2.1. Low levels of eye contact

People look each other in the eye for a variety of social reasons – to acknowledge each other, communicate emotion, and coordinate visual attention.Footnote ³ Autistic people tend to engage in eye contact much less frequently than non-autistic people. Indeed, in infancy, reduced eye contact is one of the features that distinguishes infants who are later diagnosed with autism from those who are not (Elsabbagh & Johnson Reference Elsabbagh and Johnson2010; Oner et al. Reference Oner, Oner and Munir2014; Zwaigenbaum et al. Reference Zwaigenbaum, Bryson, Rogers, Roberts, Brian and Szatmari2005). One explanation for why autistic children and adults infrequently engage in eye contact is that they are not motivated to do so. For example, Chevallier et al. (Reference Jones, Quigney and Huws2012b) describe “a spontaneous disinterest in mutual gaze” (p. 235) as part of a suite of behaviors that indicate “diminished social orienting” (p. 231), which is in turn taken to indicate diminished social motivation. But there are alternatives to this social explanation.

Although sustained eye contact is often assumed to be a universal behavior among non-autistic people, there are striking cultural differences in the extent to which individuals engage in it (LeVine et al. Reference LeVine, Dixon, LeVine, Richman, Leiderman, Keefer and Brazelton1994). For example, among the Gusii of Kenya, “conventions of adult conversation involve much less mutual gaze” than among Westerners (Richman et al. Reference Richman, Miller and LeVine1992, p. 617). Among the Navajo, “looking someone in the eye while they are speaking is a form of rudeness and causes the Navajo speaker considerable discomfort” (Connors & Donnellan Reference Connors and Donellan1993, p. 273). Similarly, in China, direct eye contact was historically not common because it was considered “rude and arrogant” (Zhang et al. Reference Zhang, Wheeler and Richey2006, p. 112). Despite the fact that eye contact is less common in some other cultures, no one would propose that members of those cultures have diminished social motivation.

There are also cultural differences in the amount of eye contact between non-autistic mothers and infants. For example, Richman et al. (Reference Richman, Miller and LeVine1992) compared naturally occurring social interactions between Gusii mothers and their infants with those between suburban American mothers and their infants. For the American mothers, the most common category of social behavior directed toward their 10-month-old infants was “looking.” In contrast, the most common category of social behavior for the Gusii mothers was “touching”; “looking” was fourth of the five categories coded. We doubt that Gusii mothers are less interested in connecting with their infants than American mothers. Rather, as LeVine (Reference LeVine, Gielen and Roopnarine2004) has suggested, other behaviors, like touching and holding, may be “functionally equivalent to the verbal and visual engagement of Americans” (p. 161). This is a crucial point that is frequently acknowledged in cross-cultural comparisons but rarely in the autism literature: There are multiple ways to communicate social engagement (see sect. 5); gaze aversion may be a useful diagnostic marker for autism in Western cultures (Norbury & Sparks Reference Norbury and Sparks2013), but it does not necessarily indicate social aversion (Akhtar & Gernsbacher Reference Akhtar and Gernsbacher2008; Gernsbacher et al. Reference Gernsbacher, Stevenson, Khandakar and Goldsmith2008b).

In fact, looking away from another person's eyes has been shown to have several adaptive functions in non-autistic individuals. In early infant-caregiver interactions, for example, when infants become overstimulated, they often look away, which leads to lowered heart rates (Field Reference Field1981). In some cultures, caregivers avert their gaze when their infants are upset as a means of calming them (Dixon et al. Reference Dixon, Tronick, Keefer, Brazelton, Field, Sostek, Vietze and Leiderman1981). Doherty-Sneddon and Phelps (Reference Doherty-Sneddon and Phelps2005) argue that, because maintaining mutual gaze consumes processing resources, gaze aversion can help manage cognitive load: Non-autistic adults avert gaze when solving difficult problems (Glenberg et al. Reference Glenberg, Schroeder and Robertson1998), and training young children to avert gaze can improve their performance on cognitively demanding tasks (Phelps et al. Reference Phelps, Doherty-Sneddon and Warnock2006).

For autistic individuals too, both experimental evidence and first-person accounts suggest that gaze aversion can confer adaptive benefits. For example, in a classic study, Klin et al. (Reference Klin, Jones, Schultz, Volkmar and Cohen2002) showed autistic adolescents clips from the film Who's Afraid of Virginia Woolf and measured what they looked at and for how long. There was no relation between the amount of time they spent looking at the actors’ eyes and measures of social competence; however, the more time they spent looking at the actors’ mouths, the better was their social competence. Klin et al. argued that social information conveyed through speech may be easier for some autistic individuals to interpret than social information conveyed through the eyes. Concentrating visual attention on the channel that produces the social information they find most interpretable (i.e., the mouth) may help them gain a better understanding of the social world (see also Rice et al. Reference Rice, Moriuchi, Jones and Klin2012).

In fact, some autistic people say that they avoid looking at a speaker's face altogether to concentrate on what they are saying. For example, one autistic informant in Robledo et al. (Reference Robledo, Donnellan and Strand-Conroy2012) explained, “I can hear a person better if I don't look at their face … So, when I'm making an effort to listen, I'm not making an effort to look, so sometimes when I'm listening to somebody I might look away from them, but I might turn my ear toward them” (p. 5). This is consistent with the autistic self-advocate Kedar's (Reference Kedar2012) explanation for why he does not commonly engage in eye contact: “I can listen better if I don't look at the person” (p. 49).

Another reason many autistic individuals avoid eye contact is because they find it uncomfortable. It “feels a bit creepy, so I tend to avoid it” (Higashida Reference Higashida2013, p. 25); “I can look but it's not pleasant” (Kedar Reference Kedar2012, p. 49); doing so feels “strange and uncomfortable” (Tammet Reference Tammet2006, p. 75); and “It is painful for me to look people in the eye” (Robledo et al. Reference Robledo, Donnellan and Strand-Conroy2012, p. 5) (for several additional first-person accounts, see McGlensey Reference McGlensey2016). Rather than indicating a lack of social interest, gaze aversion may be a strategy that some autistic people use to focus or to reduce or avoid stress.

There may be circumstances where learning to engage in eye contact more frequently can be beneficial (see sect. 4.2). But for now, the important point is that relatively low levels of eye contact do not necessarily signal a lack of social interest; in fact, gaze aversion is used adaptively by both autistic and non-autistic individuals to manage affective and cognitive resources.

2.2. Infrequent declarative pointing

Autistic children point less often than non-autistic children (e.g., Mundy et al. Reference Mundy, Sigman, Ungerer and Sherman1986), a behavioral difference that is sometimes interpreted as indicating that autistic toddlers lack the motivation to share experiences with others. In this section, we describe weaknesses with this social motivation interpretation of reduced pointing in autism and discuss alternative explanations.

