In early June, Ally Betchan and her family made the monthly trek from their small central Texas town to a therapy center in Austin, hoping that she could learn to communicate. Like nearly 30 percent of people with autism, Ally is severely disabled and does not speak.

Ally, 22, sat quietly in a small room next to her instructor, Soma Mukhopadhyay, a sprightly 63-year-old who, by contrast, talked almost nonstop. More than 30 years ago, Ms. Mukhopadhyay taught her nonspeaking autistic son, Tito, to write and type independently, creating a communication method that supporters hailed as transformative and critics have challenged ever since.

Ms. Mukhopadhyay held up a clear plastic sheet marked with the alphabet, prompting Ally to make up a story. As Ally tugged rhythmically at her purse, she slowly pointed at letters to spell “DONNA KNOWS,” and then seemed to get stuck, pointing to a jumble of letters.

“I’m so lost,” Ms. Mukhopadhyay said, shaking the sheet and pressing her to try again. As Ms. Mukhopadhyay occasionally tapped under the letter board on her thigh or leaned in the direction of a letter, Ally eventually spelled: “CARING HURTS.”

“‘Donna knows caring hurts’ — that is a life lesson,” Ms. Mukhopadhyay said, nodding in agreement. Then, Ally jabbed many letters in quick succession, but distinctly: “SHE LOVES THOSE WHO CARE FOR HER.” Sitting beside her, Ally’s mother, aunt and grandmother smiled.

Ms. Mukhopadhyay’s technique, called the Rapid Prompting Method, or R.P.M., is one of several intended to help nonverbal people learn to communicate using letter boards held in midair by another person. At the core of these assisted spelling methods is a radical assertion: that nonspeaking autistic people, many of whom have been considered intellectually disabled their whole lives, may have typical or even extraordinary cognitive abilities, obscured by motor problems and an overwhelmed sensory system that has cut them off from the world around them.

Proponents of assisted spelling say it has improved the lives of thousands of nonspeakers, some of whom have used it to write memoirs or obtain graduate degrees. Yet despite the potentially profound implications of these communication methods, there has been remarkably little scientific research evaluating them. Citing the risk that the person holding the letter board may influence the messages, and a history of such abuses with prior assisted communication methods, many medical groups have cautioned against them.

All of this has led to a growing debate dividing autistic people, families and the scientific community. The central question is less about whether breakthroughs like Tito’s are possible, but whether they are as widespread as many proponents claim. Do assisted spelling methods reliably reveal a person’s own thoughts, or do they give families a false sense of their loved ones’ inner world and capacities?

That uncertainty has led some autism experts to argue that assisted spelling requires urgent examination.

“Currently there is this incredible impasse,” said Dr. David Amaral, the research director of the MIND Institute at the University of California at Davis, with true believers on one side and skeptics on the other.

“My bottom line is that there’s enough evidence that this does work for a subset of people that we should be understanding why it works, how it works and for whom,” he said. “We’re not doing that now — there just isn’t any science.”

‘Real’ or ‘Fake’?

Despite the lack of studies on the effectiveness of assisted spelling, the methods have steadily increased in popularity over the last decade. In April, a book written by a nonspeaking autistic man named Woody Brown, who was trained by Ms. Mukhopadhyay and communicates via a letter board held by his mother, shot up the New York Times best-seller list and sparked debates over the legitimacy of his authorship. (An interview on The Today Show showed Mr. Brown, who had received an M.F.A. from Columbia University, appearing to point at random letters as his mother spelled eloquent prose.)

This year, Health Secretary Robert F. Kennedy Jr. appointed several spellers and their parents to an autism panel tasked with setting federal research and policy priorities. In New York, a fight over a proposed Communication Bill of Rights that would enshrine into law the right to use assisted spelling has drawn support from relatives of Helen Keller. And perhaps most visibly, “The Telepathy Tapes,” a viral podcast released in 2024, made the extraordinary claim that some autistic spellers have telepathic abilities.

That podcast is what brought the Betchan family to Ms. Mukhopadhyay’s office last year. After hearing about assisted spelling on the show, the family began to wonder if Ally, silent for more than two decades, had something to say.

They rushed to the Dollar Tree to buy supplies for their own makeshift letter board. Then the family peppered Ally with questions: What town are you from? What street do you live on? How do you spell your mom’s name? “She knew them all,” said her mother, Elisha Betchan. “That’s when I thought there may actually be something to this.” Soon afterward, they started working with Ms. Mukhopadhyay, paying $110 for each 45-minute session.

Though Ally’s family is aware of the fraught debate over the legitimacy of assisted spelling, they said it was irrelevant given their firsthand experience. “We see what we see and we know,” said her aunt, Cherie Kocian. “We don’t need data to know that it’s real.”

Many of the parents of the six spellers I observed felt similarly. “It gives parents like us access to something that our children have never had,” said Ahmed Hashmi, whose son, Omer, 33, started using assisted spelling in 2024. He is now attending college with his father’s assistance.

