostly dating to the 1990s, have consistently shown that it is the facilitator, not the facilitated person, who controls the typing.
Some arguments against authorship tests invoke the Observer Effect: the disturbance of what’s being observed by the act of observation. For example, according to Sheehan and Matuozzi’s pro-FC paper (Sheehan & Matuozzi, 1992), Nobel Prize winning physicist Arthur Schawlow, himself the father of an FC user,
described certain experimental efforts to investigate facilitated communication validity[i.e., message-passing tests] as analogous to looking for a ping pong ball on the floor of a dark room by shuffling your feet around. If you touch it even slightly it is not there anymore.
It’s not clear how message-passing tests could cause the equivalent of a dislocation of a ping pong ball, but this objection would apparently rule out such tests as hopelessly unreliable.
Some arguments against authorship tests focus on psychological effects that purportedly invalidate their results. Rigorous testing, allegedly, is inherently hostile; facilitated individuals, allegedly, sense the researchers’ skepticism about their linguistic capabilities. Their performance is further undermined, allegedly, by “stereotype threat”: negative stereotypes that can undermine test performance in marginalized groups, in this case negative stereotypes about the capabilities of minimal and non-speakers with autism. All this conspires, allegedly, to create memory retrieval difficulties so prohibitive that the facilitated person is unable to come up with the words for the pictures they’re shown during testing—even words as simple as “hat” and “flower.”
One problem with these arguments—besides the lack of evidence for them—is that they don’t explain how it is that, as in some of the 1990s message-passing tests, the facilitated person is able to type the words that label what their facilitator saw. Why would a facilitated person, allegedly underperforming due to a hostile environment, be able label these words, but not the words that label what they saw and their facilitator didn’t see?
I’m aware of only two proponents of FC/RPM/S2C who attempt to address this question: Rosemary Crossley, credited with introducing FC to the English-speaking world in the late 1970s, and Cathie Davies, a once-frequent commenter on this blog whose comments here ceased three years ago, right around the time that Crossley passed away. In those comments (here, here, and here), Davies acknowledges that in the rigorous message-passing tests of previous decades, facilitated individuals often typed what the facilitator saw, not what they saw and their facilitator didn’t.
The results of message passing studies should, by now, be very predictable. There is little point in replicating such studies, as I am not aware that the results are widely contested.
But for her, the question is “how are those results to be interpreted?” Critics, she claims, have produced only one alternative interpretation to facilitator control: “the straw man – ESP.” Davies, if she’s still around, is apparently not an enthusiast of the Telepathy Tapes.
For Davies, neither facilitator control nor telepathy explains why the facilitated person types “flower” when only the facilitator saw the picture of the flower. Furthermore, like Westcombe, she faults the method (asking the autistic person to label an external stimulus like a picture) rather than the claim it’s investigating (the validity of FC). For Davies, the method is at fault for being unreliable, and the researchers are at fault for drawing faulty conclusions.
Your preferred experimental design may be adequate to demonstrate that many FC users do not pass messages under controlled conditions. However, such studies have no power to explain why this may be so. We do not currently know enough about the relevant population to design experiments able to distinguish between competing explanatory hypotheses.
Thus, she concludes, all conclusions based on this method (prompting the autistic person to label a picture that the facilitator didn’t see) are “speculative.”
One reason they’re speculative, Davies says, is:
[T]he principle of the underdetermination of theory by evidence (Quine, 1951). Experimental data is typically consistent with a broad array of competing theoretical explanations.
But how is data generated by tests that prompt the autistic person to label a picture that the facilitator didn’t see ambiguous in any significant way? The ambiguity, Davies claims, comes from the fact that the tests are “closed-item” tests—that is, they solicit a particular word or phrase rather than an open-ended response. This, she claims, is not representative of the type of communication that FC is all about—namely, open-ended communication:
[S]ubjects’ performance in message passing or responding to closed questions says little about their capacity for self-motivated communication on topics important to themselves: the type of communication most commonly reported in non-experimental studies and, arguably, the type of communication most valuable to the communicators.
She adds:
[T]he task demands for self-motivated communication are different from those for message passing under controlled conditions, particularly in relation to processing exteroceptive (“outside-in”) sensory information [presumably by this Davies means tasks like labeling a picture].
