There are real and fake smiles that can be distinguished visually, if you know how, and it turns out that there are real and fake laughs. William Hudenko, a clinician and researcher, patiently explained this to me at IMFAR 2009. Real laughs are "voiced" versus fake laughs that are "unvoiced" and these two kinds of laughs can be distinguished acoustically.
Hudenko et al. (in press) describe voiced laughs as having a "tonal, song-like quality" and as "strongly associated with positive affect," while unvoiced laughs are "largely atonal and noisier." Instead of reflecting a person's genuine emotions, unvoiced laughs are hypothesized to reflect various social signals.
When nonautistics laugh, about half the time their laughs are unvoiced. So how do autistics laugh?
Hudenko and his colleagues collected a lot of laughs from autistic children and two groups of nonautistic children (matched on chronological age, and matched according to vocabulary test age equivalents). Laughs were elicited in a 10-minute "laugh assessment sequence" in which "an examiner playfully interacted with each child."
The results? Autistics laughed just as much as nonautistics. The sole difference between autistic and nonautistic laughs was in proportion of voiced laughs. While on average, 97% of autistic children's "laugh episodes" were voiced, only 63% and 47% (age-equivalent and chronological age groups respectively) of nonautistic children's laughs were voiced. And about half the autistic children produced only voiced laughs.
You can find some autistic and nonautistic laughs here.
In their discussion, Hudenko et al. put forward this view:
...children with autism routinely produce fewer types of laughs than typically developing children because their laughter is more closely linked to their internal experience of positive affect.If this is so, then the autistic children in this study expressed more positive emotion--more genuine happy affect--in interacting with another person than did the nonautistic children. Hudenko et al. also refer to an earlier study which found, in the typical population, more positive responses to voiced versus unvoiced laughter. Nonautistics prefer voiced laughter.
Given prevailing standards in the autism literature (arising from prevalent standards of autism advocacy), no one should be surprised at how Hudenko et al. interpret their findings. The authors imply, in the absence of any evidence in this direction, that all this happy, genuine, engaging autistic laughter is unlikely to be socially "appropriate." Unfortunately, according to the authors, autistics "are not using laughter in a socially subtle manner." And here is the paper's unfounded concluding sentence:
In fact, by using laughter in a less social manner it may be that this expressive pattern actually contributes to the social deficits exhibited by children with autism instead of serving to facilitate connections with others.But the story doesn't quite end there. I ran into Dr Hudenko at IMFAR because he and one of his colleagues had a poster (abstract is here), a follow-up of sorts. In this new and as yet unpublished study, recordings of voiced and unvoiced autistic and nonautistic laughs were played for 135 nonautistic college-aged students. The students were asked to rate their "affective response" to each laugh on a scale from strongly negative to strongly positive. In a different task, the students were asked whether each recorded laugh came from an autistic or nonautistic child.
The results? The students rated their responses to autistic laughs as being significantly more positive than their responses to nonautistic laughs. Interestingly, this held true even when voicing--whether laughs were voiced or unvoiced--was accounted for. And when asked to do so, the nonautistic students could tell autistic and nonautistic laughs apart. The students performed above chance on this task, while only about one-fifth of them believed they could make this distinction.
So Hudenko et al. (in press) contend that autistics' way of laughing is defective and detrimental--a presumed contributor to autistics' presumed social deficits. This in turn implies that ideally, autistics would not have such engagingly positive, genuine and distinctive laughs, and instead should have the only "right" kind of laughter, the kind which characterizes nonautistics. But according to Dr Hudenko's IMFAR follow-up study, "improving" autistics this way would result in their laughter being less preferable to nonautistics than is currently the case.
Hudenko, W., Stone, W., & Bachorowski, J. (2009). Laughter Differs in Children with Autism: An Acoustic Analysis of Laughs Produced by Children With and Without the Disorder Journal of Autism and Developmental Disorders DOI: 10.1007/s10803-009-0752-1
Addendum: This post is included in the 5th edition of Scientia Pro Publica (more information here), hosted by Pro-Science.