La causalité chez l'enfant (Children's understanding of causality)


There are only three ways in which the birth of the notion of cause can be interpreted, if we put on one side for the moment a sociological interpretation to which we return later. The notion of cause results from either external experience, which is from associations imposed by things themselves; or from the feeling of internal experience arising from personal activity; or lastly from our perception of a relationship between things, or between things and oneself, and which then simply results from our deductive capacity. Hume defended the first of these three possibilities: causality, for him, comes from the habits that we acquire through the pressure of things. Maine de Biran supported the second possibility: causality is born out of volition and muscular effort. The third possibility is supported by famous rationalists in general.

Here, we are not engaging in philosophy, but child psychology. So, we will not discuss the basis of these possibilities – that is, we will not ask ourselves where they lead for the theory of knowledge. Our task will be simply to investigate their value as working tools in the analysis of the child's mind. Interestingly, all three possibilities are very fruitful.

When we examine the young child, aged 2 to 7, we constantly have the feeling that Hume and Maine de Biran are both right, although they contradict themselves. When we examine the evolution of causality as a whole, up until the age of twelve to fifteen, we have the impression that the concept of cause is an instrument of knowledge that reason constantly modifies and rebuilds. We shall briefly explore this paradoxical situation and then find a working hypothesis that will lead us to a description of the facts.

When we observe young children, we continually have the conviction that Hume is right: that is, the totally empirical and phenomenist manner in which the young child establishes cause and effect relations. So, he notices that the lights of bicycles and automobiles are switched on at night, and then concludes that it is the lamp that makes the bicycle move. If we ask the child how is it possible that this can happen and how the lamp makes wheels spin he does not know and does not care in the slightest. For the young child, as Hume says, ‘anything can produce anything’. However, at the same time the child behaves in a curious way; he believes that the lamp has causal force. He believes that it is alive and that it moves the bicycle forward on purpose. He says, ‘nice lamp’ or ‘naughty lamp’. In short, he conceptualises it as a kind of person. Now, this is something that cannot be explained if causality is born only from experience and from habit. Indeed, before any experience, the child already perceives things in a particular way. He projects his feelings on the whole universe. He does not distinguish the subjective from the objective: he is much less close to things than we ourselves are.

So Maine de Biran is also right. From the cradle, the child experiences his own muscular sensations, his states of pleasure and sadness, his desires and his disappointments. But he cannot discover the external world, that is, he cannot experience the resistance of things, without seeing things in a link with his self. This then gives birth to the notion of force, animism, in short of all elementary forms of causality, which are a great deal closer to the psychological reality than to the objectively conceived physical reality. Only, here again, serious difficulties arise. On the one hand, the child understands his self (his thinking, etc.) only by analogy to what he observes in the external world. On the other hand, the young child has only a very fragile consciousness of his self. He discovers things well before he finds out about his own person; also he discovers the person of others before his own. So, it must be said that he continually confuses the subjective and the objective. It is thus wrong to say that first he discovers his self and that only then he conceives things by analogy to this self.

Thus we can propose a working hypothesis, concerning the first years of the child that we will try to verify later on. We note that Hume and Maine de Biran contradict one another. Hume is right to say that children's understanding of causality starts with empirical trial and error, but he does not explain that the child understands things in relation to his self. Maine de Biran is right to say that there are subjective elements in primitive causality, but he does not explain why the child is interested in things well before discovering his self. This being the case, we shall simply take note of this paradox and suppose the following. We should acknowledge that at the beginning of mental life, there is no boundary between the self and the external world. This means that the newborn will not be able to know things in themselves but that, with any primitive knowledge, elements stemming from things and those stemming from the body are indissolubly linked for consciousness. In other words, for the baby, knowing consists of assimilating things to schemas from one's own action, in such a way that, for consciousness, things appear to have qualities, which in fact stem from the organism. If we continue with these assumptions, it seems that we could understand the whole development of the concept of causality. Indeed, two essential consequences can be deduced from this starting point, and observations will show how well founded these two consequences are.

