Reciprocity in autism research, and why it matters

Imagine a game of catch between a toddler and a professional baseball player.  In a reciprocal game, the ball will pass steadily back and forth between the players.  That doesn't necessarily mean that both players are performing the same movements or doing the same amount of work.  The toddler's throw may be so weak that the baseball player must run forward to catch it, while the latter must grade his throw so that it doesn't go over the toddler's head.  The toddler and the professional are not making equal contributions, but as long as the ball goes back and forth between them, they are having a reciprocal game.  Reciprocity is about acknowledging and responding to someone else's actions, not necessarily about how well this is done.

If the game ends quickly, we do not blame it on the toddler not throwing the ball far enough or missing the catch.  We instead point out that the catcher didn't stand close enough to catch the toddler's throw, or that he threw the ball too hard or too fast or too high or too far off to the side for the toddler to reach.  When a skilled adult plays with a young child, we expect the adult to adjust for the child rather than the reverse.

A typical social interaction between an autistic child and a neurotypical adult is similarly unbalanced, but we view it quite differently.  When such an interaction fails, we do not ask what the adult did to support their less-skilled conversational partner.  Instead, we blame the autistic child for their lack of skill.  We even say that the child "lacks reciprocity."  Indeed, "lack of social or emotional reciprocity" is one of the symptoms of Pervasive Developmental Disorder listed in DSM-IV.

Yet reciprocity, by definition, takes two. It is illogical and impossible to locate a failure of reciprocity to a single person.  To say reciprocity exists in a single person is like saying that there is one "equal" angle.

Autism researchers, however, do not necessarily realize this problem.  For example, Courtney Hale and Helen Tager-Flusberg (2005)1 measured "contingent discourse" as an indicator of reciprocal communication in 57 4-14 year old children conversing with their parents in the lab.  These children all had the ability to speak in phrases (however, their group IQ was well below average at 77; the range was 42-118). Autistic children's utterances were coded as "contingent" (continuing the topic of a parent's previous utterance), "noncontingent" (changing the topic), or "imitation."  The authors assumed that children with more communicative competence would be more likely to continue a parent's topic, while those with less would be more likely to change topics.  They found that autistic children who produced more noncontingent utterances also had stronger autism symptoms, indicated by higher ADOS scores.

Those familiar with the conversation of neurotypical children will already spot a glaring problem: typically-developing children tend to lead rather than follow parents in conversation, and given their short attention spans, change topics quite frequently.  Notably, this "contingent discourse" study did not include a typically-developing control group, so we don't know how impaired even typically-developing peers would appear according to this analysis.

But there's an even bigger problem: only the child's responses to the parent were coded for contingency.  There were no measures of parents' tendency to follow the child's topic.   This would be like setting up a conversation between a toddler and a professional baseball player, measuring how many throws the child made direct to the adult rather than off to the side, not  measuring the adult's throws at all, and claiming to measure "reciprocal throwing and catching."  Real reciprocity goes both ways. One cannot know whether or not a conversation was reciprocal without measuring the contributions of both participants.

Such faulty assumptions about reciprocity have more than academic importance.  If parents and teachers operate from them, they may miss opportunities to foster autistic children's social development. Just as a toddler learns to throw and catch better when able to play with someone who adapts to their skill level, so an autistic child can gain conversational experience and skill through talking with an adult who sensitively adapts to them.

A review by Morton Ann Gernsbacher2 catalogs numerous intervention studies showing that autistic children, like typically developing ones, benefit when others are responsive to them.   

Researchers taught four typically-developing preschoolers to either initiate interaction with three autistic peers or to respond to the interaction their autistic peers initiated (i.e., be reciprocal).  When the typically-developing preschoolers reciprocated their initiations, the autistic preschoolers responded more positively (Odom & Strain, 1986)3.

When an experimenter manipulates an object in the same way as an autistic child, the child makes much longer and more frequent eye contact with the experimenter (Tiegerman & Primavera 1984)4. Similarly, when their mothers imitate their manipulation of toys, autistic children not only make longer and more frequent eye contact, but engage in more exploratory and creative behavior with the toys, and show more positive affect (Dawson & Galpert 1990)5.  Autistic children display more positive social behaviors such as looking, vocalizing, smiling, and playing with the experimenter after repeated sessions where the experimenter imitates their behavior.  They are also more likely to sit next to and touch the experimenter (Field, Field, Sanders, & Nadel, 2001)6.

The few intervention programs that have trained parents in responsiveness to their children led to improvements in the children's social behavior (Aldred, Green, & Adams 2004)7.  Parents were taught to join their child's perseverative play, follow their lead, read their behavior to determine their interest, match their child's pace, and learn about their child's temperament and behavioral style.  Even though the child received no treatment, children's interactive behavior increased on average 50% in affect, persistence, interest, cooperation, initiation, and joint attention--the last of which improved 84% (Mahoney & Perales, 2003)8. These children also increased their social competence scores on a standardized test, the Infant Toddler Socioemotional Assessment (ITSEA).

The DSM highlights a real difficulty many autistic kids have in responding to others' overtures in a desired way.  These difficulties may, in part, persist because we do not adapt our interaction style and expectations to their level of skill as we do with typically-developing kids.  Autistic children lack much-needed experience with reciprocal interactions, not only because of their level of conversational skill, but because of our lack of responsivity to their needs.  But if we increase our responsiveness, we can open the door to truly reciprocal interactions, fulfilling their social needs and demonstrating the responses we desire.