For this reason, Davies claims, closed-item tests are actually less rigorous than tests that elicit self-motivated communication.
She adds that closed-item testing is more susceptible to facilitator influence than open-ended communication:
Research has demonstrated that facilitator influence is more likely in the context of closed questions with simple answers (Wegner, Fuller, & Sparrow, 2003).
The facilitated individuals in Wegner et al.’s studies were verbally fluent adults (at least inasmuch as they were Harvard undergraduates); this limits the applicability of Wegner et al.’s findings to minimally and non-speaking FCed autistics. Nevertheless, Davies claims that these findings further undermine the assumption “that tests involving closed questions with simple answers are the most rigorous tests.”
But the more open-ended the question, the harder it is to blind the facilitator to its answer, and the harder it is to verify that the answer produced by the facilitated person is the correct answer. Thus, closed questions are necessary conditions for rigorous tests.
As for Davies’ claim that “responding to closed questions says little about [the] capacity for self-motivated communication,” the linguistic skills involved in picture-labeling (basic expressive vocabulary skills) are a prerequisite for open-ended communication. If a person isn’t able to produce the word “flower” when asked to label a picture of a flower, how are they able to produce more sophisticated words and string them together into the sentences that regularly appear as FCed output? And how is the cognitive process of producing the word “flower” in response to a question about a picture more challenging than producing the word “flower” in an open-ended, “self-motivated” description of a walk you took through the park? Davies might claim that flowers and walks in the park might not qualify as “topics important to themselves.” Message-passing tests, however, can be, and have been, adjusted to include objects related to such topics, and the results remain the same.
But Davies also claims that in “closed-item testing” the facilitated person may be less sure of him or herself and so may seek cues from the facilitator:
[W]hen unsure what is required of them or anxious about presenting a wrong answer - FC users may actively and intentionally seek cues from their facilitators.
She adds:
It would, however, be difficult to characterise this behaviour as “control” by facilitators!
For Davies, in other words, if a facilitated person types “flower” when only the facilitator saw a picture of a flower, this may be evidence, not of control by the facilitator, but of “cue seeking” by the facilitated person.
As an example of cue seeking, Davies cites a subject with “complex communication needs” in a study by Kezuka (1997):
It appeared that J had learned to scan the keys with her pointing finger until slight changes in force from the assistant signalled her to drop her finger to the letter below” (p.576).
What Kezuka means by “learned” here is unclear: some learning is non-conscious and includes conditioned responses to facilitator cues which, in turn, is the basis for facilitator control. But arguably, even conscious learning about facilitator cues, and adherence to those cues, count as facilitator control. If a teacher shushes a student, and the student, as a conscious response to being shushed, stops talking, the teacher is arguably the one in control. There is, in other words, little reason to believe that the scenario presented by Kezuka doesn’t involve facilitator control.
Besides claiming that they “cannot distinguish between ‘control’ by facilitators or ‘cue seeking’ by communicators” and that they do not measure the “self-motivated communication skills” that are “representative of FC practice,” Davies cites one more problem with message-passing tests: their consistently negative results:
Regardless of the reason, testing FC users’ ability to communicate using a task that is so clearly problematic for many must be questioned.
In other words, what most people would consider to be evidence of facilitator control—the consistently negative results of message-passing tests—Davies considers, instead, to be evidence against the validity of the tests.
This line of reasoning, of course, would call into question the validity of any method that consistently returns negative results—for example, a method that consistently returns negative results for claims that the earth is flat.
Davies cites the following from Shadish et al (2002):
[I]f a study produces negative results, it is often the case that program developers and other advocates then bring up methodological and substantive contingencies that might have changed the result. For instance, they might contend that a different outcome measure or population would have led to a different conclusion. Subsequent studies then probe these alternatives and, if they again prove negative, lead to yet another round of probes of whatever new explanatory possibilities have emerged.
Shasish, as quoted by Davies, eventually goes on to say that there may be a consensus that the negative results are real, but “that this process is as much or more social than logical.” In other words, non-objective and non-rigorous.
Returning to the question of authorship, Davies states: “I would like to say ‘get a new test.’” And what would be a better test than asking the facilitated person to label a picture that the facilitator didn’t see?
