Who Gets to Speak? The Institutional War Over Nonspeaking Autistic Communication

Beau was 21 years old the first time he communicated a full sentence. He was in a session with a practitioner who had been working with him for some time, going through a lesson about Ernest Shackleton — the polar explorer who led his crew across Antarctic ice after their ship was crushed. The practitioner asked him: if you could name your family's boat anything, what would you name it? He spelled out: Absolute Love.

“That was the first time I got a glimpse into his inner world,” his mother Julianne said. “It has been so fun to get to know my son. So exciting. And I'm constantly getting to know him better.”

Beau is nonspeaking and autistic. He had spent more than two decades in a world that had largely decided, without asking him, that there was nothing much going on inside. His school kept him in a self-contained classroom doing remedial work — coloring, naming pictures, counting to 100 — until he aged out at 21. No one had suggested to his family that what they were seeing might be a motor-planning disorder rather than an intellectual disability: a disconnect between what his mind wanted and what his body could execute.

The method that finally gave Beau a voice is called Spelling to Communicate, or S2C. It is one of several related methods — along with the Rapid Prompting Method (RPM) and the broader category of letterboard-based communication — in which a nonspeaking person points to letters on a laminated alphabet board held by a trained communication and regulation partner, spelling out words, sentences, thoughts.

To the American Speech-Language-Hearing Association (ASHA), the private membership association that credentials speech-language pathologists, accredits the graduate programs that train them, and sets the profession's practice standards, it is something closer to dangerous: an unvalidated, “not recommended” method that its position statement warns may produce messages reflecting the partner's intent rather than the speller's own.

This stance creates immediate, practical roadblocks. Schools routinely ban letterboards from classrooms, insurance companies deny coverage for training, and courtrooms bar nonspeaking individuals from taking the stand in their own defense.

The bill of rights that wasn't

New York Senate Bill S7792 (NY A07363) began as a communication bill of rights. Championed by Assemblymember Angelo Santabarbara, whose own son Michael is nonspeaking and autistic, the Assembly version passed every committee and floor vote unanimously: 9-0, then 28-0, then 133-0 on the full Assembly floor. Nonspeaking advocates led the charge, backed by groups like I-ASC's Spellers and Allies Advocacy Network and Elizabeth Bonker's Communication 4 ALL. The goal was simple: ensure that no school or group home could strip a disabled person of their chosen way to communicate.

Then it moved to the Senate, where the companion bill was amended to define “validated” as methods that are “evidence-based, empirically supported and has demonstrated results in autonomous communication.” The amendment is a single clause, but it effectively excludes any method that does not yet have an established research base — which is precisely the situation S2C is in.

Reporting by City & State New York documented that the New York State Office for People with Developmental Disabilities — the agency that oversees group homes and services for disabled people, and that stood to lose oversight authority under the original bill — had raised concerns about whether a message produced via letterboard truly reflects the thoughts of the person with a disability or the person holding the board.

Santabarbara was unambiguous about what was happening: “The agency's happy with her version. They don't want to change what they have in place. Their bill sets us back — they will have complete control over everyone's communication device, everybody's communication method — and exactly the opposite of what everybody was fighting for.”

When Lora, a nonspeaking speller in California, heard about the amendments, she was quick to respond. “This wants to rob voices in New York, which will rob voices everywhere,” she spelled.

Lora started spelling five years ago, after her mother read a book about a boy's experience with the method. Within a month of daily practice, everything shifted. “One of the first words that she spelled was ‘heuristic,’” her mother said. “And I thought: how did she know this?” Lora's school laughed at the method; her mother pulled her out. She earned an accredited high school diploma through homeschooling and has since completed college coursework in sociology and statistics.

Asked what it felt like when her communication wasn't respected or supported, Lora spelled: To say it is mortifying is an understatement. It strips us of our humanity.

The body that won't cooperate

The controversy over letterboards hinges on a neurological reality: impaired performance of skilled, purposeful movement. The clinical term used in autism research is dyspraxia. In a 2007 study in Developmental Medicine & Child Neurology, Dziuk and colleagues at Johns Hopkins and the Kennedy Krieger Institute found that autistic children continued to show significantly poorer praxis than controls even after accounting for basic motor skill, and that praxis performance strongly predicted the social, communicative, and behavioral impairments that define the diagnosis.

Spellers and the clinicians who work with them describe the disconnect as extending across all voluntary movement — initiating a point, stopping one, holding a sequence together — which is why the partner's role is framed as motor and regulatory support rather than interpretation.