The first consequence is that the young child will be at the same time closer to and further away from things than we are ourselves. The child will be closer to things than we are, in the sense that, given that his thinking is not yet aware of itself (because it proceeds by empirical trial and error and not yet by deduction), he will not be able to free himself from immediate appearances to build a deeper rational reality. However he will be further away from things than we are, in the sense that he will constantly mix into his own subjective conceptions, elements from which we adults have managed to free ourselves. Thus, we can say that the young child is at the same time closer to and further away from his self than we are: closer because he confuses self with everything else and further away because he is not aware of it. It is in this sense that Hume and Maine de Biran are both right even though they appear to contradict one another.

The second consequence is that while developing, the child will free himself from the empirical appearance of things as well as from the subjective character of his thought. So, to consider an example, he will no longer believe that the sun and the moon follow him when he goes for a walk and will instead come to appreciate that he is experiencing a simple illusion. At the same time, he will no longer believe that stars are alive and conscious like us. Indeed, by separating his self from things, so to speak, he will need to develop a better understanding of appearances: he will no longer be able to believe that the moon is following him if he discovers that the moon simultaneously seems to follow Peter, James and John. Rather, he will be forced to consider the moon as immobile or as moving independently from us. By separating his thought from things, the child will also stop conflating things with his self. He will no longer believe that inanimate things possess consciousness, life and animated forces. In this case, causality, envisaged dynamically, and not just at its origins, is a product of reason: by separating his self from things, that is, separation at the same time from empirical appearances, and also from subjective illusions, the child is obliged, little by little, to construct a reality deeper than that of immediate experience; and all causality consists of such a construction.

Given this, we will set out to analyse the facts. We will try to describe as objectively as possible the stages of children's understanding of causality, and we will try to verify the extent to which our working hypothesis is well founded. Since causality is relevant to nearly all mental processes in the child, we will need to limit ourselves to only one category of explanation, the explanation of movements, even if, occasionally, we venture into related fields.

We are going to ask children how clouds, stars, rivers and wind move forward, what waves and drafts are, why clouds stay in the sky and boats on the water, while stones fall to the ground or sink to the bottom of water. By classifying the responses, we will establish the existence of four main stages in the child's understanding of causality.

During the first stage, numerous manifestations of which we have shown exist until about the age of three to four with some traces lingering until later in development, children offer explanations that are magical and phenomenist: things are linked to certain gestures that we make without any link that is spatial or intelligible.

During the second stage, which spans from approximately the age of three to eight, the cause of movements in nature is moral, and, by analysing this type of causality, we will find a mix of animism and artificialism. This moral causality does not exclude some knowledge of some physical links, but the physical and the moral are still confounded.

During the third stage, which spans from approximately the age of eight to eleven, the cause of movements becomes principally physical; however children's notions of physics at that age are impregnated by dynamism: bodies are moved by substantial forces, as in Aristotle's physics, and the child only has a vague idea of the mechanism itself, with its rational principles of conservation.

Finally, around the age of ten to eleven, a mechanical explanation of movement appears, as well as certain rational principles such as the conservation of weight.

As we can see, this developmental progression seems at first to be linked with what we said previously. On the one hand, causality seems to be progressively de-subjectivised: firstly it is magic, then moral, then it becomes dynamic, and finally mechanical. However, at the same time, rational construction seems to progressively replace empirical association: phenomenism is succeeded by a hidden moral order, then by some forces that the child tries to make intelligible, and then finally by logical principles that experience simply suggests, without imposing them, and that can only be constructed through reason.

I. The primitive magico-phenomenist stage

In a few days, I will have the honour to analyse, in front of the British Psychological Society, the concept of causality in a baby just a few months old. I will try to show that once the baby perceives a link between one of his own gestures and a movement in the external world, he attributes efficacy to this gesture and then uses it to reproduce the movement previously perceived in the outside world. So, a baby will half close his eyes in front of an electrical lamp cord to turn the lamp on, or he will carry out a hand gesture to make an object move from a distance. Thus, it seems that the primitive causality is at the same time both phenomenist and quasi-magical because the baby does not make any distinction between his self and the external world and considers his muscular efforts as extending into things themselves.

Curiously, we find many remains of these primitive behaviours in the three and four-year-old child. We will analyse just two of them.

First, there are the most primitive explanations of the movement of stars and clouds. The child discovers very soon that when he walks, the stars and often the clouds seem to follow him. Thus there is a link between the child's walking and the movement of the stars and clouds that his experience imposes, and which, as such, is of a phenomenist nature at least in its origins. In view of such an observation, we can take at least three possible attitudes, which are interesting in terms of causality because they reveal clearly the orientation of the child's mind.