Non-autistic children begin pointing with their index finger around 12 months of age (e.g., Bates et al. Reference Bates, Camaioni and Volterra1975; Camaioni Reference Camaioni1997; Carpenter et al. Reference Carpenter, Nagell and Tomasello1998; Liszkowski et al. Reference Liszkowski, Brown, Callaghan, Takada and de Vos2012). Researchers have traditionally suggested that infants are motivated to point for two primary reasons: to obtain things (“proto-imperatives”) and to share experiences with someone (“proto-declaratives”). For example, an infant may point to a desirable toy because she wants it; in this case, obtaining the toy is the goal and the adult is merely the means to that goal. Or an infant may point to a helicopter overhead to direct the adult's attention to that interesting event; in this case, sharing the experience with the adult is the goal and the helicopter is the means to that goal (Bates et al. Reference Bates, Camaioni and Volterra1975; Brinck Reference Brinck2004).

Declarative points are generally considered to be more socially, cognitively, and communicatively sophisticated than imperative points are. Someone who points to share an experience is assumed to be (1) motivated to inform the addressee of something interesting and (2) sensitive to whether the addressee has registered the intended message. For example, typically developing 12-month-old infants in one study pointed more at an interesting event (e.g., a puppet popping out of a display) if the experimenter expressed positive emotion toward the infant but did not actually look at what the infants were pointing to than if she expressed positive emotion and looked (Liszkowski et al. Reference Liszkowski, Carpenter, Henning, Striano and Tomasello2004). These kinds of findings suggest that infants who engage in declarative pointing do so because they want to affect their addressee's attentional state – specifically, they want them to notice something the infant finds interesting or noteworthy (e.g., Baron-Cohen Reference Baron-Cohen1989; Bates Reference Bates1976; Camaioni Reference Camaioni1997).

Imperative points, in contrast to declarative points, are viewed as purely instrumental (e.g., Baron-Cohen Reference Baron-Cohen1989; Camaioni Reference Camaioni1997). To produce imperative points, some understanding of others as causal agents – as individuals who can be directed to carry out desired actions – may be necessary. But the assumption is that imperative gestures can be produced for no purpose other than obtaining a desired object; affecting the mental state of the addressee is not the primary goal (but see Grosse et al. Reference Grosse, Behne, Carpenter and Tomasello2010; Shwe & Markman Reference Shwe and Markman1997).

It is commonly asserted that autistic children point to obtain things, but not to share experiences. For example, DeMarchena and Eigsti (Reference DeMarchena and Eigsti2014) write that “protodeclarative pointing is reduced in young children with ASD, while protoimperative pointing is not” (p. 375). Camaioni (Reference Camaioni1997) writes, “children with autism have severe difficulties in producing … declarative, but not imperative, pointing” (p. 222). And Whalen and Schreibman (Reference Whalen and Schreibman2003) have suggested that autistic children are “profoundly impaired” on declarative pointing, but not on imperative pointing (p. 456). If autistic children do point imperatively but not declaratively, it would be consistent with the assumption that they lack the motivation to share experiences with other people (Chevallier et al. Reference Jones, Quigney and Huws2012b; Liszkowski Reference Liszkowski2005). The logic is as follows: Because they point imperatively (and, perhaps, to direct their own attention; Goodhart & Baron-Cohen Reference Goodhart and Baron-Cohen1993), they are capable of producing the pointing gesture when they are motivated to do so; their failure to point declaratively must therefore be due to reduced motivation to share experiences with others.

But there are both empirical and theoretical problems with this line of reasoning. As noted earlier, there is no doubt that as a group, autistic children point less frequently than non-autistic children (e.g., Mundy et al. Reference Mundy, Sigman, Ungerer and Sherman1986). But the data on the dissociation between imperative and declarative pointing are not nearly as conclusive as they are sometimes made out to be. Specifically, an autistic child who does not point declaratively also tends not to point imperatively. For example, in one highly cited study (Baron-Cohen Reference Baron-Cohen1989, Experiment 3), 4 of the 10 autistic preschool-aged children observed during a 45-minute play session pointed imperatively but not declaratively, which is consistent with the notion that autistic children are uniquely impaired on declarative pointing. But the remaining 6 autistic children did not point at all, suggesting that pointing simply may not be a part of the behavioral repertoire for many or most young autistic children.

Similarly, in Baron-Cohen et al.’s (Reference Baron-Cohen, Cox, Baird, Swettenham, Nightingale, Morgan, Drew and Charman1996) report on the development of the Checklist for Autism in Toddlers (CHAT), 9 of the 10 parents of toddlers who were later diagnosed as autistic reported that their child did not engage in either type of pointing. Later work on the development of the Modified CHAT confirmed this finding (Robins et al. Reference Robins, Fein, Barton and Green2001): Parents of 82% of toddlers who later received a diagnosis of autism reported that their child did not point declaratively, and 72% reported their child did not point imperatively.Footnote ⁴ In short, autistic children as a group are not specifically impaired on declarative pointing, which undermines the claim that autistic children do not point declaratively because they are not interested in sharing experiences with other people.

An alternative explanation is that, as Gernsbacher and colleagues (Reference Gernsbacher, Stevenson, Khandakar and Goldsmith2008b) have argued, “it is the core act of pointing and its underlying motor demands rather than any deficit in intentionality or desire to share experience that underlies autistic children's lower frequency of [declarative] pointing” (p. 42). Indeed, autistic children and adults have well-documented and widely acknowledged difficulties with the planning, coordination, and execution of intentional movements in a variety of domains (e.g., Fournier et al. Reference Fournier, Hass, Naik, Lodha and Cauraugh2010; Glazebrook et al. Reference Glazebrook, Elliott and Szatmari2008; LeBarton & Iverson Reference LeBarton and Iverson2016; MacDonald et al. Reference MacDonald, Lord and Ulrich2014; Torres & Donnellan Reference Torres and Donnellan2015).

