Is Autism Statistically Linked to Early Non-Maternal Child Care?

Maxson J. McDowell Ph.D., L.M.S.W.

A printer-friendly version of this article

For information about the author

You are hosting a dinner party. When you have gathered all the ingredients the gas supply in your oven fails and you cannot cook the soup, chicken, vegetables, or pie. You serve salad. Hungry and unhappy, your guests leave early. Your party has been spoiled by a single early failure which caused a cascade of subsequent failures. Autism is like that dinner party. It represents a cascade of failures in psychological development, a cascade which is pervasive because it begins very early. What remains a mystery, in spite of decades of research, is the primary deficit which causes the cascade.

In some forms of autism, intelligence and language may be normal or superior. On average, mathematicians and scientists score higher than normal for symptoms of mild autism (Baron-Cohen et al. 2001, pp. 8, 10-11). Physical development is normal in autism. But a child suffering from autism treats others as though they were inanimate objects. The child does not make eye contact and appears not to know that others have thoughts and feelings which the child could share (Hobson et al. 1999, p. 55). Thus psychological and social development is defective. Any abnormality in the brain of a person with autism may be a secondary consequence of severe limitations in early social experience (Trepagnier 1998).

A biological primary deficit?

Mothers are often blamed for their children's problems. Bettlelheim said that autism was due to 'refrigerator mothers', a charge which tortured a generation of mothers of children with autism (Pollack, 1997). But there is no evidence that autism is caused by cold mothers. The pendulum has swung and it is now widely agreed (again with little evidence) that autism is caused by a biological deficit, that is, by genes or by some chemical or biological injury to the brain, or by both genes and injury acting in concert.

Though there is presently much focus on genes, several facts show that genes cannot be the whole story. The incidence of autism has risen sharply in recent decades (Byrd et al. 2002) while the gene pool has not changed over this time period. Although some alleles are statistically linked to autism, a child may be autistic without those alleles and a child who does have those alleles may not be autistic. For identical twins (who have identical alleles) one may be autistic and the other not (Rodier 2000, pp. 59, 63). Thus the alleles linked to autism are neither necessary nor sufficient to cause it. The foregoing shows that genes set the stage, making a child more susceptible to autism, but that an as-yet-unknown environmental factor or factors must trigger autism.

The incidence of autism has risen dramatically in many nations. The co-ordinated increase points to a single cause. It is highly improbable that one biological or chemical factor in the environments of many different nations has changed rapidly enough to account for this increase. Only a widespread behavioral change could fit this pattern.

A psychological primary deficit

At least two prejudices discourage consideration of a psychological cause for autism. One is the fear of blaming mothers. Another is the fear of psychology itself. A scientist or medical doctor may fear psychology because it suggests that he or she does not understand his or her own mind. This fear seeks explanations which reassure that the mind can be understood like the workings of an automobile: an undesirable mental condition (autism, for example) is due to a mechanism somewhere in the brain, a circuit of neurons which is malfunctioning and which might be fixed with drugs or with surgery.

But science itself has shown that the mind or, to introduce a term which is more useful for this discussion, the personality1 is not a mechanism. Holland (1998, pp. 229-31) and others have demonstrated conclusively that every living system is a self-organized dynamic system or a complex adaptive system rather than a mechanism. Since the personality is a living system, albeit a psychological one, the personality must also be a self-organized dynamic system.

An analogy may be helpful. A classroom includes, desks, doors with hinges, widows, lights with switches, blackboard, intercom and so on. It is a mechanism organized from the outside by designers and builders. It continues to exist whether or not energy flows through it. Intuitively you can see that this is not a good model for a personality. A class, however, organizes itself. Given a few initial components (a teacher with sufficient authority, willing students, adequate subject matter and regular meetings) a productive class will evolve spontaneously. Like all dynamic systems it only exists while energy flows through it, in this case from its participants. (If everyone in the room fell asleep the class would temporarily cease to exist.) A class accumulates a history, evolves a unique character and then dies a natural death, existing thereafter only in the memories of those who knew it. A class is a good model for a personality.

If one of the class's initial components fails, for example if the teacher does not have enough authority, this will trigger a cascade of subsequent failures which create an unproductive class. The initial failure is social and the resultant cascade of failures are also social, that is, they are all the same in kind. Non-social factors (for example a fire or a snowstorm) might sabotage the class but these would act indirectly, by increasing the probability of a failure of one of the initial social components.

In general, the primary event which triggers a cascade is the same in kind as the events which comprise the cascade. Autism results from a cascade of psychological failures in the developing personality. It follows that the primary deficit, the necessary and sufficient cause which triggers the cascade, must also be psychological. Genes and other biological or chemical factors, none of which are either necessary or sufficient, must act indirectly by increasing the probability of of the primary deficit.

The image of the mother's eyes2

In a recent paper (McDowell 2004) I presented evidence that a very early step in the self-organization of the personality is the internalization of an image of the mother's eyes. What follows is a summary of some of the evidence. For a complete account see Autism, early narcissistic injury and self-organization: a role for the image of the mother's eyes?