One of these attitudes is the critical one. We notice that trees or a wall move behind us when we walk beside them and that landscapes flee from us when we are on a train. So the child, who is at first fooled by these appearances, quickly frees himself from them. Regarding the stars, he could wonder what the moon will do if he goes in one direction and his brother goes in the opposite one. In this way, he will end up considering the movements of the stars that follow us as pure appearances.

The second possible attitude is the animist one. If the sun and the moon follow us, it is that they want to follow us. They look after us, watch us, keep an eye on us, wait for us when we enter a shop, and so on. But, if they wanted to, they could go away. If they follow us, it is therefore because they wish to do so.

The third possible attitude is the magical one, or the attitude of believing in one's own role and own personal efficacy. If the stars follow us, it is because we force them to follow us. They obey us. We make them go forward, or stop, and adjust their speed. They cannot disobey us.

The first of these three attitudes is definitely the most cognitively demanding, because it assumes critical thinking. Therefore, it is quite natural that it is the one that appears last. Indeed, it is only around the age of eight, on average, that it becomes common (in three quarters of the children), that is, in our third stage, the stage of dynamism.

The second and third of these three attitudes seem logically to be of equal difficulty. Nothing in our experience tells us whether the moon follows us because it wants it, or whether it follows us because we force it to do so because it follows us all the time. So, it is interesting to wonder which of these two modes of thought is most primitive. The statistical results are very clear. The child begins by believing in his own personal efficacy and by thinking that that he forces the stars to obey him. The animist attitude only appears second.

Chronologically speaking, responses favouring a belief in personal efficacy are thus characteristic of our first stage, and are particularly frequent around the age of four. Answers, which reveal that the stars follow us because they want to, are especially frequent between the ages of five and eight, and are characteristic of moral causality, as we will see later.

We move on to the second example, which will serve at the same time, as a counter-proof. To verify our previous analyses of beliefs, we have imagined a display presenting the child with a convergence between one of his gestures and an external movement. The task is comparable, therefore, to the case of the stars that seem to follow us. We show the child a pipette full of water that naturally excites his interest. Then we say: ‘When you want the water to fall, make a little sign with your finger’. The child moves the finger and, without allowing him to see how we do it, we let a small amount of water fall from the pipette. The child repeats the gesture and once again we allow some drops to escape, and so on.

Here, as previously, three orientations of mind are possible in the presence of this mystery. The child can have a critical attitude and say: ‘I don't understand it, but there must be a trick. It must be you and not me who makes the water fall’. Or, the child can take a purely phenomenist attitude, and say: ‘There is a link between the movement of my finger and the fall of the water, but I have nothing to do with it. Probably, my finger moves the air, or something similar is happening’. Or, finally, the child can take the attitude of believing in personal efficacy, and say: ‘It's me who makes the water fall because I forced it to fall and it obeys me’.

Interestingly, it is once again the belief in personal efficacy that is the most primitive. Young children, from age three to four, are immediately convinced that they have power over the water in the pipette. If it is, for example, a boy who does the task, and a girl pupil subsequently does the finger sign without success, the boy merely says: ‘She doesn't know how to do it. It's not for girls; it's only for boys … ’. On the other hand, around the age of seven the child considers that there is a trick behind it and that it is the experimenter who makes the water fall. Between the two stages, we can see a progressive decrease in the belief in personal efficacy. Thus, we can see that these results parallel those relating to how children explain the movements of stars.

In short, it seems that for the baby, from six to eight months to around two years old, primitive causality takes the form of a kind of belief in the causal value of the gesture, a belief that is encouraged by the successes obtained during all empirical or phenomenist experiences. It seems, furthermore, that this form of causality reappears each time the opportunity is given to the child until a time that is difficult to specify; nonetheless these remaining traces of primitive causality are frequent until around the age of four.

