When “Evidence” Requires Harm: GLP Beyond the Gold Standard

Reframing proof through lived trajectories, ethical research, and the refusal to sacrifice autistic children for corporate science.

What ABA calls “no evidence” is really refusal to look. To demand RCTs for GLP is to demand harm. Our scripts, mitigations, and lived literacy are already the proof—evidence sung, not manufactured.

Introduction: Setting the Stage

I have written before about the tricks that get played with the word “evidence.” In No Evidence Means We Refused to Look, I pulled apart the way enclosure works—how the bar is set at randomised control trials (RCT), how everything else is dismissed as anecdote or bias, how autistic knowledge is swept aside because it does not fit the template. In When the Evidence Is Lived, I turned the lens inward, showing what it meant to grow into literacy through a gestalt mind. That piece was not theory but testimony—evidence inscribed in a body, in shelves of books once hostile and now home.

Together those essays set the stage for what comes next. Because if the gatekeepers demand an RCT, if the establishment insists that only their gold standard can count, then it is worth asking plainly: what would it take to run such a trial? What would the experiment look like if gestalt language processing were forced into the narrow channels of behaviourist science?

The question is not abstract. The ABA and Verbal Behaviour traditions rely on precisely this manoeuvre. They hold up their RCTs as the final word, as though documented histories of harm could be washed clean by the glow of statistical significance. They point to the absence of trials in GLP and declare the field empty, implausible, unworthy of trust. And so the challenge becomes sharper—what kind of trial would satisfy their demands, and what would it reveal, not only about GLP children but about the politics of evidence itself?

It is a thought experiment, but one worth dwelling in. Because to map it out carefully is to show the impossibility at its core. To imagine the mechanics is to glimpse the ethical fracture that the system tries so hard to conceal. And to hold that fracture up to the light is to see science not as neutral discovery but as power—the power to define what counts, the power to erase what does not, the power to claim that absence is emptiness rather than refusal.

This essay, then, begins at that uncomfortable edge. Not to legitimise their demand, but to expose it. To say: here is what your gold standard would look like, and here is why it could only be built by harming the very children you claim to protect.

The Testable Questions

If one were to imagine such a trial, the first question would be deceptively simple: what exactly are we testing? On paper it could be framed as a comparison between a GLP-informed intervention and either ABA’s Verbal Behaviour model or the usual business-as-usual instruction offered in schools. The design is neat—two arms, two groups of children, one protocol to rule them all. Yet the neatness is an illusion, because the outcomes that matter to gestalt learners are not the same as the outcomes that dominate behaviourist research.

For gestalt processors, progress shows itself in mitigations—in the way a once-rigid line is softened, reshaped, bent into new contours. It shows itself in recombinations—phrases drawn from different scripts joined together to make something novel, personal, alive. It shows itself in contextual flexibility—whether an echo once bound to one setting can migrate and make sense elsewhere, whether the child can carry meaning across thresholds. It shows itself in prosody—the rise and fall of the voice, the melody that carries emotional truth long before syntax is stable. And it shows itself in affect regulation—scripts used to calm, to protest, to laugh, to connect. These are not secondary features. They are the substance of growth, the language of a gestalt trajectory making itself known.

But the yardsticks that reign in ABA research are altogether different. They measure mean length of utterance as though longer strings of segmented words prove more competence. They count prompt compliance as though the ability to follow instructions on command is synonymous with language. They tally discrete responses, chart fidelity to drills, calculate speed of segmentation. In their world, the echo is a nuisance, the script a failure, the melody a distraction from the real work of chopping speech into analysable parts.

Here lies the mismatch, and it is not a small one. If the trial were to be designed with analytic measures as the only admissible outcomes, then the result is already written: the GLP approach will fail to show “evidence,” because the things it nurtures are invisible to the chosen instruments. It is more like bringing a sundial and a farmer’s almanac into the pool, then insisting that because the swimmer does not match the shadows or the seasons she is not really moving at all. The instrument itself predetermines the verdict—built for another world, another rhythm, blind to the motion in front of it.

So when we ask what could be tested, the answer is twofold. Yes, we can imagine outcomes for GLP, outcomes that capture the subtleties of mitigations and recombinations, the real markers of growth for gestalt learners. And yes, we can describe the analytic measures that would be imposed instead, the rulers calibrated only for a different kind of child. But once the difference is clear, so too is the injustice. The choice of measure is never neutral. It carries within it the politics of recognition and erasure—what will be allowed to count, and what will be dismissed as noise.

A Small Digression …

I do not write from the outside. I am not a casual observer throwing stones at the citadel. I am a PhD, and I chair an independent Institutional Review Board registered with HHS. To reach that point was not a short journey. It took years—hundreds of credit hours—just to reach the place where a dissertation proposal could be written and defended. It took study in theory and method, training in design and analysis, coursework in statistics and epistemology, long apprenticeships under supervisors who asked for rigour as well as imagination. It took the design of a study that was judged sound enough to proceed, the collection of data, the careful weaving together of findings into analysis, and the final defence before a committee who had the power to say yes or no.