In firsthand accounts, too, autistic individuals describe occasionally or regularly experiencing a disconnect between a movement they would like to make and the movements they are actually able to make (e.g., Grandin Reference Grandin, Schopler and Mesibov1992; Robledo et al. Reference Robledo, Donnellan and Strand-Conroy2012; Williams Reference Williams1992). For example, an autistic 12-year-old explained his movement difficulties as follows: “I may wish to say something or do something. I can't get myself to move … It is confusing because I can initiate some things like eating or getting things in the house sometimes. I don't have an insight into this aspect of my illness. I just know it's there” (Kedar Reference Kedar2012, p. 63). At 13, the same child wrote of difficulties he had in getting his body to do what he wanted: “I am telling my hand to raise in class, or my feet to run, or my fingers to move on the piano, however they don't listen to me. Either they don't move, or they move badly. It's really rotten” (p. 101). Another autistic individual explained that, as a child, he “could not point at objects for many reasons. The most important reason is that I had very little sensation of my body” (Mukhopadhyay & Biklen Reference Mukhopadhyay, Biklen and Biklen2005, p. 133). Thus, some autistic children may have difficulty pointing because of motor and/or sensorimotor challenges, not because of a deficit in social motivation.

Furthermore, among autistic children who point, many do, in fact, point declaratively. For example, Mundy et al. (Reference Mundy, Sigman, Ungerer and Sherman1986) showed that although autistic preschoolers pointed less frequently overall in a standardized assessment than non-autistic children, the autistic children produced, on average, as many declarative as imperative points.Footnote ⁵ Additionally, in a study in which parents of autistic children were asked to rate the frequency of declarative pointing, more than 25% selected “a few times/week,” “a few times/day,” or “many times/day” (Allison et al. Reference Allison, Baron-Cohen, Wheelwright, Charman, Richler, Pasco and Brayne2008).Footnote ⁶ Given that many autistic children at least occasionally point declaratively, one cannot argue that they are not motivated to share experiences with others.

One could argue that because autistic children engage in declarative pointing less often than non-autistic children, they are less motivated to do so. But this neglects the previously mentioned motor challenges that may make pointing difficult for many autistic children. Furthermore, it is possible that autistic children are similarly motivated to share experiences, but that there are differences in the kinds of things they find interesting or worth sharing. For example, in one commonly used assessment of nonverbal communication (the Early Social Communication Scales; Mundy et al. Reference Mundy, Delgado, Block, Venezia, Hogan and Seibert2013), autistic toddlers are less likely than non-autistic toddlers to point declaratively at a moving wind-up toy or to pictures in a picture book. The behavioral difference is indisputable, but the social interpretation that is frequently offered – a lack of motivation to share experiences (e.g., Mundy Reference Mundy2016) – is disputable: It is possible that some autistic children simply do not find these toys or events as interesting as most non-autistic children do.

Some empirical support for this possibility comes from a study in which autistic children were less likely than non-autistic children both to engage in declarative communication about toys and to explore the toys themselves (O'Neill & Happé Reference O'Neill and Happé2000). This is consistent with reports from some autistic individuals, who describe focusing on and finding enjoyment in different things than non-autistic people. For example, one autistic adult wrote that “Lots of times I'm surprised by what other people said they saw and heard, because it is not what I saw and heard … I don't know why my head picks things to focus on, but I know it is usually not the same things other people pick to focus on” (Jones et al. Reference Jones, Quigney and Huws2003, p. 119; see also Rubin Reference Rubin and Biklen2005). As a 12-year-old, Kedar (Reference Kedar2012) explained, “I'm not entertained by the ordinary things that most people enjoy” (p. 45).

The production of declarative gestures, including pointing, has been linked to a number of positive language outcomes (for reviews, see Colonnesi et al. Reference Colonnesi, Stams, Koster and Noon2010; Goldin-Meadow Reference Goldin-Meadow2009). When a child points to an object, many parents in Western, middle-class families use it as an opportunity to offer the object's name (e.g., Goldin-Meadow et al. Reference Goldin-Meadow, Goodrich, Sauer and Iverson2007). Clearly, there can be benefits if a child and adult attend to the same referent when it is labeled (Tomasello & Farrar Reference Tomasello and Farrar1986). But as with eye contact, pointing is just one of several ways joint attention – a shared focus on the same object or event – can be achieved (Akhtar & Gernsbacher Reference Akhtar and Gernsbacher2007; Gernsbacher et al. Reference Gernsbacher, Stevenson, Khandakar and Goldsmith2008b), an issue we consider in more detail in section 5.

Although it is possible that the reduced frequency of declarative pointing in autistic children is somehow caused by a reduced motivation to share experiences, alternative explanations are possible, including differences in motor ability and in what they find worth sharing.

2.3. Elevated levels of motor stereotypies

Most autistic people engage in motor stereotypies: rhythmic, repetitive movements that (from a naïve observer's perspective) appear purposeless (Bodfish et al. Reference Bodfish, Symons, Parker and Lewis2000; Goldman et al. Reference Goldman, Wang, Salgadd, Greene, Kim and Rapin2009; Seltzer et al. Reference Seltzer, Krauss, Shattuck, Orsmond, Swe and Lord2003). For example, they might flap their hands or arms, rock their bodies, wiggle their fingers in front of their eyes, or spin objects.Footnote ⁷ Many non-autistic people also engage in motor stereotypies when anxious or bored, twirling their hair or a pencil, biting their nails, drumming their fingers, or tapping their feet. Motor stereotypies are also common early in typical development, where it is thought that they may help infants transition from uncoordinated motor activity to voluntary motor control (Thelen Reference Thelen1981). But whereas they are infrequent beyond the preschool years in most non-autistic children, they tend to persist or become more frequent among autistic children (Harrop et al. Reference Harrop, McConachie, Emsley, Leadbitter and Green2014; MacDonald et al. Reference MacDonald, Green, Mansfield, Geckeler, Gardenier, Anderson, Holcomb and Sanchez2007). In MacDonald et al. (Reference MacDonald, Green, Mansfield, Geckeler, Gardenier, Anderson, Holcomb and Sanchez2007), for example, 2- and 4-year-old non-autistic children were observed engaging in motor stereotypies, on average, for 4.8% and 2.1%, respectively, of a 10-minute sample; in autistic children, the average percentages were 6.9% and 20.2%.

When as much as one-fifth of an autistic child's time is spent engaging in behaviors that appear meaningless to many observers, questions arise about whether that time would be better spent doing other things. Repetitive behaviors in autism have been called “debilitating” (South et al. Reference South, Ozonoff and McMahon2005, p. 155), “disrupting” (Goldman et al. Reference Goldman, Wang, Salgadd, Greene, Kim and Rapin2009, p. 36), a “prominent impairment to the daily life of affected individuals” (Langen et al. Reference Langen, Durston, Kas, van Engeland and Staal2011, p. 356), and a “major barrier to learning” (Leekam et al. Reference Leekam, Prior and Uljarević2011, p. 562). Additionally, because motor stereotypies look odd (Smith & Van Houten Reference Smith and Van Houten1996), they can be a source of social stigma (Cunningham & Schreibman Reference Cunningham and Schreibman2008).Footnote ⁸ For these reasons, reducing or eliminating them is an explicit goal for many clinicians, teachers, and parents (e.g., Lanovaz et al. Reference Lanovaz, Robertson, Soerono and Watkins2013; Rapp & Lanovaz Reference Rapp and Lanovaz2014; Rapp & Vollmer Reference Rapp and Vollmer2005).