By comparing modern primates we can trace human evolution. Monkeys practice frequent joint attention, that is, they derive crucial information about their environment by following the direction of another's gaze. Humans (the descendants of ancestral monkeys) have evolved shared attention or intersubjectivity, that is, a mutual awareness of thoughts and feelings (I know that she sees it and she knows that I see it). Simultaneously, the appearance of the eyes has evolved: the scelera has changed from brown (in monkeys) to white and has become much more exposed, which provides more visual contrast between the colored iris and the sclera. Human eyes, therefore, give a much better signal of gaze direction. Thus the appearance of the eyes (their image) has evolved in a way that can only support intersubjectivity.

Experiments have shown that an adult human identifies complex emotional states in another person by observing the region of the other's eyes. A person with autism avoids looking at the region of the other's eyes.

At birth the range at which an infant's eyes can focus is very limited. This range coincides with the distance, during breast feeding, between the infant's eyes and the mother's eyes. While breast feeding, the mother spends about 70% of her time looking at the baby's face. A newborn cannot recognize whole forms, but gazes preferentially at sharp visual edges. In the mother's face, the eyes provide the most dramatic visual edges, for example between iris and sclera. An infant prefers to look at visual stimuli which move in response to the infant's movement, that is, at the visual edges of the mother's eyes.

At about six weeks the infant begins to maintain continuous eye contact with the mother. Continuous eye contact stimulates the mother to increase markedly the duration and repertoire of play with her baby. Such maternal play is crucial for the infant's subsequent development. An infant who, because of blindness or cranial nerve palsy, or because of institutionalization, cannot maintain continuous eye contact with the mother is statistically very likely to develop autism.

At about nine months an infant develops an intersubjective sense of self. If the infant could speak he or she might say 'I see that my mother sees me as lovable, so I see myself that way too.' Thus the infant's sense of self develops through continuous maternal eye contact.

To maintain continuous eye contact the infant must learn to recognize the mother's eyes as distinct from the other features of the mother's face. To recognize the mother's eyes, the infant must first internalize an image of them (when I recognize something, I match what I am currently sensing with a previously internalized image of that thing).

I propose that a very early failure to internalize the image of the mother's eyes leads to a pervasive cascade of developmental failure. A very early failure to internalize the image is the primary deficit in autism. The genes and other biological factors which are statistically linked to autism all increase the probability that the infant will fail to internalize the image of the mother's eyes.

Non-maternal child care

Recently the incidence of autism has increased dramatically. This increase has coincided with a dramatic increase in the use of non-maternal child care in early infancy (Rideout et. al., 2003). Early non-maternal child care has been shown statistically to predict problem behavior at 54 months (National Insititute of Child Health and Human Development, 2003). Non-maternal child care, which includes the use of television, video and computer games, must on average afford less eye contact than maternal care. Reduced exposure to the mother's gaze would make it more difficult for the infant to internalize the image of the mother's eyes.

The forgoing leads to a prediction which can readily be tested: autism may be statistically linked to early non-maternal child care. It is to be hoped that the current study in Norway, The Autism Birth Cohort (Manning 2004), will confirm or refute this prediction. Manning's description of this study, however, suggests that it is focussed on biological causes.


My hypothesis has implications both for public health (preventing autism) and for investigating the organization of the personality. It suggests that we can identify some of the earliest components out of which the personality self-organizes.


1. By 'personality' I mean all the internal processes and systems which comprise the psychological life of a person.

2. The terms 'mother' and 'maternal' refer to the very early mothering function, which may sometimes be fulfilled by a person other than the mother herself. For the purposes of statistical measurement, the term 'non-maternal child care' includes any caretaker other than the mother.


Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., & Clubley, E. (2001). The autism spectrum quotient (AQ): evidence from Asperger syndrome/high-functioning autism, males and females, scientists and mathematicians. J. Autism Developmental Disorder, 31(1): 5-17.

Byrd, R. (2002). Report to the legislature on the principle findings from The epidemiology of autism in California: a comprehensive pilot study.

Hobson, R. P., Lee, A. & Brown, R. (1999). Autism and congenital blindness. J. Autism Developmental Disorders, 29(1): 45-56.

Holland, J. H. (1998). Emergence from chaos to order. Reading MA: Perseus.

Manning, A. 2004. U.S. will pay for study to seek cause of autism. USA Today, February 24.

McDowell, M. J. 2002. The image of the mother's eye: autism and early narcissistic injury.

_____ 2004. Autism, early narcissistic injury and self-organization: a role for the image of the mother's eyes? J. Analytical Psychol., 49(4): 495-520. (Also online at

National Institute of Child Health and Human Development (2003). Study of early childcare and youth development: does amount of time spent in child care predict socioemotional adjustment during the transition to kindergarten? Child Development, 74(4): 976-1005.

Pollack, R. 1997. The Creation of Dr B.: a Biography of Bruno Bettleheim. New York: Simon and Schuster.

Rideout, V. J., Vandewater, E. A., & Wartella, E. A. (2003). Zero to six: electronic media in the lives of infants, toddlers and preschoolers. Report #3378, Henry J. Kaiser Family Foundation.

Rodier, P. M. (2000). The early origins of autism. Scientific American, February: 56-63.

Trepagnier, C. (1998). Autism etiology: a face-processing perspective. Brain and Cognition, 37(1):158-60.