Peter Hobson is professor of developmental psychopathology at London鈥檚 Tavistock Clinic and University College, researching mainly into child development. Unusually for a field that prefers people to wear one hat, he is also a psychiatrist and works as a psychotherapist for adults. He caused a bit of a stir when his team noted the clinical similarities between sighted autistic children and congenitally blind children. Vision, they argued, could be key to social and cognitive development.
What is it about those emotional connections in the early months that sets the stage for thinking?
Infancy is terribly important to all of us in the sense that this emotional linkage between us is what carries an infant forward and allows the infant to build upon its foundations in order to develop. If that isn鈥檛 going properly, yes, you can catch up later, but something fundamental is missing. The child with autism is unable to make use of that environment in a way that he or she can then build thinking and rationality.
Advertisement
Might people understand failure in the early environment as implying that bad parenting causes autism? Presumably you鈥檙e not keen on that?
Quite explicitly, I am absolutely not keen on that, and it is easy to get muddled up. We need to make that clear. I say in the book that genetics is very important in many cases of autism. For example, the incidence in identical twins is much higher than in non-identical twins, and the sex difference is four-to-one to the boys,
Autism is a difficult label. How do you characterise it?
Autism is a syndrome, a constellation of clinical features that just happen to go together. It鈥檚 terribly important not to become too concrete. It is simply a set of features 鈥� in particular, a profound abnormality in interpersonal relations, difficulties in communication including language, peculiarities in thinking, and one or two other difficulties.
Are there other conditions that may or may not be classical autism but which also teach us a lot?
In the book I highlight two special cases: congenital blindness and the abandoned children in the orphanages of Ceaucescu鈥檚 Romania. Each of those cases can lead to substantially increased autism, but not all blind children and not all Romanian orphans are affected. I emphasise that although the syndrome is there to be studied in blind children, it may well have a different natural history. Autism in the blind is the same in some senses and also different from autism in the sighted. In particular, blind children may show a capacity to develop out of their autism, although it鈥檚 true that many sighted children with autism do also develop.
But surely the Romanian orphans are an unusual and special case?
Yes, a very unusual and special case. It鈥檚 an absolutely drastic absence of human engagement in the case of the orphans. We are talking about something that is very extreme. There are many, many children who have unfortunate backgrounds but they don鈥檛 become autistic. But still the fact that autism can arise here does pose a serious challenge to traditional accounts of autism.
And the congenitally blind children?
My view on that is that, for most infants, other people lift you out of your own rather set way of experiencing the world. But congenitally blind children have difficulty with that because they can鈥檛 see facial and other expressions. And then there鈥檚 something more specific, which is that not only can a blind person not see your expressions and feelings and link them with theirs, but also they can鈥檛 know how you relate those expressions to a shared world. Quite literally they don鈥檛 see how the meaning of objects in their world can have different meanings for others.
What happens then?
It means that blind children can get relatively locked in their own world and less able to be flexible, less able than other children to play creatively, and less able to adjust their language. But this is probably a different kind of autism. One of the interesting features of blind children is that they echo back what people say. Why does this happen? Also they don鈥檛 play symbolically very well. Why should this be? For most two-year-olds, if you say something like 鈥淒o you want a bath?鈥� they realise you are talking to them. When they want a bath they don鈥檛 just repeat that phrase wholesale, they say 鈥淚 want a bath鈥�. That is something blind children find very difficult. Blind children can be like autistic children: they hear the expression and they link it with having a bath, so when they want a bath they will say 鈥淒o you want a bath?鈥� They haven鈥檛 understood the speech roles because vision gives you access to the switches in viewpoint. I am keen not to overdo the argument because they can be helped to connect. But if you left such children in institutions there would be many who failed to develop normally.
That sounds like Darwin鈥檚 work on expressions. Do you think much of this comes pre-wired?
Part of the thesis is indeed that we are pre-wired for certain things, such as reading expressions, but we are not pre-wired for what other people think. It鈥檚 both a psychological matter and a philosophical matter. That is, you have to have something that鈥檚 shared with other people in order for human thought to get off the ground. Expressions of emotion and the reactions to expressions of emotions are vitally important in that story.
How does this all connect to ideas about language being necessary for symbolic thinking?
There鈥檚 a big debate about whether you have thought and then you communicate through language, or whether you have communication and that gives you thought. Now I don鈥檛 think there is a strict opposition and I agree that symbolising, of which language is the most impressive manifestation, is indeed part and parcel of thinking. Symbols, once you understand how they work and how to use them, can crystallise thoughts so you can do mental things that you couldn鈥檛 otherwise do. Which to a large extent 鈥� and of course you can have other symbolic systems such as gesture 鈥� is what makes thought what it is. A lot of the way we think depends upon language 鈥� not language as a separate thing, but language as it emerges out of communication. So language itself is not only something separate that just adds on, but something that crystallises a potential in interpersonal exchanges.
Does that separate you quite strongly from those who argue that something separate emerges from language itself?