“Apraxia is like having a body that must do what it wants in spite of what my brain wishes,” Lora spelled. “Spelling set my voice free but my body loves to rebel and betray me at every turn.”

This matters for how we understand nonspeaking autistic behavior. The “behaviors” that have been treated as problematic for decades — self-injury, repetitive movements, apparent noncompliance — may in many cases be involuntary expressions of a body that won't cooperate with its owner's intentions, not chosen behaviors to be extinguished.

Betsy Hicks Russ, whose son Joe began spelling at around thirty, spent decades assuming he didn't understand what was going on around him. “The system that is in place is basically set up to make it look like that is the case — the way that assessments are done, the way that schools teach — they are made to reinforce these old thought patterns.” When Joe finally began spelling in open sentences, the first thing that came through was forgiveness.

The science that gets excluded

Some of the most prominent research supporting spelling-based communication comes from Dr. Vikram Jaswal, a developmental psychologist at the University of Virginia. In the first of two peer-reviewed studies, his team used eye-tracking to examine whether nonspeaking autistic spellers were initiating their own letter selections. Participants were spelling at roughly one letter per second with accuracy above 90%, and were consistently looking at letters before pointing to them — on average about half a second in advance.

“If the board is being moved underneath their finger, you wouldn't expect any systematic pattern in what they're looking at,” Jaswal explained. “But we found that they were looking at the letters they were about to point to, and doing it faster than the human information processing system could accommodate if they were responding to cues.”

The second study used a reaction-time paradigm to assess literacy without requiring explicit demonstration. Participants tapped letters as they pulsed on a screen; when the sequence spelled a sentence they'd just heard, they were measurably faster than when it was random. Roughly half showed evidence of foundational literacy — in a population that educational professionals estimated, at rates of fewer than 10%, could spell even a basic sentence.

Both studies drew published rebuttals in Evidence-Based Communication Assessment and Intervention, a journal co-edited by Ralf Schlosser, who also co-authored ASHA's position statement. Both were written by Katharine Beals, who co-founded FacilitatedCommunication.org. Beals first submitted her eye-tracking critique to Scientific Reports as a Matters Arising challenge; an editor declined it in December 2020 after consulting external advisers. It appeared in Schlosser's journal five months later.

Critics treat message passing as the decisive test of authorship. The problem, as Jaswal points out, is that no such studies have been successfully conducted — not because spellers fail them, he argues, but because the design presents significant challenges for a population with motor difficulties, high anxiety, and no experience with adversarial testing conditions.

“I'm puzzled by who decided message passing was the gold standard,” Jaswal said. “It's not applied to other AAC methodologies. No one subjects people learning to use picture-based communication to message-passing tests. And we know from research that with practice, nonspeaking autistic people can improve at this kind of study. That suggests the problem is lack of practice, not lack of ability.”

A systematic review often cited by opponents — conducted by Schlosser and colleagues — screened thousands of publications on methodological grounds but excluded all studies from its final evidentiary set, reaching a conclusion of “no evidence” by limiting the criteria to message-passing studies, of which there are none. “They designed the review to find what they already knew,” Jaswal said. “That the methodology they favor hasn't been applied. That's not the same as saying there's no evidence.”

The industry behind the science

Applied Behavior Analysis — the dominant therapeutic approach for autistic children — is now a more than $4.4 billion industry. Medicaid alone spends over $2 billion annually on ABA services, and all 50 states mandate insurance coverage. A single Board Certified Behavior Analyst overseeing a case bills $120 to $250 an hour for supervision; Registered Behavior Technicians bill $50 to $80 an hour for 10 to 40 hours a week. A single child's program can run $26,000 to over $166,000 a year.

Families and practitioners consistently report that once someone can express what they need, want, and feel, self-injurious behavior and dysregulation tend to decrease. Heather Burroughs, a former school psychologist and mother of Devyn, a nonspeaking 19-year-old, names the financial logic plainly: “If you can communicate, your behaviors go down. I hate to say it because it sounds so evil, but if you could get 40 hours a week when you're non-verbal and have bad behaviors, you're not gonna get that if you're communicating and regulated and have coping skills.”

Entrenchment is not the same as evidentiary standing. A Cochrane systematic review found the quality of evidence for Early Intensive Behavioral Intervention to be “low” or “very low” across outcome measures. Two peer-reviewed studies by Bottema-Beutel and colleagues found that 84% of ABA intervention papers had at least one author with an undisclosed conflict of interest; in the updated analysis, 93% of studies had such an author, only 8% disclosed any conflicts, and 93% of no-conflict statements were false.