This is not incidental. It matters because the training is as much about ethics as it is about numbers. When I step into my role on the IRB, I am carrying not only my own research history but also a deeper body of knowledge dedicated to protecting human subjects. Especially children. Especially those already made vulnerable by power. We train to notice what others might prefer to ignore: coercion dressed up as consent, harm disguised as “minimal risk,” the weight of historical precedent when research goes unchecked.

And the history is heavy. The entire apparatus of IRBs was built in the wake of atrocity—the experiments carried out in the name of Nazi race science, the human trials inflicted on prisoners and children in Japan, the long shadow of Tuskegee. Out of those horrors came a recognition, fragile and contested, that human beings cannot be reduced to raw material for data. That children, in particular, cannot be shuffled into control groups for the sake of clean results. The training is designed to keep those lessons alive. To remind us that behind every protocol number is a person, and that the first duty of research is to do no harm.

So when I speak about what an RCT in GLP would entail, I do so with that lens. I know what the field requires for a trial to be deemed valid. I also know what is required ethically to make research defensible. And I know, perhaps more sharply than most, what happens when the balance tips—when the hunger for proof outweighs the obligation to protect.

The Hypothetical RCT: A Cold Mirror

But I digress.

Let me return to the cold mirror of the trial itself, the one that would finally satisfy the gatekeepers. In their world, it begins with random assignment—whole classrooms split by the coin toss. One set of children receives a GLP-informed approach: scripts welcomed, mitigations scaffolded, recombinations celebrated. The other set is left in the familiar confines of ABA drills, their speech chopped into units, their gestures timed and corrected, their compliance tallied. Neat symmetry, at least on the page.

Yet before the first assignment can even happen, the recruitment hurdle appears. GLP is not recognised in U.S. diagnostic categories. There is no code, no checkbox on an intake form, no official designation to pull from the database. So the screening must be retrofitted—teachers and parents interviewed about whether the child uses echoes, mitigates lines, weaves prosody into regulation. Naturalistic language samples are collected and coded for mitigations, recombinations, contextual flexibility. Only then, after this unsanctioned labour, would the cohort even exist. Already the system is resisting recognition, ensuring that GLP children remain uncounted.

Once recruited, the trial would adorn itself with all the trappings of validity. Sample size powered to detect differences. Fidelity checks with observers rating videos to ensure therapists and teachers are following the manual. Statistical models stacked to account for clustering, for attrition, for noise. The kind of apparatus that looks, from a distance, like science in its purest form.

On paper, this is the study that would silence critics, the one that would finally “count.” But step closer and the violence becomes visible. To design it is to withhold language supports from half the children. To randomise is to declare their voices negotiable. To insist on fidelity is to codify erasure as protocol. This is the mirror the system refuses to hold up to itself—that the only way to produce the evidence it demands is by harming the very children who are supposed to be protected.

The Limitations and the Harm

The first and most glaring limitation is that there is no true equipoise. In medical research that principle means uncertainty—researchers cannot ethically withhold a treatment if there is already reason to believe it will help. With GLP, that uncertainty is not genuine. We know that welcoming scripts, honouring mitigations, and supporting recombination allow children to thrive. To place some in an arm where those supports are stripped away is not neutral—it is to risk harm knowingly.

Children are not laboratory materials to be shuffled for statistical neatness. They are not interchangeable units to be dropped into a control group for the sake of a cleaner dataset. Each child carries a way of speaking, a way of meaning, a way of belonging that deserves respect. Randomising them as if they were identical samples denies that truth and turns them into fodder for a design that flatters the journal more than it protects the person.

Here the ethical inversion shows itself in full. Behaviourist interventions—drills, compliance regimes, programmes with documented histories of trauma—still count as evidence because they have RCTs stamped upon them. The harms to autistic children are ignored, the trials are praised for their rigour, and the cycle continues. Meanwhile GLP, dismissed for lacking the same stamp, is treated as suspect not because it fails children but because we refuse to harm them in the same way. What is named an ethical strength—protecting children from deprivation—is recast as a fatal weakness.

And then there is recruitment bias, baked in from the start. Because GLP has no diagnostic code, no institutional recognition, the population is made invisible. Screening has to be improvised, patchy, reliant on teachers and parents who notice the patterns. The result is predictable: the population is declared too elusive, the trial unfeasible, the evidence absent. But this absence is not natural. It is engineered invisibility—a refusal to see, dressed up as methodological constraint.

