Communication styles

On tests of figurative language, autistic groups score lower on average than non-autistic groups. A meta-analysis of 41 studies found a moderate difference (Hedges' g = −0.57).¹ Metaphors were relatively harder for autistic participants than irony and sarcasm, and age did not explain the size of the gap.¹ A second analysis by the same group, of 14 metaphor studies, found a similar gap.²

How the groups are matched matters. When they were matched on language ability, the difference was small and not significant.¹ A separate group looked only at studies matched on age and on verbal or overall IQ. For metaphor it still found a medium-to-large gap in accuracy, and slower answers in the studies that timed them; verbal ability explained most of the variation, and the better participants' verbal skills, the smaller the gap.³ For idioms and proverbs (11 studies, 235 autistic and 224 non-autistic participants) the gap was medium, unrelated to age, and tended to be smaller in people with higher verbal ability.⁴ In a 2012 commentary, written before these meta-analyses, Gernsbacher and Pripas-Kapit argue that autistic people who find figurative language hard also find language in general hard, and that there is no figurative-language difference specific to autism.⁵ The later evidence is split: the language-matched analysis fits that view, while the IQ-matched analyses still found gaps, and a study of adults described below found differences in implied meaning with no difference in grammar.^1,3,7

In everyday talk, much of what people mean goes unsaid. One research group built online tests of implied meaning for adults. With core language ability taken into account, autistic adults (66) were over twice as likely as non-autistic adults (118) to choose an unexpected reading of an implied meaning, and over five times as likely to answer "do not know".⁶ A follow-up suggested the "do not know" answers reflected a preference for certainty and explicit communication: when the question was more constrained, autistic participants often reached the expected reading.⁶ Their preregistered test of this idea did not bear it out.

In that study of 102 autistic and 109 non-autistic adults, the team expected confidence to be much lower than accuracy would explain; instead, autistic participants were slightly less accurate and slightly less confident by similar amounts on implied meaning, and did no worse on grammar.⁷ In an exploratory comparison, 49 adults with self-reported reading difficulties showed the reverse pattern.⁷ Discomfort with uncertainty went along with lower scores and with more self-reported difficulty in conversation, a correlation only.⁷ In another of their samples, autistic adults had stronger vocabularies than non-autistic adults but lower scores on a combined measure of understanding communication. Among the 71 autistic adults, lower scores moderately predicted more face-to-face communication difficulty, both self-reported and rated by researchers, though the two ratings did not agree with each other.⁸ In an exploratory analysis the authors say was too small for comparing genders, autistic women tended to score higher than autistic men, though they reported similar everyday communication difficulties.⁸

Irony gives mixed results. In a study of 24 autistic adults acting out what they heard, indirect requests (such as "it's cold in here" meaning "close the window") were understood as well as by non-autistic adults, while irony was clearly harder.⁹ When reading written irony, autistic adults were as accurate as non-autistic adults, though they spent longer rereading both ironic and literal passages.¹⁰ Another reading study found that autistic adults did not show the pattern non-autistic adults showed when anticipating how a criticized character would feel, or what the critic intended, after ironic rather than literal criticism.¹¹

In an online focus group of nine UK autistic adults aged 23 to 70, being honest, accurate and direct mattered a great deal, and participants were frustrated when others were not.¹² The authors report that participants saw euphemism and sarcasm as getting in the way of communication. Interactions with a clear purpose, such as a task, a shared activity or passing on information, were easier than others.¹²

A Conversation Questionnaire was built from what autistic people themselves said they find hard. In 312 autistic, dyslexic and non-autistic people, its items fell into two groups: not knowing what to say, and habits that can disrupt a non-autistic style of conversation. Both separated autistic from non-autistic respondents.¹³ In a study that followed five autistic adults talking with non-autistic partners for 4 to 12 months, the autistic participants named strategies that helped them, about the setting and about the type and structure of talk.¹⁴