The endorsed communication alternatives don't fare much better under scrutiny. A meta-analysis published in an ASHA journal found PECS produced small to negative gains in speech with concerns about maintenance and generalization. A systematic review tracking progression found most users never advance past the earliest phase — handing over a single picture. A meta-analysis of randomized controlled trials found no statistically significant effect on language development itself.

PECS is not an open method: it is a trademarked, commercially licensed system owned by Pyramid Educational Consultants, an approved ASHA continuing-education provider, meaning SLPs who pay for its workshops can apply that attendance toward the certification hours ASHA requires. ASHA produces over $75 million in revenue a year, nearly 70% from dues paid through the credentialing pipeline. That pipeline is what makes ASHA's endorsement commercially valuable.

Speech-generating devices require costly hardware, licensed software, ongoing clinical training, and continual reconfiguration. A 2024 multi-stakeholder study out of Murdoch University found that up to 50% of AAC users abandon their devices, with device fit rated a top-three barrier by every parent-carer surveyed.

Spelling-based methods have no comparable billing infrastructure. S2C is trademarked and I-ASC is its exclusive training source, as HALO is for RPM — but the cost is front-loaded: once a parent learns to serve as a communication partner, there is no recurring billable relationship, no reassessment cycle, no hardware to replace.

The classroom as battleground

By far the biggest obstacle faced by every family interviewed was the school — not individual teachers or aides, but the institutional machinery that assesses a child's perceived intellectual capacity and decides what accommodations are worth providing.

Heather Burroughs recognized this pattern long before the fight over letterboards. She had spent years administering IQ tests. “I would get the number and think it helped people get what they needed. Then I had Devyn. Her motor system was profoundly affected, she couldn't rely on speech, and I realized those scores weren't measuring what she knew. They were measuring what her body could demonstrate.”

In 2012, Burroughs entered a legal battle with her daughter's elementary school over Devyn's service dog, Hannah. The district concluded Devyn could never truly be Hannah's handler, and required Burroughs to supply a separate adult even though Devyn already had a one-to-one aide and nurse. Burroughs fought for eight years. When the U.S. Department of Justice joined the case, the resulting federal settlement required the district to recognize that students with disabilities may remain the handlers of their own service dogs while receiving reasonable modifications from staff.

“People see someone who needs assistance and assume the thoughts, choices, or abilities must belong to the person providing the support,” she said. “But that's backwards. The support exists so the person's own intent can be expressed. The real question isn't whether someone performs every physical step independently. The question is whether the support preserves that person's agency, intent, and control. That's what the ADA has always been about.”

Alex Le Pape's family spent 17 months making at least 33 documented requests to the Lower Merion School District in Pennsylvania to allow Alex to use a letterboard in school and to train staff to assist him. They sent videos. They brought in his speech therapist, his psychiatrist, his behavior analyst. They offered to pay for staff training themselves. The district declined, citing ASHA's 2018 position statement explicitly.

Alex, 16 when this began, developed self-injurious behaviors as the conflict escalated, and his psychiatrist recommended he be removed from school. The family filed for due process. The hearing officer excluded videos of Alex communicating and refused to allow him to testify, ruling that permitting his testimony would amount to deciding in advance that the method was legitimate.

The Third Circuit, in a 33-page precedential opinion issued in June 2024, held that “whether or not the letterboard was effective communication for Alex was a disputed material fact, and therefore a question for a jury, not a hearing officer or a judge applying deference to an administrative record.” A school district cannot unilaterally declare what counts as effective communication for a disabled student. But Alex Le Pape is now 23. He is not going back to high school.

The courtroom door that closes

There is another dimension to this story that has received almost no attention: what happens when the person whose voice has been silenced is also the victim of a crime.

Beau was 14 — before he had any access to spelling — when he was sexually assaulted in a public pool. He had enough spoken words to tell a caretaker what had happened. The police were called. But because Beau could not withstand cross-examination, there was no prosecution. The perpetrator was not charged.

“Predators know this,” Beau spelled. “They know that if there is not a secondary person that sees what happened, or a camera, they won't get prosecuted.”

This is not an isolated case. Elizabeth Zielinski, a special education advocate in Virginia, has encountered cases in which spellers were victims of crimes but could not offer testimony in a form courts would accept. Jennifer Binder-Le Pape described at least two such families in her network.