The cause of motor stereotypies in autism is unknown, but one proposal suggests that it may be related to a lack of social motivation in infancy. In a recent review, Leekam et al. (Reference Leekam, Prior and Uljarević2011) noted that non-human animals raised in restricted or deprived environments can develop stereotyped motor movements, such as pacing, body rocking, and compulsive grooming (e.g., Lewis et al. Reference Lewis, Tanimura, Lee and Bodfish2007). They suggested that autistic infants may also suffer from a restricted environment, specifically, a “self-imposed” restricted environment caused by their “extreme social withdrawal” (p. 577). According to Leekam et al., this restricted environment could have a number of cascading effects, including the development of motor stereotypies (see also Schultz Reference Schultz2005).

But the link between perceived social withdrawal and motor stereotypies is tenuous at best. Among autistic toddlers and young children, the level of impairment in the social domain (as measured by observation and parent report of conventionally expected social behaviors) is not correlated with the frequency of repetitive and restricted behaviors (RRBs), the umbrella category of behaviors of which motor stereotypies are a part (Harrop et al. Reference Harrop, McConachie, Emsley, Leadbitter and Green2014). A large population-based study investigating autistic-like traits reported very low correlations (0.1–0.3) between social impairments and RRBs in 7- to 9-year-old children, even when considering only those with extreme scores (Happé & Ronald Reference Happé and Ronald2008).

Furthermore, when Leekam et al. (Reference Leekam, Prior and Uljarević2011) suggest that the ultimate cause of motor stereotypies could be autistic infants’ “extreme social withdrawal” (p. 577), the implication is that these infants are not motivated to interact with people. Leekam et al. do not describe the particular behaviors they believe constitute evidence for extreme social withdrawal in infancy, but we imagine they might include things that are also routinely interpreted as indicating diminished social motivation (e.g., failing to engage in sustained eye contact, not responding when called by name, not actively seeking out adult interaction). As we have been arguing, however, explanations unrelated to social interest are possible for these kinds of behaviors. An infant may be perceived as lacking in social interest, but this does not necessarily mean that the infant actually is lacking in social interest (see sect. 5).

Lack of social motivation has also been implicit in some explanations for why motor stereotypies persist. Many scientists and clinicians believe that motor stereotypies can be brought under voluntary control (e.g., Rapp & Vollmer Reference Rapp and Vollmer2005), and individuals who do not learn to suppress them are sometimes thought to lack the motivation to do so (Miller et al. Reference Miller, Singer, Bridges and Waranch2006). The underlying assumption seems to be that individuals who want to connect with other people would not engage in apparently meaningless behaviors that can cause others to ostracize them. Although some autistic individuals have described using stereotypies because they want to be left alone (Joyce et al. Reference Joyce, Honey, Leekam, Barrett and Rodgers2017), others have reported that they simply cannot control these movements even though they would like to (e.g., Fleischmann & Fleischmann Reference Fleischmann and Fleischmann2012; Robledo et al. Reference Robledo, Donnellan and Strand-Conroy2012; Tammet Reference Tammet2006). As one autistic informant explained, “I want to stop doing anything that doesn't look normal” (Robledo et al. Reference Robledo, Donnellan and Strand-Conroy2012, p. 6).

For some autistic people, motor stereotypies can serve important self-regulatory and even communicative functions. For example, one informant explained that “one of my most interesting and prevalent repetitive behaviors (stims, ‘stereotypies’) is rubbing objects (e.g., door knobs) because of the unpleasant sensations they leave on my hands – I keep trying to ‘rub’ the touch off” (Jones et al. Reference Jones, Quigney and Huws2003, p. 118). A study of autistic young adults found that many described engaging in stereotypies as a coping response to anxiety or uncertainty (Joyce et al. Reference Joyce, Honey, Leekam, Barrett and Rodgers2017), which is consistent with data showing a link between parent reports of their child's anxiety and parent and clinician measures of stereotyped behaviors (Sukhodolsky et al. Reference Sukhodolsky, Scahill, Gadow, Arnold, Aman, McDougle, McCracken, Tierney, White, Lecavalier and Vitiello2008). In terms of communication, Julia Bascom (Reference Bascom2011), executive director of the Autistic Self Advocacy Network, has explained that she expresses emotion through hand flapping: Friends “can ‘read’ my flapping better than my face … I wish everyone could look at my hands and see I need you to slow down or this is the best thing ever or can I please touch or I am so hungry I think my brain is trying to eat itself” (italics in original).

Non-autistic individuals can engage in motor stereotypies that look very similar to those produced by autistic individuals, but when they do, their behaviors are not assumed to reflect deficits in social motivation. For example, Harris et al. (Reference Harris, Mahone and Singer2008) studied 100 non-autistic children referred to a pediatric neurology movement-disorders clinic specializing in tic disorders. These children (on average, 8 years of age) were otherwise typically developing but engaged in many of the same kinds of stereotypies as autistic children (e.g., hand flapping, rocking), often multiple times a day, and sometimes for several minutes each time. Social motivation was not mentioned as a possible cause. The way in which the very same unusual repetitive motor movements are characterized “appears to depend more on the underlying diagnosis of the patients than the movements themselves” (Edwards et al. Reference Edwards, Lang and Bhatia2012, p. 181; see also Leary & Hill Reference Leary and Hill1996).

In one study offering a particularly compelling demonstration of how laypeople interpret motor stereotypies differently depending on diagnosis (Sperry & Symons Reference Sperry and Symons2003), mothers of young autistic children watched several 10-second home-movie clips of infants between the ages of 9 and 18 months and were asked to rate how intentional the infants’ behaviors were (i.e., “On a scale of 1–6, how much do the children mean to do what they are doing”; Feldman & Reznick Reference Feldman and Reznick1996). Half of the mothers were told that all of the children in the clips had a diagnosis of autism, and half were given no diagnostic information. Unbeknownst to the mothers, in fact, all of these infants had later received a diagnosis of autism. For clips showing a child engaged in stereotyped motor movements (e.g., arm flapping, leg kicking, body rocking), there was a striking effect of condition: Mothers who had been told that the infants were autistic rated the behavior shown in those clips as less intentional than mothers who had not been given any diagnostic information.

According to autistic individuals, some motor stereotypies are involuntary behaviors and some are intentional, and the same individual may engage in stereotypies for a number of different reasons. But engaging in them does not necessarily have anything to do with one's interest in connecting with other people.