Small talk is often described as hard for autistic people. In 95 interviews with autistic young people and adults during COVID-19 lockdowns, mostly women and mostly in Australia, participants described missing casual contact with acquaintances and strangers as well as their close relationships.¹⁵ Comparing these accounts with earlier research on non-autistic people (the study had no non-autistic group), the authors suggest these "weak ties" serve similar purposes, including wellbeing.¹⁵ Many valued casual contact built around an activity or hobby, and many said they "suck at small talk".¹⁵

In conversations between two autistic adults or two non-autistic adults (7 pairs of each, all German speakers), the timing of turns did not clearly differ overall, and both kinds of pair preferred very short gaps. In the first part of the conversation, autistic pairs left much longer gaps.¹⁶ Noise is a separate difficulty: in interviews with nine autistic adults, most described real trouble following speech against competing sounds, and said this was distinct from social difficulty, though the two could add up.¹⁷

The channel matters too. In a survey of 245 autistic adults, email was preferred for dealing with services, and phone calls were very unpopular. With friends and family, both face-to-face and written messages were preferred, and the preference for face-to-face depended on how close and accepting the relationship was.¹⁸

Echolalia is repeating other people's words, straight away or later. It is common in autistic children and found in some older autistic people.²⁰ A scoping review of 28 studies found it serves several purposes for autistic people, that clear definitions are lacking and views on its purposes conflict, and that many interventions still aim to reduce or stop it.¹⁹ A systematic review of 15 intervention studies in autistic children found that interventions generally reduced echolalia, but the studies were of very low quality, defined it inconsistently, and none treated it as meaningful.²⁰ An earlier narrative review found that differences in how echolalia is defined and understood have led to very different ways of managing it.²¹

Adults have rarely been asked. In interviews with five autistic adults, echoing was described mainly as a way to regulate oneself and less often as a way to interact. They also linked it to music: hearing the sound of words more than their meaning, and echoing songs.²²

A popular approach called "gestalt language processing" treats delayed echolalia as the raw material from which some children build language. A critical analysis concludes that the term is poorly defined, that there is no evidence it names a real clinical group, and that the related therapy protocol lacks empirical support.²³ Two researchers replying to a call to adopt it argue that doing so is premature, and that delayed echolalia should be kept separate from the wider set of ideas.²⁴

Scripts in a looser sense, prepared lines and planned conversations, appear in research on camouflaging. A camouflaging questionnaire built from autistic adults' accounts, and tested in 354 autistic and 478 non-autistic adults, includes the item "I have developed a script to follow in social situations (for example, a list of questions or topics of conversation)".²⁵

In interviews with 15 US autistic advocates, interests mattered to people for their own sake and also helped them communicate and connect with others.²⁶ In interviews with 20 autistic adults, speaking and non-speaking, themes included sharing what one values and simply being present together; one participant said "we infodump a lot about special interests to each other".²⁷ In the UK focus group, one participant said close friends and other autistic people were more tolerant of long, passionate talk about special interests.¹²

Interviews with 22 UK autistic adults aged 18 to 57 found preferences for parallel play, being close by without needing to interact, and for playing with other autistic people. Solitary play was valued as restorative.²⁸ A 2025 position paper argues that research on autistic communication should pay more attention to conversation built around shared interests.²⁹

Autistic writers use "situational mutism" for speech that is available in some situations and not in others. Khadija Gbla writes that in situations where they are overwhelmed and anxious, "as a trauma response my brain can decide that not speaking is the best course of action", and that "sometimes I just can't speak, not because I don't want to, but because I can't."^c20 Kimberly Gerry Tucker, who has a diagnosis of selective mutism, describes speech that "takes a freefall down my throat like a runaway elevator".^c21 Other writers say "going nonverbal" for the same loss of speech, not for an inability to understand language or to communicate another way: Danielle, interviewed by Kala Allen Omeiza, has selective mutism and does not speak "especially in times of stress".^c22

