Why no one can learn their first language without engaging in joint attention behaviors
In my last post I wrote about how FC proponents need to believe that autistic individuals (particularly non/minimal-speakers who use FC) are as socially motivated as non-autistic individuals. In this post, I’ll be returning to joint attention, the subject of several of my earlier posts. Like social motivation, joint attention is a topic that FC proponents are committed to having beliefs about. That’s because of the correlations between FC and minimal speech, minimal speech and autism severity, and autism severity and joint attention.
Those who are subjected to FC are minimal speakers; in autism, speech difficulties are correlated with autism severity. Autism severity is a function of severity of core autism traits; core autism traits include joint attention behaviors; core joint attention behaviors include following people’s eye gaze to attend to what they’re attending to.
What’s problematic for FC proponents are the myriad studies linking joint attention behaviors to language skills. Language skills include not just speaking skills, but also skills in comprehension and written language. If joint attention is a prerequisite for language skills in general, the severely autistic individuals who are subjected to FC are unlikely to have the skills to intentionally spell out the messages attributed to them or to understand what they’ve typed. This, of course, seriously undermines the legitimacy of FC.
But every once in a while an article comes along that claims that joint attention is not necessary for language acquisition, for example Akhtar and Gernsbacher (2007), which we critiqued here. Most recently, there’s an article by Schaeffer et al. (2023) entitled Language in autism: domains, profiles and co-occurring conditions. In this article, Schaeffer et al. review the literature on the various language challenges and the degree to which they are consequences vs. comorbidities of autism. While acknowledging that joint attention may facilitate language acquisition, Schaeffer et al. argue that it isn’t crucial. This conclusion is based on the following claims:
Joint attention may not be impaired at birth, but instead declines between 2 and 6 months.
Joint attention behaviors don’t predict subsequent language outcomes.
The language difficulties in autism instead result from “difficulties integrating speech sounds and mouth movements… as part of early atypical sensory perception.”
Many autistic children acquire fluent language.
There are reports of autistic children who acquire languages through watching TV, which does not involve joint attention.
An alternative route to learning language, instead of joint attention and social interaction more generally, is statistical learning, or “the ability to effortlessly detect patterns and regularities in (auditory, visual or visuo-motor) input without explicit instruction or intention to do so.”
But while there is a grain of truth to these claims/observations, each one is problematic.
As far as the first claim goes, what’s unimpaired at birth isn’t joint attention, but looking at people’s eyes. According to Jones & Klin (2015), whom Schaeffer et al. cite on this:
[E]ye looking--is not immediately diminished in infants later diagnosed with ASD [autism spectrum disorder]; instead, eye looking appears to begin at normative levels prior to decline. The timing of decline [at somewhere between 2 and 4 months] highlights a narrow developmental window and reveals the early derailment of processes that would otherwise have a key role in canalizing typical social development.
Among the processes derailed by declining eye looking is joint attention, which develops closer to 9 months. So much for joint attention starting out unimpaired.
As far as the second claim goes, it is true that other factors besides joint attention predict language outcome in autism. Two factors that have proved important are imitation skills and non-verbal IQ. Furthermore, there are exogenous factors like language interventions: What kind of language instruction did a child experience? How early did it begin? How intensive was it?
But the existence of other factors does not mean the joint attention isn’t also a factor. Which factor is most important varies, a function of at least four additional factors:
Which subpopulation within the autism spectrum we’re talking about (e.g., minimally speaking or only moderately language delayed)
Which aspects of language we’re talking about (e.g., speech or comprehension; basic nouns or phrases and sentences)
Which phase of language learning we’re talking about (e.g., early language learning or later-stage learning)
Whether we’re talking about first-language learning or second-language learning.
In Schaeffer et al.’s article, all of these factors come into play, and the result is a very confused discussion.
Some of the studies Schaeffer et al. cite find joint attention to be one of several factors in language outcome, with non-verbal IQ being the main other player (Anderson et al., 2007; Bennett et al., 2015; Thurm et al., 2015; Weismer & Kover, 2015). Anderson et al. find that, while joint attention is crucial for those who are the most language impaired, for those who are the most linguistically advanced, only non-verbal IQ predicts outcome. Bennett et al., which only measured correlates of joint attention, and did so indirectly via a parent survey (the “social” domain of the Vinland Adaptive Behavior Scale, or VABS-II), found social factors to be only important early in language development. Weismer & Kover looked at autism severity as measured by the calibrated severity score of the ADOS (the Autism Diagnostic Observation Schedule, a standard measure of autism symptom severity, including diminished joint attention). They found autism severity to be a significant predictor for language comprehension and production (where production means speech). But they also found that cognitive skills are a significant predictor, specifically, of language production. Thurm et al., finally, focused on minimally verbal preschoolers and on production—i.e., speech. They found that the social component of the calibrated severity score of the ADOS predicted who would develop “phrase speech”, or the ability to spontaneously combine multiple words into meaningful phrases. But they also found that when they took non-verbal IQ scores into account, only these were predictive of phrase speech.