2.4. Frequent echolalia

Echolalia involves the verbatim repetition of part or all of another's utterance and can include words and phrases that do not appear to be relevant in the current context. For example, on meeting a therapist for the first time, one young autistic child repeatedly exclaimed, “Got a splinter!” (Prizant Reference Prizant2015). It has been estimated that 75% of autistic children engage in echolalia (Rutter et al. Reference Rutter, Greenfield and Lockyer1967), and in the DSM-5, it is listed alongside motor stereotypies as an example of a restricted, repetitive pattern of behavior (American Psychiatric Association 2013). To a naïve interlocutor, these violations of conversational conventions can make it appear as though the individual's utterances are not communicative and that the individual is therefore not interested in social engagement or communication.

Because echolalic speech is so common in autism and can appear to be meaningless, it (like motor stereotypies) has been a target for behavioral interventions designed to reduce or eliminate it (for a review, see Neely et al. Reference Neely, Gerow, Rispoli, Lang and Pullen2016). Indeed, some have argued that echolalia is not just meaningless but actually problematic: Some have argued, for example, that repetitions of (apparently) irrelevant words and phrases interfere with language development (Valentino et al. Reference Valentino, Shillingsburg, Conine and Powell2012) and contribute to communicative breakdowns (Neely et al. Reference Neely, Gerow, Rispoli, Lang and Pullen2016). In this section, we argue that echolalia should not be dismissed as meaningless merely because a listener is unable to immediately decipher its meaning. We describe how autistic and non-autistic children and adults use it communicatively – to connect with other people – as well as for self-regulatory purposes.

If a young non-autistic child is asked, “Is your sister bothering you?” and responds, “Bothering you,” this would likely be interpreted as an affirmative reply (“Yes, she is bothering me”). These “frozen phrases” are structurally and perhaps functionally equivalent to echolalic speech and are seen as playing an important role in language development. Typically developing children use unanalyzed chunks of spoken language to convey meaning (Bretherton et al. Reference Bretherton, McNew, Snyder and Bates1983). In the case of “bothering you,” a young child may not yet be able to separate that phrase into its constituents or know how to slot in a different pronoun, but the child is using that phrase to communicate a meaningful proposition (Bloom Reference Bloom1973; Nelson Reference Nelson1981). Indeed, the use of frozen phrases in typical development is seen as a step on the road to productive speech.

Just as frozen phrases give way to productive multiword utterances in typical development (Lieven et al. Reference Lieven, Pine and Barnes1992), echolalia can serve as a stepping stone to productive use of grammar for speaking autistic individuals (Blanc Reference Blanc2012; Gernsbacher et al. Reference Gernsbacher, Morson and Grace2016; Manning & Katz Reference Manning and Katz1989; Roberts Reference Roberts, Arciuli and Brock2014). A sentence or phrase that starts as fully echolalic can become modified over time as some elements are replaced by others. For example, one autistic child regularly repeated, “One day in Teletubbie land, all of the Teletubbies were very busy when suddenly a big rain cloud appeared.” Later, this boy (named Bud) described his father's returning home by saying, “One day in Bud's house, Mama and Bud were very busy when suddenly Daddy appeared” (Dawson et al. Reference Dawson2008, p. 766).

Non-autistic adults also repeat utterances verbatim for various communicative functions. For example, they repeat what someone else has said to express agreement (e.g., Speaker 1: “Let's go.” Speaker 2: “Let's go.”). They use repetition to express incredulity (e.g., after a child asks for ice cream for breakfast, a parent might respond, “Ice cream for breakfast?!”) and quote lines from television shows or movies to comment on the similarity between the current situation and the one depicted on film (e.g., “No soup for you!” from Seinfeld). In all these cases, the speaker's use of repetition is intended to be communicative, and a listener who shares the relevant common ground would interpret it as such.

Why is something that occurs regularly in typical development and in non-autistic adult speech often seen as aberrant and meaningless in autism? There are at least three reasons. First, autistic people use echolalia more often than non-autistic people do (van Santen et al. Reference van Santen, Sproat and Hill2013).Footnote ⁹ Second, autistic individuals who engage in echolalia likely also engage in other behaviors that are frequently perceived as meaningless (like motor stereotypies; sect. 2.3). Finally, it may be more difficult to decipher the meaning behind a given instance of autistic echolalia.

This last point about meaning may be the most problematic. The tension between when and whether echolalia should be considered meaningful can be seen even in Kanner's (Reference Kanner1943) original description of autism. Some of the autistic children he described engaged in “parrot-like repetitions” (p. 228) that could not always “be linked up with immediate situations” (p. 227). He considered some of these utterances to be meaningful, as when he inferred that one child “expressed agreement by repeating the question literally, echolalia-like” (p. 220; also p. 243) or that the same child was asking his mother to pull off his shoe even though he said “pull off your shoe” (p. 219). But Kanner considered other examples of echolalia (e.g., “You'll fall off the bicycle and bump your head”; p. 227) to be meaningless, writing, for example, “None of [the child's] remarks was meant to have communicative value” (p. 227).

But just because a listener is unable to decipher the meaning of an utterance in a particular context does not mean that the speaker did not have a meaning in mind (e.g., Stiegler Reference Stiegler2014). Sometimes, the intended meaning can be understood only by someone who knows the speaker well and is motivated to take the time to carefully study the contexts of use. For example, the child introduced at the beginning of this section, who repeated “Got a splinter!” when meeting a new therapist, had sometime in the past experienced a painful splinter and used that phrase to communicate her anxiety (Prizant Reference Prizant2015). One autistic boy echoed the phrase “UPS [United Parcel Service] is here” as a clever means of getting his father's attention (Light et al. Reference Light, Roberts, DiMarco and Greiner1998, p. 166). Another repeated, “Chicken Little thought the sky was falling, but the sky is not falling” when his mother was despondent over the death of a friend (Gralow Reference Gralow2008). Listeners who did not know these individuals well or know what was going on in their lives would probably mistakenly consider these phrases meaningless.Footnote ¹⁰

Anecdotal evidence suggests that when interlocutors impute meaning to echolalic speech they had previously considered meaningless, the effects can be profound. For example, Suskind (Reference Suskind2014) describes how his young autistic son, Owen, would repeatedly say “juicervose,” an apparently meaningless sequence of sounds. One day, however, Suskind and his wife recognized that “juicervose” was Owen's approximation of the phrase “just your voice,” a line that figures prominently in The Little Mermaid, one of his favorite Disney movies. In that movie, a witch offers a mermaid human legs in exchange for her voice, explaining, “It won't cost much: just your voice!” Putting these pieces together, the Suskinds concluded that Owen's use of “juicervose” was an attempt to draw attention to his own lost ability to speak.