We need to add, before we leave this first stage, that magico-phenomenist causality would appear to correspond to a representation of things that reminds us of the famous notion of ‘participation’ so well described by M. Levy-Bruhl in the mentality of so-called primitive people (Where participation is concerned with people's feelings towards things around them as well as with the evaluation of the thing's relation to the speaker). I do not wish to raise here the question of the genesis of these participations of the baby; a question that is very complex because it touches on the development of the concept of things. However, it is worth noting the explanations four-year-old children give to account for the appearance and disappearance of certain phenomena, especially difficult to conceive of as things, permanent and identical to themselves, such as the explanation given for drafts and shadows.

Using a fan to create a draft, we ask the child how the draft that he just experienced was formed. Let us make sure, however, that the experiment is done in a room where the window was closed before the child arrived. From the age of six to seven, the child gives some explanations that do not interest us here, and which amount to saying that the fan has created air by its own movement, because the child does not know that there is motionless air in the room. However, children of the age of four and often also five give us a much more curious interpretation. Naturally, they do not know, and a fortiori, that the room is full of air and so they think that the fan, by its movement, created the appearance of a draft, which, at the same time, comes from the wind that blows outside the window. Indeed, the child, feeling the draft, says straightaway: ‘that's wind’, and when asked where the wind comes from, he claims that it comes from ‘outside’. It is pointed out to the child that the window is closed, but that does not bother him. He maintains that the wind from the fan is nothing other than the wind from outside, which is guided by the fan when it is working. When the fan stops, the wind leaves, when the action is repeated, it comes back.

It is the same for shadows. The child naturally thinks of a shadow as a substance, and not as an absence of light. But it is also a substance that, like air, has the power to move itself instantly. When one produces a shadow on the table with a book, the four-year-old child recognizes fully that it is the book which produces the shadow but, at the same time, he claims that the shadow comes from outside. There is a lot of shadow under trees; there is also some in the sky, in the clouds that produce the night. So, when one puts a book on the table, this shadow from the trees or the sky is channelled there. It disappears when the book is lifted, and returns when it is put down again!

It is clear that these strange beliefs simply come from the child establishing an identity where we only make a comparison. Instead of saying: ‘this shadow of the book is similar to the shadow of the trees’, or ‘This draft is similar to the wind’, he says: ‘This shadow is that of the trees’, and ‘This draft is the wind’. But, it is really difficult not to perceive in these responses a stage prior to our properly logical identifications, since here the identity is established without any worry about spatial contact: the shadow, which was in the sky, is now under the book, and returns then to the sky, and so on. One could almost argue that things are simultaneously in different places, if it were not better to simply say that the child establishes a link without caring at all for geometrical connections. Thus, there is something here that is similar to participation. As was seen earlier, such representations are the natural extension of magico-phenomenist causality, as just described, and whose essential characteristic is precisely not to take into account spatial relations.

II. The stage of moral explanation

If our analysis of the first stage of explanation in the child is correct, we have verified one part of our working hypothesis: causality starts with the maximum of phenomenist empiricism together with the maximum of subjectivist dynamism. The child who believes that he makes the sun move forward is, in fact, closer than we are to the direct experience of things, but at the same time he attributes to his personal efficacy greater powers than he will do later. At the limit, the baby projects his self on to the whole universe while being very susceptible to the influence of his immediate experience.

What will become of this situation during the second stage of causality, roughly from the age of three to seven to eight years? On the one hand, the child will continue to be seduced by the appearance of things, but a little less than during the first stage. Thus, he will continue to believe that the stars follow him, but he will believe it less easily for clouds. He will continue to put stars and clouds on the same spatial plane, but he will situate them higher in the sky than previously. He will no longer believe that he only needs to close his eyes to produce the night, but he will still think that if everybody goes to bed in the middle of the day it would become dark straightaway. In short, he will remain very phenomenist but to a lesser degree than previously. When it comes to the subjective aspect of causal links, the child's belief in his own personal efficacy will gradually diminish into a link that is simply moral.

In fact, the sooner the child learns to distinguish his self from that of others, the less he will attribute to the efficacy of his own gestures. By learning to imitate others, and then, thanks to language, by learning to obey his parents, the child has an essential experience which will impact upon his representation of the world. He will conceive of the universe as a vast society of living beings subject to a set of duties and constraints, and causality will then be construed as coercion, half-physical and half-moral, analogous to the control that adults exert over their children.