No, I think something does emerge. The argument goes like this: the first thing is communication, then language crystallises something in it, then it feeds back to give the detailed properties of thought. Language emerges as a system, and that leads to new thoughts emerging from language. I think what I disagree with is those who, like Chomsky and Pinker, want to separate out the meanings in language from the syntax in some important way. On balance what language is is a system of symbols to do things in relation to other people, to communicate. And much of the structure of language is implicit in non-verbal communication. In a hundred years or even fifty, I reckon that we鈥檒l have found that the area you need to explain by innate language structures has diminished quite substantially.
Tell me more鈥�
Let鈥檚 look at the route by which Chomsky and others arrive at what they think. What Chomsky did was to abstract language structure from everyday intercourse in the world, both cognitively and socially. Then he wanted a mathematically elegant characterisation of the structure. He said, if you look at the input, there is a poverty in the stimulus, you couldn鈥檛 possibly learn grammar from this, therefore you must have certain innate structures. There are some other arguments for that; and I don鈥檛 think that it is entirely wrong. But there is a very different way of looking at it.
Which is?
If you begin with communication, you start off with general principles true of animal communication. Then you have the rather more sophisticated matter of a human dealing with and understanding the world of objects as well as people. Put those together and add our ways of living with each other and acting with each other and feeling with each other, then out of that you can distil quite a lot of the structure of language. For example, rather than starting with nouns and verbs or subjects and sentences, you can actually think about children perceiving people as agents, acting on the world, and understanding what goes on between them. If you can symbolise these things, then much of grammar can build on that.
Back-track for a minute. How do we get to symbolic thinking in the first place?
That is one of the major issues that I attempt to tackle. Basically the question is: what anchors symbols? Take the symbol grounding problem, which is a problem for people who model thought on computers. The problem is that the symbols in a computer mean something to us but they mean nothing to the computer. If we鈥檙e going to explain human thought, we鈥檝e got to be able to interpret our own symbols. Now the problem is that computers don鈥檛 have the right linkage to the world 鈥� they don鈥檛 have a social life, or at least not one that we know about. So the account of how humans get symbols in the first place pivots around a story in which your linkage to the world and to each other is central.
How does that work for a child?
The idea is that the infant starts out by first connecting with other people. Then at the end of the first year of life 鈥� quite possibly by a biological process at that stage but also by learning through other people and by inhabiting a shared world 鈥� the child is lifted out of himself to see that there are different perspectives on things. By virtue of being powerfully affected by other people鈥檚 attitudes, a child comes to understand that there are objects out there that he is related to, and that these can have more than one meaning. In the middle of the second year, the child realises that what has meaning for me can have meaning for you, that there is a sound like 鈥渃at鈥� which is just a sound but that it can also have a shared meaning.
Is it just to do with meaning?
No. It鈥檚 also the fact that meaning is grounded in interpersonal relations. And the child discovers that they can make new meanings because they understand that鈥檚 what human beings can do. This takes place on two levels. One is that we are actually trying to get others to do things: obviously I鈥檓 trying to influence you to write accurate things in your article. The other is to do with the shared world, in this case the topic of thought, what thought is about. So what happens around the middle of the second year, when the child has already started realising that you can get things done with words, is that a new thing happens and the child comes to understand that the meaning for me and the meaning for you are interchangeable. When he or she uses a word, it is a candidate for my use; and when I use a word, it will be understood in the same way as I would understand it if you were to use it. It鈥檚 moving into the position of the other.
That sounds like a key to understanding autism.
It could well be a clue. In normal development, symbols arise from communication and common reference to the world, and my understanding that as a human being I can give things new meanings. That鈥檚 where we come to the French semioticians in that you鈥檝e got to have this understanding of the relationship between the signified and the signifier, between the symbol and that which it stands for. That is what this story is attempting to explain. And autism may arise when these processes go awry.
You will annoy people by concentrating on feeling as the main driver of the development of thinking. Why have you done that?
It鈥檚 not to stimulate disagreement, it鈥檚 just that I have a view on things which does highlight aspects of the mind that other people marginalise. Emotion and cognition are connected in an essential way. I start with a partly personal and partly intellectual belief that human interpersonal relations involving feelings are terribly important for the development of thinking. The feeling dimension seems to be not only important to our social existence, but also to the mind as it develops throughout life.
How do you juggle being a psychoanalyst for adults, a psychiatrist, and a psychologist studying children? What do your colleagues make of you?
It means that I鈥檓 a bit of an eccentric. I do feel that I am in each domain, it鈥檚 not that I water them down. But there is a big issue about methodology. The methodologies of experimental psychology, of psychiatry, of psychoanalysis, are all wildly different. You get people in each field who disparage the others. I happen to have respect for all three, not because I am wishy-washy and I think that everybody鈥檚 got to be friendly, but because I do think they all have their own great strengths. I have sometimes got bruised by colleagues because people constantly think that you are reducing one set of ideas to the other, but on the whole I have managed to be tolerated.
The Cradle of Thought is published by Pan Macmillan