Whether they were correct that Owen intended “juicervose” to have this or any meaning is almost beside the point. If they were correct, it was a clear instance of Owen successfully communicating with his family. If they were not correct, it was still a breakthrough: From that moment forward, Suskind (Reference Suskind2014) writes, they began actively encouraging Owen to use dialogue from Disney movies to express his thoughts and feelings (and they also communicated with him in the same way). As an adult, Owen now uses non-echolalic speech to communicate, but he continues to recite dialogue from movies to convey particular meanings and to make sense of particular situations, just as non-autistic people do. It is impossible to know for sure, but it seems likely that the Suskind family's willingness and ability to see meaning in Owen's echolalia as a child played an important role in his subsequent linguistic, cognitive, and social development.

Indeed, in the literature on typical development, the meaning parents ascribe to their child's behavior influences how they treat the child, which is thought to have important downstream consequences (e.g., Reddy & Trevarthen Reference Reddy and Trevarthen2004). Parents of typically developing infants often react to early vocalizations as if they were intended to communicate something, responding in ways that are thought to promote communicative development (Snow & Ferguson Reference Snow and Ferguson1977). For example, Gros-Louis et al. (Reference Gros-Louis, West and King2014) found that mothers in a Western culture frequently responded to the “meaningless” vocalizations made by their 8- to 14-month-old infants by expanding upon them: If the infant babbled “da-da-da,” the mother might respond with “Da-da is working. I am mama” (p. 392). Gros-Louis et al. found that these kinds of maternal responses predicted an increase in infants’ vocal production. Similarly, treating echolalia as intentional communication is likely to signal to an individual that conversational partners are interpreting their attempts to communicate as meaningful, which may reinforce the individual's desire to communicate.

Some instances of echolalia may not be intended to communicate anything to other people, but this does not necessarily mean that they are meaningless. Like motor stereotypies, some autistic people use echolalia as an adaptive, self-regulatory strategy, repeating a phrase to assure themselves that things will be okay, for example (Prizant Reference Prizant2015). This is not so different from the mantras that some non-autistic people repeat to calm themselves in stressful situations (see, e.g., Eddie Murphy's repetition of “Keep it together” in the movie Bowfinger). Some autistic individuals have also reported using echolalia as a means of keeping material in short-term memory (Higashida Reference Higashida2013), the same strategy non-autistic individuals use when trying, for example, to remember an address or phone number long enough to write it down.

Some autistic people have reported that they occasionally repeat words and phrases involuntarily (Rentenbach & Prislovsky Reference Rentenbach and Prislovsky2012; Robledo et al. Reference Robledo, Donnellan and Strand-Conroy2012). But echolalia is also clearly used by autistic (and non-autistic) people communicatively, as a creative means of connecting with other people.

2.5. Summary and key points

We have described four behavioral differences commonly associated with autism: Compared with non-autistic individuals, autistic individuals are less likely to engage in eye contact or point, and they are more likely to engage in motor stereotypies and echolalia. We have explained how each of these behavioral differences has been interpreted by some scientists as reflecting diminished social interest or motivation. Using arguments from logic, existing quantitative data, and the testimony of autistic people themselves, we have described alternative explanations for each behavioral difference. Here we synthesize two key points from the previous sections.

The first is that most of the unusual behaviors documented in autism have also been documented among non-autistic children and adults (Bishop Reference Bishop1989). When non-autistic people engage in these behaviors, they are not attributed to deficits in social motivation; to the contrary, they are often considered to be adaptive responses to a particular situation. Take reduced levels of eye contact. As noted, many autistic people do not consistently engage in eye contact. But many non-autistic people also refrain from eye contact when they are trying to concentrate or control their emotions. Similarly, many autistic people repeat words or phrases. But so do typically developing children when they are learning to communicate, and so do non-autistic adults when they are trying to emphasize a point or self-regulate. Finally, most autistic individuals engage in motor stereotypies, rhythmically moving parts of their bodies or engaging with objects in unusual ways. But stereotypies are also not unique to autism: They are ubiquitous in typical development, where they are considered essential to motor development, and common among non-autistic adults, who use them to combat anxiety and boredom.

The second key point is that many autistic people have explained that they do not intend their atypical behavior to reflect anything about social interest. For example, they report that averting gaze allows them to concentrate and regulate their emotions, just as it does for non-autistic people. They report that motor stereotypies and echolalia both serve self-regulatory and communicative functions, just as they do in non-autistic people. They describe experiencing an occasional or regular disconnect between movements they would like to produce and those they can actually produce in the moment: a disconnect that could affect their ability to engage in pointing and other behaviors that are conventionally interpreted as indicating social interest.

4.1. Research

Consider, as an example, a recent study by Moriuchi et al. (Reference Moriuchi, Klin and Jones2017) designed to investigate whether autistic toddlers do not engage in eye contact as often as non-autistic toddlers because they find it aversive (see sect. 2.1) or because they do not find eyes engaging or informative. Recall that eye contact is one behavior conventionally assumed to reflect social motivation because it represents an obvious example of orienting toward another person (Chevallier et al. Reference Jones, Quigney and Huws2012b). Moriuchi et al. hypothesized that if autistic toddlers found eye contact aversive, they would look away more quickly if they happened to look at a videotaped actress's eyes than if they happened to look at her mouth or other stimuli in the scene.Footnote ¹¹ The results did not support the gaze aversion hypothesis as it was operationalized in that study: Just like non-autistic toddlers, autistic toddlers looked away from eyes as quickly (or slowly) as they looked away from mouths or non-face stimuli.

The authors also hypothesized that whereas non-autistic toddlers would seek out the videotaped actress's eyes when she was behaving in an especially socially engaging manner, autistic toddlers who were averse to gaze would avoid the actress's eyes at those times. Results again did not support the gaze aversion hypothesis: Autistic toddlers spent less time overall than non-autistic toddlers looking at eyes, but they tended to seek them out at the same times as the non-autistic toddlers. The authors concluded that their results “contradict the hypothesis that children with ASD actively avoid looking at the eyes early in life” and instead suggest that autistic children have a “passive insensitivity to social signals in others’ eyes” (Moriuchi et al. Reference Moriuchi, Klin and Jones2017, p. 33).

We find the second part of this conclusion puzzling. That autistic toddlers in this study tended to seek out the actress's eyes at the same times as non-autistic toddlers (though less frequently) suggests a sensitivity to social signals, not an insensitivity. It suggests that, like non-autistic toddlers, autistic toddlers were motivated to look at the actress when she was most likely to be providing interesting or important social information. As the authors note, this is not consistent with the prediction one would make if they had an aversion to looking at others’ eyes. But it is also not consistent with the prediction one would make if they were passively insensitive to social signals in others’ eyes.