To verify that, we asked children aged from seven to eight about the movement of the wind, clouds, stars and rivers. In this way, we will find a really clear result confirming what the sadly missed Prof. James Sully had outlined in his well known ‘Studies on childhood’: it is that these bodies are conceived of as big well-behaved children, who accomplish their duty correctly.

In fact, during this stage of moral explanation, we find two different types of answers, which characterise two successive sub-stages. During the first sub-stage, which spans mainly from three to six years, the child represents things as living beings directly obeying Man (or God, which amounts to the same because they define God as a ‘Mister’). During the second sub-stage (from six to eight years), things are subordinated one to another and the whole of nature is subordinated to moral rules, which are still anthropocentric, but in which explicit and mythological artificialism of the beginning of the stage no longer appears directly. From the point of view of causality, described here, the structure of the explanation is thus the same during both sub-stages.

Beginning with the first sub-stage, which is thus characterised by a mix of animism and artificialism, we asked four to six-year-old children how the wind is formed and why air moves forward. The answers are always the same. The wind ‘It's someone who has blown – Who did that? – Some people – Which people? – People who had this job’ (a six-year-old), or ‘It's when we whistle; that pushes the air out’, or it is God and people who produce air with a bicycle's pump. As for knowing why air moves forward, the cause is physical on the one hand, as we have just seen, but it is mainly moral: the wind blows ‘in order to break down trees to make a fire’ (a six-year-old), or ‘to make lots of waves that then pushes the little boats so that we don't need to row’ (a six-year-old), or ‘to bring clouds, to bring rain’, and so on. The wind is thus thought to have a set of functions in the economy of the movements of nature, as we will see subsequently.

Then we asked the child how clouds move forward. We find here, as for the wind, a series of transitions between the explanation by personal efficacy and the moral and artificialist explanation. Thus, my daughter, at twenty months, said while seeing clouds: ‘fog smoke Daddy’, expressing in this way that she assimilated the clouds to the smoke from her father's pipe. In these conditions it is obvious that the movement of clouds will very early on be explained by the action of adults. Thus, a three-year-old child said to me about clouds: ‘It is the train driver who makes them move’, because he had seen smoke coming from the funnel of a train. Other reasons will involve workers, masons, or even God, who will be intermingled with children's explanations more or less according to the example along the lines of the hypothesis that we make the clouds advance by walking, or they follow us as we walk.

Here again whatever the physical means put forward, the causality of movement very quickly becomes moral. Clouds move forward in an orderly rather than random fashion because they have a series of roles to fulfil. They move forward because they must make or forecast rain. They must, in certain cases, push the sun and the moon, hold up the sky, and so on. Above all, they have a function, really unexpected for the adult, but to which the child often gives thought: they make the night. Indeed, for the child as for the first thinkers from Greece to Empedocles, the night is a substance, a kind of black steam, which fills up the entire atmosphere. Besides, as we have already seen, it is the same thing for shadows, and children believe until around the age of nine that things also produce shadows even during the night, only we cannot see the evidence precisely because it is night time! This black substance constituting the night is conceived of as stemming from clouds, or as a big dark cloud. Consequently, one of the functions of clouds is to make the night; thus, when asked point-blank why clouds move forward, children often answer: ‘to make the night’, or simply ‘because of the night’.

The movement of stars is explained in a similar manner. On the one hand, it is humans or God who makes them move forward. On the other hand, if they move, their purpose is to give warmth and light and to keep an eye on us, to lead us. The star guiding the Wise Men to Jesus' cradle is obviously a vestige of this stage of the explanation of movement.

Finally, the movement of water in rivers is explained in a similar way. On the one hand, it is the boats or the oars that make the river move forward. On the other hand, the water of the river is conscious and alive: it flows to make boats and swimmers move forward; it also flows to go to the sea or to give us water.

In short, in all these primitive explanations of the second stage, one observes the same mix of artificialism and animism. The explanation is artificialist in the sense that it is people who constitute the first and ultimate cause of all natural activity. But, this artificialism assumes animism, in the sense that in order to serve men, things need to be alive, conscious of their duty, and intelligent enough to accomplish it. That is what I am trying to convey in claiming that the causality of this stage is essentially moral.