Mr. Hashmi argued that the anecdotal experiences of families like theirs were also a form of evidence scientists should consider. “You cannot ignore it,” he said. “Science will play catch-up eventually.”

But the potential consequences of assisted spelling becoming widely used without scientific research to back it are top of mind for Howard Shane, director emeritus of the Center for Communication Enhancement at Boston Children’s Hospital.

In the 1990s, a method called facilitated communication exploded in popularity, promising miraculous access to the inner lives of people who were profoundly disabled. The method involved facilitators physically holding the arms or hands of nonspeakers as they typed on keyboards.

In 1992, lawyers in a sexual abuse case brought by a 16-year-old autistic girl named Betsy Wheaton asked Dr. Shane to evaluate the validity of the facilitated communication method used to elicit her allegations. Ms. Wheaton had claimed, through a facilitator, that her father and brother had sexually abused her.

In blinded tests, often referred to as message passing tests, Dr. Shane independently showed Ms. Wheaton and her facilitator images, sometimes matching and sometimes not, and asked Ms. Wheaton to type what she saw. Every object correctly identified was based on what the facilitator, and not Ms. Wheaton, had been shown.

“Everyone in the room, including the guardian ad litem, whom I trusted, knew the truth,” Ms. Wheaton’s facilitator, Janyce Boynton, who has since become a vocal critic of facilitated communication and assisted spelling techniques, later wrote. “F.C. was fake, and I was not the child’s facilitator. I was the one moving her arm.”

By the mid-1990s, despite a lack of corroborating evidence for many of the claims, more than 60 sexual abuse allegations had been brought through facilitated communication, resulting in some children being removed from their homes and parents being jailed. In 2011, a former Rutgers professor who had been a facilitator for a nonspeaking man with cerebral palsy pleaded guilty to sexually assaulting him. She continues to claim that he typed his consent to have sex with her.

Dr. Shane and other researchers went on to conduct dozens of blinded tests of facilitated communication, overwhelmingly finding that the nonspeakers were able to give the right answer only when their facilitators knew it.

“I never found anyone that passed,” Dr. Shane said.

An Untested Theory

While experts like Dr. Shane disagree, practitioners of the newer forms of assisted spelling insist that their approaches, which do not rely on touch, bear no relation to facilitated communication.

Ms. Mukhopadhyay’s method made headlines in the early 2000s when she and Tito moved from Bangalore, India, to the United States. While Ms. Mukhopadhyay had trained Tito to communicate using a letter board, he later moved on to writing and typing on his own. At age 11, Tito had published an autobiography that included works of poetry.

Language deficits in people with severe autism, Ms. Mukhopadhyay theorized, were a motor problem, not a cognitive one. Having a partner hold the letter board, she argued, provided motor support and steadied the dysregulation and sensory overload so common in people with autism, allowing them to communicate.

Many autism experts dispute Ms. Mukhopadhyay’s motor theory. They point to data estimating that more than a third of autistic children are considered intellectually disabled. And they question how nonspeakers, many of whom weren’t taught to read and write at school, could know how to spell in the first place.

Even without physical contact, the mere existence of a communication partner raises the possibility of influence. Skeptics frequently cite the example of Clever Hans, a horse whose owner at the turn of the 20th century claimed he could do complex math and could read and write in German. Later investigations found that the horse had accomplished these remarkable feats by instead becoming highly attuned to subtle cues from his owner.

Ms. Mukhopadhyay and other practitioners have also raised suspicions because they refuse to take part in blinded experiments like Dr. Shane’s that could reveal whether the communication is truly the subject’s own, arguing that such experiments put undue stress on severely disabled people to prove themselves to skeptics.

“We have a long history of questioning people’s communication,” said Elizabeth Vosseller, a former speech-language pathologist who trained under Ms. Mukhopadhyay before founding her own method called Spelling to Communicate based in Herndon, Virginia, which has more than 900 practitioners in 31 countries. Ms. Vosseller pointed to the recent history of deaf people being dismissed as cognitively impaired to argue that society should “presume competence” in people who do not speak.

A forthcoming systematic review found that not a single published study has evaluated independent authorship in assisted spelling.

The only blinded test of assisted spelling was published in 2018 before quickly being retracted. The study, which The Times obtained, found that a 17-year-old speller could respond to questions correctly only when his mother was aware of the answers. It also stated that the family had “liquidated their savings” to pay for assisted spelling trainings, including in Herndon. After the study’s publication, the family claimed it had violated their privacy and threatened legal action unless it was removed.

The refusal to conduct or participate in blinded studies is “the biggest red flag,” said Alison Singer, the director of the Autism Science Foundation and the parent of a minimally speaking daughter who does not use assisted spelling. The risk that the words might not be those of the nonspeakers is “horrifying,” Ms. Singer said. “Our children already can’t speak, they can’t communicate, and now you’re going to hijack their voice?”