None of these findings contradict longstanding conclusions about the special role that joint attention plays where it matters the most for language learning: at the earliest stages of language acquisition, where kids learn basic nouns. Learning the meanings of these early words—“Daddy,” “cup,” “dog,” “fish”—means looking at what speakers who use these words are pointing to or looking at. And looking where people are pointing or looking is the essence of joint attention. To learn basic nouns on your own, without explicit instruction, you need joint attention.
Not surprisingly, therefore, those studies that focus on the learning of basic nouns in autism (e.g., Baron-Cohen et al., 1997), find a key role for joint attention. When a researcher took an object out of her bag, looked at it, and labeled it, the non-autistic participants looked up, engaged in joint attention (i.e., looked where the researcher is looking) and learned the label. Not so with the participants with autism. They instead mis-learned the meaning of the noun label, assuming it referred to a different object: the one that they had in own their hands.
True, learning basic nouns is only a tiny part of language development. But it is also a crucial one. Basic nouns are the building blocks for the rest of language. You can’t learn verbs and prepositions before you learn nouns. You can’t learn abstract nouns before you learn basic nouns. You can’t combine words into novel sentences before you learn basic nouns. And it’s in learning basic nouns that you start seeing how language is a public communication system that plugs into the real world—into concepts and categories that are based on real-world phenomena. No one can master their first language without going through an initial phase in which they learn how to map basic nouns to the objects and categories that people conventionally use them to refer to.
But Schaeffer et al. question Baron-Cohen et al.’s conclusions. They point out that the autistic participants, even though they failed to engage in joint attention, must have already learned some words. That’s because they were matched with the non-autistic participants on vocabulary skills. And so, Schaeffer et al. conclude, autistic individuals must be able pick up the meaning of words without joint attention. But they overlook four important details:
Joint attention isn’t an all-or-nothing phenomenon. Individuals with autism often do show joint attention behaviors; just significantly less frequently, sometimes considerably less frequently, than non-autistic people do.
In particular, even if an autistic person never engages in joint attention when someone labels one object when they’re preoccupied with another (as was the case in the Baron-Cohen experiment), they may still engage in joint attention when they aren’t preoccupied with another object and/or when someone explicitly calls their attention to something (e.g., “Hey, look—a fish!”, accompanied by an explicit pointing gesture).
Even those who rarely engage in joint attention may still learn words through explicit instruction (including through explicit pointing, but also through computerized instruction, where joint attention isn’t an issue).
Finally, the participants in Baron-Cohen et al.’s study were compared to children with cognitive impairments, not to typically developing children.
Schaeffer et al. also note that “there is no unambiguous evidence that intervention techniques specifically targeting joint-attention skills have an effect on language outcomes.” But there’s a simple reason for this that has nothing to do with a lack of connection between joint attention and language acquisition: joint attention skills turn out to be extremely difficult to teach.
Now let’s turn to Schaeffer et al.’s third claim: that the language difficulties in autism instead result from “difficulties integrating speech sounds and mouth movements… as part of early atypical sensory perception.” These “sensory” difficulties, though, may be superficial manifestations of deeper social difficulties: multiple studies, for example Jones and Klin (2008), show that autistic infants as young as six months old have reduced rates of orienting to social stimuli, including to voices and mouths. These difficulties are precursors to difficulties with joint attention.
Schaeffer et al.’s fourth claim, that many autistic children acquire fluent language, overlooks how this lends support for a role of joint attention in language skills. Greater linguistic fluency is associated with milder autism, which in turn is associated with more frequent joint attention skills. The more frequent the joint attention, the faster the acquisition of those basic nouns that are the foundation for the rest of language learning, and the sooner the rest of language learning is set in motion.
Schaeffer et al.’s fifth claim concerns autistic individuals who acquire languages exclusively by watching TV. Watching TV, of course, doesn’t involve real-world joint attention behaviors. Nonetheless, there are several problems with the language-learning-via-TV-watching data as grounds for conclusions about the role of joint attention in language learning.
The first is that none of Schaeffer et al.’s references on language-learning-via-TV-watching involve what most people are referring to when they talk about joint attention in the context of language learning: individuals learning their first language. Most people, even children, learn second languages quite differently from first languages, and often in ways that involve fewer-to-no joint attention behaviors.