The standard objection is a case from the 1990s in which a facilitator's use of facilitated communication produced messages that did not originate with the student and led to false accusations against the family. It was a real harm, and families who use spelling methods do not minimize it. But several things are worth holding alongside it: the facilitator had approximately 30 minutes of training and would not meet current practitioner standards; the method involved a partner applying backward resistance at the communicator's hand, wrist, or elbow, which is meaningfully different from modern S2C practice in which the partner holds the board but does not touch the speller; and, as Jennifer Binder-Le Pape observed, the device at the center of that case was an early prototype of the modern AAC device, not a letterboard at all.

And fourth — an argument made by disability scholar Rua Williams that Jaswal cited — we might ask whether a false accusation is a greater harm than an inability to accuse, and whether the standard of protection extended to support partners is proportional to the protection withheld from nonspeaking people themselves.

Zielinski shared a recent win: a due process hearing in Fairfax County, Virginia involving a student who uses a letterboard, in which the hearing officer did not question the authenticity of his communication and the student testified on his own behalf. In a follow-up press event, video of the board was displayed behind him at triple speed. It did not move.

The standard unevenly applied

Independent authorship in spelling-based communication is a reasonable evidentiary concern. Critics are not wrong that influence from communication partners is a genuine methodological risk that needs to be studied and minimized.

But the standard is not applied to any method ASHA endorses. PECS begins with a two-person procedure in which a second adult stands behind the learner and physically guides their hand to pick up a picture and place it in a partner's palm, with that assistance faded over subsequent trials. Partner-assisted scanning, a standard AAC technique, has a partner present options aloud or by pointing while the user signals a selection. The vocabulary available on a speech-generating device is initially selected and programmed by clinicians or communication partners.

Institutions also do not ordinarily require conclusive evidence or universal effectiveness before allowing disabled people access to communication supports. PECS retains its endorsement despite outcome research showing most users never progress past handing over a single picture. Speech-generating devices retain theirs despite abandonment rates as high as 50%.

The question is not only whether letterboard communication has met an authorship standard that is not routinely applied to endorsed methods. It is also whether, in practice, the method is beneficial for the people who choose to use it — and how institutions should respond while the scientific questions remain unresolved.

When Beau was able to tell his doctors about symptoms he was experiencing, his family learned he had ulcerative colitis. When Alex Le Pape could tell his psychiatrist about his internal experience, he was able to get the psychological support he needed. When Joe could tell his mother what he needed, she finally had the gift, as she put it, of not guessing anymore. When Lora began communicating with her professors, she got an A in statistics.

“The gold standard,” Jennifer said, “should be giving them agency and a good life.”

What presumption of competence actually means

The phrase has a specific intellectual history. Anne Donnellan grounded it in what she called the criterion of the least dangerous assumption: when you cannot determine with certainty what a person understands or is capable of, you choose the assumption whose consequences are least harmful if you turn out to be wrong. Assuming competence and being mistaken costs little. Assuming incompetence and being mistaken costs a person their education, their voice, and, sometimes, their safety.

That principle has never been applied consistently to nonspeaking autistic people. “We certainly extend the presumption of competence to young children who are not autistic,” Jaswal said. “We should extend it to everybody, and recognize that people are on different developmental paths, and they're going to take different amounts of time, and we need different kinds of support. But if we don't provide those opportunities, then they're never going to acquire the kinds of skills they would like to.”

Fewer than 10% of nonspeaking autistic people are estimated by educational professionals to have the literacy skills to spell a sentence. Jaswal's research suggests the real number is between one-third and two-thirds.

“I always knew I was smart,” Lora spelled. “But the world didn't know. It felt validating to be heard for the first time.”

About this reporting

This article was reported over three months and draws on original interviews with nonspeaking autistic spellers Lora and Beau, their parents, Heather Burroughs, Jennifer Binder-Le Pape, Elizabeth Zielinski, and Dr. Vikram Jaswal of the University of Virginia. Court documents cited include Le Pape v. Lower Merion School District, No. 22-2931 (3d Cir. 2024). Research cited includes Dziuk et al. (2007), Bottema-Beutel et al. (2021, 2026), and Reichow et al. (2018, Cochrane Database).

The piece has been updated since original publication, including clarifications to the account of Heather Burroughs's federal service-dog case, an added section on the published rebuttals to Dr. Jaswal's studies, a clarified description of ASHA's institutional interest (ASHA does not earn revenue from PECS), and a corrected description of the motor differences underlying letterboard use — dyspraxia is the term used in autism research, and “full-body apraxia” is not a clinical designation.