The closest adult research links the experience to setting and to emotion. A survey of 165 autistic adults aged 18 to 35 found that a little over half said the setting changed how they communicated: noisy or unpredictable places went with losing words, falling back on scripts and withdrawing from the conversation, while quiet, familiar places made open talk easier; participants reported that stress, frustration and anxiety may make communication harder, and that positive emotions improved fluency and confidence.³⁰ A scoping review of 18 studies of communication during distress found that 11 described ways of communicating without speech, including gestures, facial expression and behavior such as screaming and hitting; about half the studies included autistic people's own accounts, and the things reported to help were a calm, supportive setting and aids such as a pain scale.³¹

Selective mutism, the clinical diagnosis, and autism overlap, though the diagnostic manuals (ICD-10 and ICD-11) do not allow both diagnoses together. In a Norwegian register of 1,682 children and young people aged 3 to 18 with a selective-mutism diagnosis, 11.7% also had an autism diagnosis, and the authors argue that the rule against diagnosing both may delay or misdirect diagnosis.³² A systematic review of selective mutism in adolescents and young adults, pooling 40 studies and 54 participants aged 10 to 24, found that it starts later, takes longer to be diagnosed and is harder to treat than in children.³³ A case series described four girls first diagnosed with selective mutism in preschool, all of average-range intelligence, who were later diagnosed autistic.³⁴ These studies are of young people with selective mutism, not of autism in particular, and that literature is small; the adult evidence on situational mutism remains the survey above.

Unreliable speech has also been described as a motor difficulty rather than a cognitive one. An autoethnography by three minimally speaking autistic self-advocates describes apraxia, a difficulty with planning movement, as a broad motor difference, and argues that unreliable speech should be read as a motor difficulty, not as a lack of understanding.³⁵ A separate study of 54 low or minimally verbal autistic people aged 4 to 18, rated by speech-language pathologists, reported signs of a motor speech disorder, including reduced clarity and less precise consonants and vowels, and the speech difficulties went with language skills; the authors say the finding needs replication.⁴⁵

The label covers more than one mechanism. Selective mutism is tied to anxiety, speech can also be lost in distress or shutdown, as the survey and review describe, and unreliable speech can have a motor cause. Writers use one term for all of them.

Families of autistic children are still advised to use one language. A 2026 perspective paper says service providers continue steering multilingual families toward the dominant language at home, which cuts children off from family and community, and traces the advice to outdated ideas rather than evidence.³⁶ Thirteen Hebrew-English bilingual autistic children, interviewed for one study, described bilingualism as helping them connect at home, at school and in worship, and said they did not want to give it up.³⁷

Studies of children mostly find no harm. A scoping review of 15 studies of autistic children found that bilingual children did better than monolingual children on measures of working memory, flexible thinking and holding back impulses in most of the studies, though results were mixed for tasks that measured something similar in a spatial form, and some parent-rated measures showed no difference; all 15 studies looked at children at one time.³⁸ Among 396 children and adolescents aged 5 to 16, autistic children with a second language had better parent-rated executive function than autistic children without one, with no negative effect even among the 18 who also had an intellectual disability, and no difference in everyday communication; the authors say that subgroup was too small to be sure.³⁹ Among 168 autistic and 262 non-autistic children, better second-language proficiency went with better attention and memory.⁴⁰ Language itself has rarely been compared. Among 23 autistic and 26 non-autistic bilingual children aged about 4½ to 9½ learning English as a second language, the autistic children scored lower on most vocabulary measures, and more second-language input went with higher scores in both groups; a cluster analysis found most of the autistic children overlapping with the non-autistic children, and a smaller group with a lot of input and very low vocabulary.⁴⁶ A preprint that pooled 11 studies of 809 children reported a small-to-moderate association between bilingualism and executive function; it has not been peer reviewed, so it cannot carry the section on its own.⁴¹

The adult evidence is thinner. A survey of 459 adults, 132 of them autistic, found that people who used more than one language had lower scores on a self-report measure of autistic social traits, with a large difference in the autistic group; the study looked at people once, so it cannot show that one caused the other, and it found no link with the communication part of the measure.⁴² A second survey, of 177 multilingual adults, 98 of them autistic, found that the autistic adults reported knowing and using more languages than the non-autistic adults, but found moving between languages harder.⁴³