Toward Ethical Research Designs

There are, of course, ways to generate knowledge without turning children into test subjects of harm. The most obvious is the stepped-wedge design, where every cluster eventually receives the intervention. Nobody is permanently consigned to the control arm. The comparison still exists—early versus later—but the ethics shift. It becomes a matter of timing rather than deprivation. Another is the longitudinal cohort, following children naturally over years, collecting language samples, observing mitigations, noting recombinations as they appear in the flow of ordinary life. Rich data, thick description, the kind of evidence that breathes rather than dissects.

Practice-based registries offer another path. Teachers and therapists already keep records, already record sessions, already log growth. Gather those across classrooms and clinics, standardise the way mitigations and recombinations are coded, and suddenly the evidence is not anecdote but corpus. The field grows in the open, not behind laboratory glass. And then there is the work best done in partnership: autistic-led mixed-methods studies, where lived experience is not tacked on at the end but braided through the design. Numbers and narratives, together, showing how language lives.

This is not an empty landscape. Marge Blanc’s 2012 book—the big brown book still resting on the shelves of countless speech therapists—has guided practice for more than a decade. It has become the touchstone for GLP-affirming therapy, not because it won an RCT but because it resonated with the realities therapists saw in their rooms. In my own field, I have written three books to help teachers recognise and support gestalt learners—No Place for Autism?, Holistic Language Instruction, Decolonising Language Education. They are not laboratory reports. They are invitations to small-scale action research, to classrooms made more affirming because teachers were willing to listen differently. That body of work is evidence in its own right.

Proof? Perhaps not in the narrow sense the journals demand. But confirmation bias? No. Ask the children whose scripts were welcomed instead of silenced. Ask the parents who saw their child’s language bloom when meaning was centred. Ask the professionals who shifted their practice and watched the difference unfold. Their testimony is not bias—it is the ledger of lived evidence, and it weighs more than any void engineered by enclosure.

Scarcity of RCTs, then, is not a fatal flaw but a reflection of ethical strength. It means we have refused to harm in order to produce numbers. It means we have chosen to protect children rather than subject them to deprivation for the sake of clean data. That refusal is not weakness. It is the real measure of what counts as care.

Analysis: The Politics of Evidence

The absence of RCTs is not a neutral gap, waiting patiently to be filled. It is an absence engineered, curated, and then weaponised. GLP is not dismissed because it lacks reality—children’s mitigations and recombinations, their scripts turned into bridges, are as real as any utterance. It is dismissed because to recognise it openly would threaten the market position of ABA and its many satellites. Verbal Behaviour has had half a century to entrench itself, and the so-called “Science of Reading” now tours with the same corporate fanfare. Together they have built their enclosures, fenced off the terrain, declared what counts and what does not long before I ever entered an American classroom as a marooned youngling. The field was already marked, the ground already carved into plots owned by others.

Evidence hierarchies, then, are not about truth but about territory. They do not simply ask what works—they ask who gets to say what works. They decide which forms of knowing are admissible, which measures are valid, which outcomes can even be named. And in that narrowing of the aperture, the full spectrum of autistic language is dimmed into invisibility. The enclosure is not accidental. It is deliberate. It is how a guild protects itself, how an empire of practice and profit maintains control.

Against this backdrop, I return to my own story, and to those of my students. A child who learned to read belatedly, who found belonging not in drills but in echoes, who grew into a writer and teacher because scripts were allowed to breathe—that is not bias. That is counter-ledger. It is lived evidence, written in the body and spoken in the classroom. My students, too, keep their own records: the day a script softened into flexible speech, the moment recombination turned into laughter shared, the slow settling of literacy when meaning took precedence over phonemes. These are not anomalies. They are the real flourishing that enclosure refuses to see.

So when the guild insists on absence, I answer with presence. When they declare void, I show fullness. When they dismiss testimony as bias, I name it as the counterbalance to their narrow ledgers. Because the politics of evidence has always been about power—and the only way to break that spell is to insist that autistic lives themselves are already the proof.

Final thoughts…

The impossibility here is not that GLP lacks evidence. The impossibility is that the system defines evidence in such a way that harm becomes the price of admission. To demand an RCT in this space is not to demand rigour—it is to demand an ethics violation. It is to ask that children be placed into deprivation so that journals may have their clean comparisons and corporations their funding streams. That is not science. That is sacrifice.

The way forward cannot be to mimic the mills, to churn out quasi-experiments polished until they gleam like proof. The way forward is to insist that our bodies, our voices, our trajectories are already evidence. The mitigations and recombinations of our children, the scripts used to self-soothe and to protest, the poetry we grew into against the odds—these are not anecdotes to be dismissed. They are data, alive and undeniable.

Absence of evidence is never emptiness. It is, more often, refusal to look. A void manufactured by narrowing the aperture until nothing can pass through. But autistic lives are not voids. We are not blank spaces waiting to be filled by corporate science. We are evidence already lived and sung, shelves already inhabited, classrooms already full of language if you have the ears to hear it.

This is what the enclosure cannot contain: that we speak, and that in our speaking we undo their silence. That is the only ledger that matters, and it is written in us.