Instead of concentrating on the comparatively unremarkable cue-seeking behaviour, surely it would be better to engage with the population in coproduced, participatory research to explore what else may be going on?
Elaborating, Davies claims that:
[M]odels of evidence-based practice...demand consideration, not only of academic research, but also of stakeholder perspectives and clinical experience. The weighting given to evidence from these sources is decided through open and transparent deliberation. No one of these three types of evidence is automatically given precedence over the others.
In other words, non-objective and non-rigorous. Any research on FC that includes “evidence” from stakeholder perspectives and clinical experience will naturally include facilitators (whose vested interests and resultant biases are arguably stronger than anyone else’s) and FCed messages attributed FCed individuals (where any a priori assumptions about authenticity will lead to circular reasoning and warped outcomes).
In the end, though, Davies seems to rule out any kind of authorship testing—no matter how rigorous or “rigorous.” She claims not only that “no valid outcome measure [i.e., authorship test] is currently available,” but that, even if there were a “valid” authorship test, deploying it may no longer be feasible. That’s because:
[A] group of individuals as hostile and entrenched in their dogma as critics of FC have proven to be... people such as yourself and your colleagues... may have effectively poisoned the waters so that no such research can be conducted.
In short, Davies tries to argue
that the most rigorous tests are the least rigorous
that the consistently negative results they’ve produced show that there must be something wrong with the tests, not what they’re testing, and
that the skepticism that people such as myself and my colleagues have acquired as a result of those consistently negative results has made any future authorship testing of any sort impossible.
In the end, this shifting of blame from the problematic claims that experimental methods have consistently invalidated over to the experimental methods that have invalidated the problematic claims recalls Westcombe’s upside-down reasoning about pseudoscience.
Cathie Davies, assuming she’s still alive and that this is her actual name, appears, like Westcombe, to be unaffiliated. But some like-minded individuals are considerably more powerful—particularly if they happen to be editors at major journals, or people assigned by those editors to review papers. A paper I contributed to, which included a description of various ways to conduct message-passing tests, was assigned a reviewer who claimed that the message-passing tests “have not undergone any empirical scrutiny” and that we should “state more clearly that it is currently unknown how often these authorship check strategies will conclude that messages are not genuine when in fact they are.”
This reviewer, as we wrote to the journal editor, seemed to want us to expand our paper to include an empirical defense of basic principles of experimental design. This left us wondering what it would be like if such a demand were made of studies in general; not just studies of FC. Studies examining the validity of telepathy might use designs similar to those we describe for FC: blinding the targeted recipient of the telepathic message to the message contents and to any possible cueing from the “sender”. Are the results of such studies unreliable until their empirical designs have somehow been empirically validated and their probabilities of drawing erroneous conclusions somehow calculated? What would such a meta-validation even look like, and how would we avoid infinite regress?
But of course, all this is beside the point. The point, at least for some people, is not to explore claims and test outcomes, but to erect fortifications—especially, apparently, when they pertain to FC.
REFERENCES
Kezuka E. (1997). The role of touch in facilitated communication. Journal of autism and developmental disorders, 27(5), 571–593. https://doi.org/10.1023/a:1025882127478
Popper, Karl (1962). Conjectures and Refutations: The Growth of Scientific Knowledge (2002 ed.). London: Routledge. ISBN 978-0-415-28594-0. excerpt: Science as Falsification
Quine, W.O. (1951). Two Dogmas of Empiricism The Philosophical Review 60 (1951): 20-43.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and Quasi-Experimental Designs for Generalized Causal Inference. (2 ed.) Cengage Learning.
Sheehan, C. M., & Matuozzi, R. T. (1996). Investigation of the validity of facilitated communication through disclosure of unknown information. Mental Retardation, 34, 94-107.
Wegner, D. M., Fuller, V. A., & Sparrow, B. (2003). Clever hands: Uncontrolled intelligence in facilitated communication. Journal of Personality and Social Psychology, 85(1), 5–19. https://doi.org/10.1037/0022-3514.85.1.5
Westcombe, A., (2019). I Do Not Think That Word Means What You Think It Means: A Response to Reber and Alcock’s “Searching for the Impossible: Parapsychology’s Elusive Quest.” Journal of Scientific Exploration, Vol. 33, No. 4, pp. 617–622, 2019