Thus, one negative consequence of assuming that autistic individuals are socially uninterested is that it can lead researchers to interpret autistic participants’ behavior as indicating that they are socially uninterested, even though that interpretation is not made about non-autistic participants who behave in the same ways. Additionally, researchers tend to emphasize differences between autistic and non-autistic behavior even though, as in the Moriuchi et al. (Reference Moriuchi, Klin and Jones2017) study, the similarities may be at least as interesting and important (see Jaswal et al. Reference Jaswal, Akhtar and Burack2016).

4.2. Interventions

At some level, a focus on behavioral differences is understandable; this is what allows diagnoses to be made and support provided. But when differences between two groups are obtained, members of the marginalized group are generally assumed to lack something desirable (Medin et al. Reference Medin, Bennis and Chandler2010). Even differences that could be benign or adaptive tend to be interpreted as “deficits” in need of remediation (Akhtar & Jaswal Reference Akhtar and Jaswal2013). In autism, the practical effect is that many interventions focus on making autistic people appear more like non-autistic people with little consideration of the potential negative consequences of these efforts. These efforts may sometimes have the unintended and ironic effect of undermining their interest in interacting with other people.

For example, research targeting the reduction or elimination of echolalia has recently been described as “burgeoning” (Stiegler Reference Stiegler2014, p. 750) despite the fact that, as described in section 2.4, echolalia can serve important communicative functions. The potentially adaptive functions of echolalia do not seem to be considered in these elimination efforts: An exhaustive review of published studies on echolalia interventions noted that “none of the studies reported the function of the target echolalia behavior” (Neely et al. Reference Neely, Gerow, Rispoli, Lang and Pullen2016, p. 90). Ignoring the function of echolalia is not just a questionable practice, it is a dangerous one: If someone's unconventional attempts to communicate are ignored or, worse, discouraged, their motivation to communicate is likely to be reduced (Akhtar et al. Reference Akhtar, Jaswal, Dinishak and Stephan2016; Prizant Reference Prizant2015; Sterponi et al. Reference Sterponi, de Kirby and Shankey2015).

Motor stereotypies are also a common target for intervention, presumably because they are ubiquitous, apparently purposeless, and thought to interfere with social and academic development. But reducing motor stereotypies has not been particularly effective at increasing desired behaviors (Lanovaz et al. Reference Lanovaz, Robertson, Soerono and Watkins2013) and may result in new undesirable behaviors (Epstein et al. Reference Epstein, Taubman and Lovaas1985) that parents actually report as more problematic than the original motor stereotypies (South et al. Reference South, Ozonoff and McMahon2005). We hope that some of the techniques that were employed to eliminate stereotypies in the past, including shouting at and shaking autistic children who engaged in rocking behavior (Risley Reference Risley1968), are no longer used. But parents are still routinely advised to try to eliminate these harmless movements. For example, in a recent op-ed piece in The New York Times, one mother of a young autistic child explained that she was told by therapists, “Try to cut down on his arm flapping” and “Don't let him spin objects.” She explained, “I drove myself to tears trying. At nap time I resorted to climbing into his crib to snuggle and sing because it was the only way I could get him to stop waving his hands in front of his eyes” (O'Brien Reference O'Brien2017).

Autistic individuals have described finding efforts to prevent them from engaging in harmless and adaptive motor stereotypies as both frustrating and aversive (e.g., Bascom Reference Bascom2011). It is not difficult to see how autistic people's desire to interact with someone might diminish if that person prevents them from engaging in harmless “stims” that may be soothing, enjoyable, or something over which they have no control.

Autistic children are also regularly instructed to look other people in the eye. There may be circumstances in which learning to make eye contact more frequently can be beneficial. For example, Krstovska-Guerrero and Jones (Reference Krstovska-Guerrero and Jones2016) trained autistic toddlers to shift their gaze from an object to an adult's eyes and found an increase in other conventional communicative behaviors such as initiating requests and smiling. This, in turn, may have led caregivers of these children to behave in ways that facilitated further social and communicative development. That said, insisting on eye contact may backfire. For example, Moriuchi et al. (Reference Moriuchi, Klin and Jones2017, supplemental materials, p. 20) suggest that autistic children may be conditioned to find eye contact aversive when adults’ exaggerated bids for eye contact become associated with “non-preferred activities.” Thus, attempting to increase a behavior that is conventionally interpreted as indicating social interest could paradoxically actually undermine that interest.

5.1. Who is responsible for altered social interactions?

Recall from section 2.3 that, in the context of motor stereotypies, Leekam et al. (Reference Leekam, Prior and Uljarević2011) proposed that autistic infants experience a “self-imposed constrained environment” caused by their “extreme social withdrawal” (p. 577). As noted in that section, we cannot know whether an infant who is perceived to be socially withdrawn is actually socially withdrawn. It is true, however, that an infant who is interested in people but who – because of a neurological condition affecting her perceptual, attentional, and sensorimotor systems – does not frequently smile, coo, or engage in sustained eye contact with her caregivers will not get the same kinds of social opportunities in Western cultures as one who routinely does engage in these behaviors. Crucially, however, this would not represent a “self-imposed” restricted environment; it would represent an environment that became restricted over time as a result of the way the infant was treated by adults who expected her to show her social interest in conventional ways. All social environments arise from an interaction between how a person behaves and others respond. It is simply inaccurate to describe any child's environment as “self-imposed.”

This is a crucial point because it has important implications for intervention efforts. If an infant's restricted environment is thought to be “self-imposed” because she is socially withdrawn, then interventions are likely to focus on making her appear less socially withdrawn. For example, Leekam et al. (Reference Leekam, Prior and Uljarević2011) advocate for “active and intensive intervention [on autistic infants] that acts upon that self-imposed constrained environment to enhance brain development and reduce stereotypies” (p. 577). Similarly, Dawson (Reference Dawson2008) recommends early behavioral intervention to correct the “failure on the part of the [autistic] child to actively engage in early social interaction” (p. 776).

If, however, we take seriously the possibility that an autistic infant's restricted environment comes about, at least in part, because caregivers (mis)interpret her unusual behaviors as indicating social withdrawal and so treat her as if she were socially withdrawn, then an important target for intervention should be caregivers’ perceptions of the infant's behavior. Broadening the range of behaviors seen as indicating social interest might provide alternative ways for autistic infants to gain the experience-dependent stimulation needed to support healthy social, communicative, and cognitive development (e.g., Akhtar et al. Reference Akhtar, Jaswal, Dinishak and Stephan2016; Prizant Reference Prizant2015).