One of Schaeffer et al.’s references (Vulchanova et al., 2012), however, does provide a quite striking case study: a Bulgarian girl with autism who, by watching German-language children’s programs on a German TV channel starting at age 3, learned enough German by age 9 to show “advanced levels of German competency” and to be enrolled in a German-language immersion school. However, she was also advanced in her native language, which she learned in the real-world environment. Thus, she’d already undergone that joint attention-mediated process of learning how language is a public communication system that plugs into real-world objects and categories. Secondly, she may be a highly unusual case: Schaeffer et al. cite no similar examples. Third, her astounding feat has reportedly been replicated by another girl who isn’t autistic: a Finnish six-year-old who gained a good command of spoken English after watching English-language cartoons for two years (Jylhä‐Laide, 1994). Third, TV-mediated language doesn’t rule out something like joint attention: camera angles permitting, one can still follow the characters’ eyes or pointing gestures over to the objects they’re referring to.
Another of Schaeffer et al.’s references, Kissine et al. (2019), cite their own study-in-progress of autistic children who acquire a second language “from exposure to the internet.” While noting that their data is still preliminary, Kissine et al. mention a twelve-year-old autistic boy who attends a French-language primary school in Belgium and whose familial environment is exclusively French speaking. This boy has acquired, via English-language videos on the Internet, a level of English on par with his French. Kissine et al. also cite Christopher, a linguistic savant who has “demonstrated an exceptional ability to learn new languages from limited exposure and with a preference for written input.” But again, learning second languages, particularly as an older person, is qualitatively different from learning a first language. Nor does it require that preliminary step of plugging language into the real world via real-world joint attention. Indeed, for second language learning in late childhood and beyond, a dictionary will do just as well, especially if you’re learning through written input. And, finally, some people excel at mastering second languages, including from written or recorded input alone—whether or not they have an autism diagnosis.
The other examples of TV-mediated, purportedly joint attention-free language learning are much less impressive: they’re all about picking up another dialect of an already acquired first language rather than learning a new language. There are autistic children in Tunisia, raised with colloquial Arabic, who learn the standard dialect of Arabic by hearing it on TV (Kissine et al., 2019). And there is the young Bulgarian girl who, in addition to learning German from TV, also acquired, presumably through TV as well, a much more formal, standard Bulgarian than the local dialect that surrounded her. But picking up dialect from TV does not require joint attention; only careful listening and accurate imitation—other skills that aren’t specific to autism.
The source for Schaeffer et al.’s final claim, that autistic individuals follow an alternative route to learning language, is Kissine (2021). Kissine proposes that autistic individuals are inclined toward statistical learning, or what he he elaborates as the “domain-general capacity to detect the distribution of structural properties.” This, he proposes, is the result of the “strong systemizing skills” that some (e.g., Baron-Cohen et al., 2009) have linked to autism.
But all Kissine’s examples of statistical learning in language acquisition pertain to formal, structural aspects of language that have no more to do with joint attention than do the analogous aspects of a musical composition or a computer language:
Carving the acoustic sound stream into distinctive sounds (phonemes)
Carving a sequence of phonemes into distinct words
Identifying and internalizing a language’s complex structures.
Contrary to what Kissine suggests, none of these learning targets has a well-trodden social route for those without autism, namely, one involving “communicative, referential intentions.” True, one might get better at carving a sound stream into phonemes by imitating other people: by “mirroring and then adapting the phonological templates used by adults.” But this kind of imitation does not require joint attention. Rather, it requires a different kind of attention that, as noted above, is also diminished in autism: attending to speech sounds. But attending to speech sounds is also required for acquiring phonemes through statistical learning.
None of this diminishes the importance of joint attention in the aspect of language for which it’s most crucial: the meanings of basic nouns in one’s first language.
Kissine concludes by saying:
It could be that, because of the unavailability of socio-communicative cues, at least some autistic individuals learn language exclusively by relying on probabilistic, associative processes.
This may work fine for processing speech into phonemes and words. But as for learning the meanings of words, as Simon Baron-Cohen et al. showed, relying on joint attention-free statistical associations can frequently lead you astray. Joint attention-free statistical associations may cause you to connect together two salient, co-occurring stimuli in the environment that aren’t, in fact, meaningfully related—e.g., the object you’re holding and the word that someone happens to be saying while you’re holding it—instead of looking up at the speaker and seeing what they’re using the word to refer to. The result is an incorrect mapping between the word and the world—and a failure to learn the word’s meaning.
Putting it all together:
Joint attention is essential for the first steps of language learning: learning the meanings of basic nouns in your first language.
Cognitive skills also play a role in language learning, particularly later-stage language learning and formal aspects of language.
Imitations skills are another factor (see also Pecukonis, et al. 2019, who look at minimally speaking autistic children).
Where speech is concerned, motor skills are important as well.
Joint attention skills, while only one of many factors, are a crucial first step in first-language learning. Unless you’ve learned the meanings of basic nouns, you can’t learn verbs, prepositions, abstract aspects of language, or how to put words together into sentences. And there’s no way to learn the meanings of basic nouns without looking where people are looking when they use them—i.e., without joint attention.