One small study points the other way: in 69 autistic children followed for three years, those in bilingual homes improved less on a rating of symptom severity than those in monolingual homes. The study was small, measured severity rather than language, and has not been repeated.⁴⁴

Many autistic writers describe their style as direct. Megan Griffith writes that many non-autistic people "stretch 70 percent of the way across the chasm" and expect the listener to cover the rest, while autistic people "tend to just give 100 percent of the info".^c1 Devon Price calls this "Autistic candor" and says that, carefully used, it has mostly been welcomed at work.^c2 Sol Smith says late-diagnosed autistic adults often use their first language "more like a second language", choosing words for their literal meaning, and rarely become fluent in the meanings people expect a listener to decode.^c3

Many autistic writers discuss small talk. Daniel M. Jones writes that it should be renamed "The Terrifying Time", and offers a rule: "If it's not worth saying, I won't bother saying it."^c4 Sarah Nannery describes realizing that small talk builds emotional connection, and that "it mattered very little what I actually said; more important was the way I said it."^c5 Megan Anna Neff contrasts social questions such as "Tell me about yourself", which she says often trigger "a mini freeze response", with conversations about a topic, like two people side by side looking at a sunset.^c6 Price suggests asking people about their hobbies and dreams in place of "stilted, masked small talk".^c7

Writers disagree about scripting. Neff writes that "Some scripting can be seen as a form of masking", but that scripts for setting boundaries or asking for what you need can become tools for self-advocacy.^c6 Price describes autistic people researching places and rehearsing restaurant orders, and having to hide how much they prepare; one interviewee sums it up as "Know enough, but not too much."^c8 Smith describes lying awake rehearsing the next day's conversations as a familiar autistic experience.^c9

Writers who speak little, or not at all at times, widen the picture. Bev Harp writes that "speech is not the same thing as language, and that communication is a much larger concept still"; for Harp, writing is easy and phone calls are the hardest.^c12 Melanie Yergeau writes: "I wish you wouldn’t interpret my silence as silence."^c13 On echolalia, Yergeau quotes a parent, Ariane Zurcher, recalling advice from the autistic professor Ibby Grace: not to translate her daughter's echoed phrases word for word, but to "feel the emotion behind the words".^c14 Orion Kelly describes bursts of repeated film quotes that soothe and that work as a detachment which "protects me from taking in any more stimulation".^c15

For connecting, Neff lists "Info swapping or info dumping" and parallel play among alternative love languages, alongside sensory comfort, respect for boundaries, written notes and predictability.^c6 She describes story swapping, answering someone's story with a similar one of your own, as a natural autistic way of showing empathy that non-autistic people can read as self-focus, and suggests turning the talk back to them afterwards.^c6 Steve Silberman recounts Jim Sinclair realizing that an autistic man reciting matching county names was using his special interest to connect, in what Silberman calls "a distinctively autistic medium of cultural exchange".^c10 Bianca Toeps writes that text gives her time to think and to make her point clearly.^c11 Research has touched infodumping, parallel play and story swapping only in small interview studies; preferences for written channels and the use of scripts have larger surveys behind them.^18,25

Nonspeaking writers describe how they communicate and what others get wrong about it. Ido Kedar, who learned to type independently on a letter board, keyboard or tablet, writes that years of drills treated a problem with movement as a failure to understand words: "It's not cognitive. It's motor."^c16 Typing independently, Kedar adds, "does not mean I need to be alone", because a good communication partner "helps us to organize our motor systems".^c16 Amy Sequenzia, who types to communicate, writes that she has loud hands: "I must, since I use my hands to communicate."^c17

Mel Baggs writes that typing on a keyboard feels "a lot more free" than talking, and argues that a person who can speak but finds it "painful, difficult, and draining" should also be allowed other ways to communicate.^c18 In another essay, Baggs describes autistic people picking up "the meaning of each other's movements" without words, and writes that "we generally do best at understanding people like ourselves" (see Double empathy).^c19