An instructive analogy can be seen in the education of deaf individuals in the United States. When forced to lip-read and speak, deaf individuals have difficulty with both language development and academic achievement (Marschark Reference Marschark2006). However, when exposed to sign language – particularly early in life and from fluent signers – deaf children's developmental trajectories in those domains more closely match those of their hearing peers (for a review, see Lederberg et al. Reference Lederberg, Schick and Spencer2013). Sometimes the most effective intervention is one that broadens the range of behaviors deemed acceptable by the majority group, accommodating individuals’ unique needs and strengths rather than insisting that they behave in the conventional way (Mottron Reference Mottron2017).

5.2. A path forward

In disabilities other than autism, it is widely recognized that there are unconventional ways in which social interest can be expressed. For example, blind infants’ behavior could lead observers to assume they are not socially motivated: They do not orient toward others by engaging in eye contact; they sometimes turn away when others are talking; and they rarely use gestures like pointing, offering, or showing (Pérez-Pereira & Conti-Ramsden Reference Pérez-Pereira and Conti-Ramsden1999). But parents do not assume, on the basis of these behaviors, that their infants are uninterested in social interaction. Parents are not instructed to insist that their children engage in eye contact, pointing, or showing. Instead, parents learn “to be more patient and careful in detecting responses and signs of engagement,” to establish joint attention via touch, and to recognize “idiosyncratic movements/gestures” as bids for experience sharing (Pérez-Pereira & Conti-Ramsden Reference Pérez-Pereira and Conti-Ramsden1999, pp. 45-47; see also Bigelow Reference Bigelow2003; Fraiberg Reference Fraiberg1977).

Among typically developing children, too, there is evidence that social engagement can be shown in alternative ways. For example, although gaze following is the most commonly studied route to achieving a shared focus on an object, Yu and Smith (Reference Yu and Smith2013) found that it was not necessary. In a study of 12-month-old infants and their parents interacting with several toys, infants rarely looked at their parent's face. But because parents tended to hold and look at a toy when attempting to draw their child's attention to it, infants could (and did) successfully coordinate visual attention with their parent simply by looking at their parent's hands. In fact, there was so much redundancy between the hands and the eyes that the authors concluded that “Current approaches [to joint attention] that concentrate on looks to faces and eyes, and to teaching those looks as parts of intervention programs for individuals with various developmental delays may be making the task harder” (p. 6).

In the case of autism, Ochs and Solomon (Reference Ochs and Solomon2010) pointed out that autistic people “possess a characteristic range of possibilities for social coordination that is shaped not only by their disorder, but also by the sociocultural practices of the communities they inhabit and the interlocutors with whom they interact” (pp. 73–74). In ethnographic work observing autistic children in a variety of settings, they found that what they called “autistic sociality” was often limited by interlocutors relying on standard ways of interacting, including face-to-face body alignment and insistence on speech (see also Ochs et al. Reference Ochs, Solomon and Sterponi2005). Based on our own preliminary work, informed by autistic people and their families, we hypothesize that sociality is also limited when interlocutors fail to recognize and respond to unconventional bids for social connection, including proximity, touch (see also Escalona et al. Reference Escalona, Field, Nadel and Lundy2002), coordinated movement, repetition (see also DeThorne et al. Reference DeThorne, Hengst, Valentino and Russell2015), and rituals.

Investigating unconventional ways autistic people may show their social interest will require a multipronged approach. One important step should involve further theoretical work on the psychological construct of social motivation itself. As noted earlier, Chevallier and colleagues (Reference Jones, Quigney and Huws2012b) operationalized social motivation in terms of an individual's tendency to (a) orient toward, (b) seek out and enjoy, and (c) maintain relations with other people. This definition bears a striking resemblance to some descriptions of extraversion (e.g., Costa & McCrae Reference Costa and McCrae1980): “Extraverts have a preference for seeking, engaging in, and enjoying social interactions” (Fishman et al. Reference Fishman, Ng and Bellugi2011, p. 67). Introverts, by contrast, are described as “withdrawn, retiring, reserved, inhibited, quiet” (McAdams Reference McAdams and Kazdin2000, p. 305) – adjectives that might also be used to describe many autistic individuals. In fact, autistic people (and non-autistic people with more autistic-like traits) do score higher on measures of introversion and lower on measures of extraversion than non-autistics (Ozonoff et al. Reference Ozonoff, Garcia, Clark and Lainhart2005; Schwartzman et al. Reference Schwartzman, Wood and Kapp2016; Wakabayashi et al. Reference Wakabayashi, Baron-Cohen and Wheelwright2006). Thus, an important question is how what social motivation theorists call “diminished social motivation” relates to what personality psychologists call “introversion.”

This question is important because there is evidence that although introverts may not show their social interest in ways that have traditionally been valued or associated with social interest in Western cultures, many do seek out and enjoy social interactions; those interactions are just different from the ones pursued by extraverts (e.g., smaller groups, less small talk; see Cain Reference Cain2012). We see an interesting parallel to autism: Like introverts, perhaps autistic people show their social interest in unconventional (i.e., non-extraverted) ways. To be clear, we are not arguing that autism can be explained as an extreme form of introversion (see Grimes Reference Grimes2010); much more work would be required to make that case. But clarity and insight could be gained from a thorough examination of how social motivation is related to other psychological constructs.

Research is also urgently needed to identify and characterize the range of behaviors that can signal social interest. For example, interviews with autistic individuals and their families can identify candidate behaviors, and observations can quantify and investigate how they change over time. Experience sampling methods and experimental manipulations can shed light on both the idiosyncratic and common conditions under which those kinds of behaviors occur and what effects they have on different social partners. Cultural comparisons can investigate ways in which observers’ judgments of social interest may be mediated by cultural expectations, which could inform debates about the range of behaviors that can demonstrate social interest. Intervention studies can investigate whether caregivers can be trained to see behaviors already in an autistic child's repertoire as bids for social connection – albeit unconventional ones – and what effects this training might have on the child, the caregiver, and their relationship.

Proponents of the social motivation perspective are likely to agree that social interactions are reciprocal (e.g., Dawson Reference Dawson2008), that is, that each member of a dyad influences the other. But theoretical and empirical work in this area has focused primarily on the problems that arise when autistic individuals fail to show their social interest in conventional ways; little consideration has been given to how (mis)interpretations of autistic behavior may contribute to those problems (Dinishak & Akhtar Reference Dinishak and Akhtar2013; Gernsbacher Reference Gernsbacher2006). Investigating and capitalizing on unconventional ways that social interest may be expressed and responded to represents a domain with rich potential and should be a high priority for autism research and intervention.