REFERENCES
Anderson, D. K., Lord C., Risi, S., DiLavore, P. S., Shulman, C., Thurm, A., Welch, K., & Pickles, A. (2007). Patterns of growth in verbal abilities among children with autism spectrum disorder. Journal of Consulting and Clinical Psychology 75.594–604. DOI: 10.1037/0022-006X.75.4.594.
Akhtar, N., & Gernsbacher, M. A. (2007). Joint Attention and Vocabulary Development: A Critical Look. Language and linguistics compass, 1(3), 195–207. https://doi.org/10.1111/j.1749-818X.2007.00014.x
Baron-Cohen, S., Baldwin, D. A., & Crowson, A. (1997). Do children with autism use the speaker’s direction of gaze strategy to crack the code of language? Child Development 68.48–57. DOI: 10.1111/j.1467-8624.1997.tb01924.x
Baron-Cohen, S. (2009). "Autism: The Empathizing–Systemizing (E-S) Theory". Annals of the New York Academy of Sciences. 1156 (The Year in Cognitive Neuroscience 2009): 68–80. Bibcode:2009NYASA1156...68B. doi:10.1111/j.1749-6632.2009.04467.x.
Bennett, T. A., Szatmari, P., Georgiades, K., Hanna, S., Janus, M., Georgiades, S., Duku, E., Bryson, S., Fombonne, E., Smith, I.M., et al. (2015). Do reciprocal associations exist between social and language pathways in preschoolers with autism spectrum disorders? Journal of Child Psychology and Psychiatry 56.874–83. DOI: 10.1111/jcpp.12356
Ellis Weismer, S., & Kover, S. T. (2015). Preschool language variation, growth, and predictors in children on the autism spectrum. Journal of Child Psychology and Psychiatry 56.1327–37. DOI: 10.1111/jcpp.12406
Jones, W., & Klin, A. (2008). Altered salience in autism: Developmental insights, consequences and questions. In E. MacGregor, M. Nunez, K. Cebula, & J. C. Gomez (Eds.), Autism: An integrated view from neurocognitive, clinical and intervention research (pp. 62–82). Blackwell.
Jones W. & Klin. A (2013) Attention to eyes is present but in decline in 2–6-month-old infants later diagnosed with autism. Nature 504(7480):427–431. https://doi.org/10.1038/nature12715
Jylhä‐Laide, J. (1994) Learning by Viewing: cartoons as foreign language learning material for children‐‐a case study, Journal of Educational Television, 20:2, 93-109, DOI: 10.1080/0260741940200204
Kissine M (2021) Autism, constructionism, and nativism. Language 97:e139–e160
Kissine M, Lun X, Aiad F, Bourourou R, Deliens G, Gaddour N (2019) Noncolloquial Arabic in Tunisian children with autism spectrum disorder: a possible instance of language acquisition in a noninteractive context. Lang Learn 69(1):44–70. https://doi.org/10.1111/lang.12312
Kissine M, Bertels J, Deconinck N, Passeri G, Deliens G (2021) Audio-visual integration in nonverbal or minimally verbal young autistic children. J Exp Psychol 150:2137–2157. https://doi.org/10.1037/xge0001040
Pecukonis, M., Plesa Skwerer, D., Eggleston, B., Meyer, S., & Tager-Flusberg, H. (2019). Concurrent social communication predictors of expressive language in minimally verbal children and adolescents with autism spectrum disorder. Journal of Autism and Developmental Disorders 49.3767–85. DOI: 10 .1007/s10803-019-04089-8.
Schaeffer, J., Abd El-Raziq, M., Castroviejo, E., Durrleman, S., Ferré, S., Grama, I., Hendriks, P., Kissine, M., Manenti, M., Marinis, T., Meir, N., Novogrodsky, R., Perovic, A., Panzeri, F., Silleresi, S., Sukenik, N., Vicente, A., Zebib, R., Prévost, P., & Tuller, L. (2023). Language in autism: domains, profiles and co-occurring conditions. Journal of neural transmission (Vienna, Austria : 1996), 130(3), 433–457. https://doi.org/10.1007/s00702-023-02592-y
Thurm, Audrey; Stacy S. Manwaring; Lauren Swineford; and Cristan Farmer. 2015. Longitudinal study of symptom severity and language in minimally verbal children with autism. Journal of Child Psychology and Psychiatry 56.97–104. DOI: 10 .1111/jcpp.12285.
Vulchanova, M., Talcott, J. B., Vulchanov, V., & Stankova, M. (2012). Language against the odds, or rather not: The weak central coherence hypothesis and language. Journal of Neurolinguistics 25.13–30. DOI: 10.1016/j .jneuroling.2011.07.004.