Final Note:
The "catch" analogy came from a speech and language therapist, who taught her students that a conversation is like a game of "catch."  You catch the ball by responding to what they say, then "throw it back" by adding a question or comment to provoke a further response from them.  You drop the ball by failing to respond, and throw it somewhere off to the side by changing topic drastically without explaining your transition. This analogy struck me as a wonderful way to make the basic structure of a conversation concrete and visualizable for even very young children. 
1. Courtney M. Hale & Helen Tager-Flusberg (2005). The relationship between discourse deficits and autism symptomatology. Journal of Autism & Developmental Disorders, vol 35, no. 4
2. Morton Ann Gernsbacher (2006). Towards a behavior of reciprocity. Journal of Developmental Processes 1,  pp. 139-152.
3. S.L. Odom and P.S. Strain (1986). A comparison of peer-initiation and teacher-antecedent interventions for promoting reciprocal social interaction of autistic preschoolers. Journal of Applied Behavior Analysis 19, pp. 59-71.
4. E. Tiegerman & L.H. Primavera (1984). Imitating the autistic child: facilitating communicative gaze bvehaviors. Journal of Autism & Developmental Disorders, 14, pp. 27-38.
5. G. Dawson & L. Galpert (1990). Mothers' use of imitative play for facilitating social responsiveness and toy play in young autistic children. Development & Psychopathology 2, pp. 151-162.
6. T. Field, T. Field, C. Sanders, & J. Nadel (2001). Children with autism display more social behaviors after repeated imitation sessions. Autism 5, pp. 317-323.
7. C. Aldred, J. Green, & C. Adams (2004). A new social communication intervention for children with autism: Pilot randomized controlled treatment study suggesting effectiveness. Journal of Child Psychology & Psychiatry 45, pp. 1420-1430.
8.  G. Mahoney & F. Perales (2003). Using relationship-focused intervention to enhance the social-emotional functioning of young children with autism spectrum disorders. Topics in Early Childhood Special Education 23, pp. 74-86.


  1. I hesitate to say it, it's my own fault for thinking experts knew more than a mothers heart...but if I could go back...I would not look to science for an answer that was never forthcoming. It took me years to trust in myself and my child again. I spent 7 years in no-mom's land.

    So much of present work on autism depends on "expert analysis" and "othering". It is fear based.

    What astounds me about this analysis...is the logic.

    1. I'm sorry for what you and Ben went through. In your defense, it's hard to know that the experts *don't* have a useful answer until you've already absorbed their way of thinking and tried their suggestions, and by then much time and frustration have passed. I guess my mom was lucky because she had already learned, through dealing with me, that experts, neighbors and so on had little to offer a parent with an unusual (in my case, "gifted" and unusually intense) kid. Not everyone comes to the autism world similarly prepared to take things with a grain of salt.

      I've been reading child development books since my brother was a baby (so...about age 6?) and the dynamics for typical parenting seem quite similar to what you describe. Trust the experts and do exactly what they say. Press button, receive desired behavior; don't worry about viewing the world from your kid's point of view and understanding why they misbehave. So much fear, and with typically-developing kids, there probably isn't even much reason for it.

      As much as I love science, and think it can discover useful information about human behavior, I wish scientists would be a little more humble and realize the limits of their current knowledge. I also wish our education placed more emphasis on things like logic, close observation, and empathy for those we study. (Many of us do learn these things, but I think more informally).

      As a parent, what did you want to know, and hope the experts could help you with?

  2. I'm not sorry for what we went through. It's in the past, and I've gained understanding from it. When I knew better, I did better.

    I love science, too. The formal way of thinking allows for less "self-delusion" and opinion as fact. But as a human institution I can see in some instances it is way off base, and cruel, placing giant burdens on the backs of children, without regard to the "real world" or knowledge of long term consequences. It can be so simplified and wrongheaded, because finding "facts" is often an onerous and changeable thing, very complicated and must be reduced enough to be understood and analyzed, and shortcuts make it manageable. Scientists hold privileged access to these facts, each in their own area of expertise, whether by the nature of a superior intellect, or the advantage of a life that allows access to higher education.

    Scientists can be quite priggish in their own way because of this lucky break.

    What would I have wanted from science? I have a biological slant in my scientific leanings. Originally, I was looking for an etiological reason for my son's communication disorder. You know...a PKU type understanding. A miracle opening of the scientific mind, where a reality slipped in and led to a treatment. Something you could sink you teeth into. Your child's body has been poisoning him. You do this, and your child recovers. Every child that has this discernible maladaption can be treated in this way, and thereby avoid the worst case scenario of non-treatment. I wanted science to look to the brain to find reasons, not at the behavior to find a new way to divide humanity into new boxes...with no answers..like, "hey, we got another autistic here...too bad, that." Like people get " credits" for every kid they discover fits the label. And they keep widening the label, although they don't seem to have many answers. Job security.

    Therein lies the second thing I wanted. And I think in your reasoning you are moving towards it. Rather than codifying behavior.."this is autistic, this is not autistic" . This part of the brain lights up in autistics, so it must be defective, this part lights up in allistics, so it must be a good thing...I can't for the life of me figure out how that is helpful to the child. Unless--the child is not the priority.

    I want psychology to work with the child that is, not the child they perceive to be correct. Because of Autistic Pride, many parents have come to see that. It's a lot like being gay, not something you choose or can change.