Some parents who say their children have benefited from assisted spelling also acknowledge that it carries risks. Portia Iversen, who led the autism research nonprofit that first brought Ms. Mukhopadhyay and Tito to the United States, described R.P.M. as a “breakthrough” that allowed her son Dov to communicate for the first time.

But Ms. Iversen also recalled an incident in which an aide using a letter board with Dov threatened to call 911, claiming that Dov had spelled that he was going to jump out the window and kill himself.

“I saw how dangerous it was,” Ms. Iversen said, shuddering at the possibility that the claim could be used to take her son away. “I knew that was not Dov.”

Both Ms. Mukhopadhyay and Ms. Vosseller acknowledge that communication partners can steer responses.

“There are so many chances with the letter board where people can influence it,” Ms. Mukhopadhyay said. Often, she said, she sees parents guide the output out of enthusiasm, but she does not question their belief that the words are their children’s own. “It would be wrong for me to come in between the parent and the child,” she said.

But she brushed off the idea that these cases invalidated her method’s integrity. “If they misuse it, they misuse it — it’s up to them,” she said.

The ‘Intact Mind’

I briefly met Tito, now 37 years old, at the small townhouse where he and Ms. Mukhopadhyay live, near her office in Austin. He sat on the carpet with his mother by his side, hunched over his clipboard with a pen in his fingers. Ms. Mukhopadhyay read over his shoulder, nudging him when he became distracted or crossing out a misspelled word and urging him to start again.

I asked him what he did while his mother was at work. “Nothing,” he wrote. Then he changed the subject to something that was exciting to him. “I am going to New York tomorrow,” he wrote. Ms. Mukhopadhyay explained that she was holding a training session at a center in Long Island for older autistic adults. Tito finished his thought, writing, “I will look at the aging people and wonder how my last years would look like.”

Cases like Tito’s appear to be very rare. Of the thousands of nonspeakers who have trained with Ms. Mukhopadhyay over the years, she said only around 15 had gone on to write or type on a keyboard on their own.

But it’s cases like Tito’s that are also the clearest evidence that learning to use language is possible for some. That has led some scientists to argue that we understand too little about the causes of language difficulties in autistic nonspeakers to dismiss assisted spelling.

“Absence of evidence is not the same as evidence of absence,” said Alexandra Woolgar, a cognitive neuroscientist at the University of Cambridge. She argued that concerns about influence needed to be weighed against “the risk of underestimating people who genuinely do understand, who could do so much more if we allow them to communicate in a way that works for them.”

Dr. Woolgar is conducting her own blinded tests of assisted spelling, hoping to provide some data on the persistent questions around independent authorship. She argued that a broader unanswered question is why nearly 30 percent of people with autism have minimal or no speech in the first place.

“One possible reason is that their language is impaired,” Dr. Woolgar said. “Another possible reason has to do with motor planning or something along that very long pathway between having a thought and actually producing a word.”

To help answer that, Dr. Woolgar is conducting a study measuring electrical activity in the brains of nonspeakers to search for signs that they understand spoken language.

Only one published study has focused on the mechanisms behind assisted spelling itself, using eye-tracking data to show that participants looked at letters before pointing to them. But that study faced criticism from experts, who argued that the letter boards moved and could draw eye gaze toward different letters.

Though researchers acknowledge that there are some nonspeakers with overlooked cognitive abilities, Ms. Vosseller and Ms. Mukhopadhyay say that, except for very rare exceptions, all nonspeakers can learn to communicate. Ms. Mukhopadhyay told me that, in all her years of using R.P.M., she had “never met” someone who was intellectually disabled.

While it is difficult to test cognitive ability in people who can’t speak, the idea that no autistic people are intellectually disabled “speaks to the enormous seduction of the ‘intact mind,’” said Amy Lutz, a historian of medicine at the University of Pennsylvania whose son Jonah is minimally speaking and does not use assisted spelling.

That idea preys on parents’ deepest hopes for their children, she said.

“For some parents, providing literally a lifetime of intense, round-the-clock care to an adult child with severe cognitive, communicative and behavioral challenges can feel like pouring all your love and effort into a big bucket with a hole at the bottom,” Dr. Lutz said. “Believing that your child is actually brilliant and going to change the world, for those parents, can justify all that labor and sacrifice.”

Ms. Betchan, Ally’s mother, said that in the year since Ally has been doing assisted spelling and practicing with her at home, she has sometimes battled uncertainty over whether the words were her daughter’s own.

Sometimes, Ms. Betchan said, she gets the creeping sense that Ally is reading her mind. Other times, she attributes that feeling to its simplest explanation: that she is unconsciously guiding some of her daughter’s answers. But she is confident those doubts will go away with time.

“Just working with her every day,” Ms. Betchan said, “is enough evidence that I need for me to continue it.”