If now we pass on to the second sub-stage, to the explanations of children who are on average between six to eight years old, we find a similar type of causality, but transposed to nature itself. In other words, it is no longer people who are thought to cause the movement of things, but the things themselves. At first sight, this appears really very different since the explanation seems to have become purely physical. But, as soon as we analyse the responses more closely, we notice that nature is still conceived of as a society of living beings subject to rules; therefore only the form of the explanation has changed, while in fact the underpinning has always stayed essentially moral and anthropocentric.

Here are some examples. When one asks the child how clouds move forward, he answers that it is the night, or the rain, or the cold and bad weather, or the sun and the moon, which make the clouds move, or finally that they move on their own. All this at first appears strangely fanciful and incongruous. In reality, all these explanations amount to saying that clouds must move forward to fulfil their own function towards men. To say that the night causes the movement of clouds is again to simply state that clouds must move forward to make the night. To say that the night or the cold makes clouds run is effectively saying that clouds must come to give us rain or to announce that rain is on the way. To say that the sun and the moon are the cause of the movement of the same clouds, is saying that the sun chases away clouds just as a policeman chases away a vagrant, or that the moon accompanies clouds in their journey towards the night. In short, the bottom line is that there is a fundamental confusion between physical law and moral or social law.

Similarly, the movement of the stars is due to the clouds, to the rain, to the night, to the wind, in that stars, since they are living and conscious beings who have some specific duties to fulfil, flee from the rain to bring heat to somewhere else or appear at night to give us light.

The explanation of the movement of rivers presents some interesting particularities. One common response, for example, is that the water of rivers moves forward because of stones. In fact, this explanation has again a meaning that is both physical and moral. On the one hand, a stone, being heavy, weighs down the water and produces a current, but above all, on the other hand, a stone makes the water move forward because it presents an obstacle, that is, in the vicinity of a stone the water ‘takes a spring’ to jump over, as a horse in front of a hurdle, and it is this ‘taking a spring’ which creates the current. However this current, says the child, is necessary because the water ‘must flow’ to drive boats or to give us something to drink.

The floating of boats or the suspension of clouds offers the same puzzling mix of empiricism and morality. One child will claim both that the little boats float because they are light and the big boats because they are heavy. Similarly, another child will say that the stars stay in the air because they are light and clouds because they are heavy. The child will justify this by saying that light things are supported by water or air, and that heavy things have the strength to support themselves on their own. But, the underlying idea is again that the rules, which determine flotation, have a quality that is as much moral as physical. Clouds ‘must’ stay in the air because otherwise they would kill us in their fall. As to why stones sink to the bottom of the water and why boats float, a six-year-old child answered, and this is very representative of this stage, by saying: it is because boats are ‘more intelligent’ than stones, this is because ‘they do not do what they must not do!’

In short, until around the age of seven to eight, the explanation of movement, as with the explanation of causality in general, reveals a perpetual conflation of the physical and the moral. Movements have or have not, depending on the case, causes that are properly material. But, when they have, the physical mechanism of movement is only an instrument to serve the real cause, which is a moral cause: this real cause, the obligation of things to accomplish their function, conforms to a harmonious plan of which man is both the author and the raison d'être.

III. The stage of dynamist explanations

Up until around the age of seven to eight years, the physical causes of movement are neither necessary nor sufficient. Certainly the child already says that it is the wind that makes clouds or stars or rivers move forward. However the wind is only a means used by these things, just as birds use a breeze when flying. Without this additional help, their movement would still be the same, and this supplementary influence of the wind is not regarded as sufficient to explain the deep cause of movements.

Starting from the age of seven to eight, in contrast, any movement is conceived of as resulting from physical mechanisms that are necessary for its production. Besides, although bodies, or at least some bodies, continue to be considered as being alive and conscious, the physical cause of movement appears more and more to be a sufficient explanation for the child. However, as we are going to see, this physical cause is still conceived of in a very particular way from the ages of seven to eight until the ages of ten to eleven.

Indeed, it would be rather improbable if the childlike animism, which is so prominent until around the ages of seven and eight, suddenly disappeared entirely. We note, on the contrary, that bodies conceived of by the child as having an inherent force of motion, as opposed to those for whom movement is caused by an external force, continue to be considered as being alive and conscious. Now, we are going to see that most natural beings appear to the child as enjoying the capacity to move at least partly on their own, as opposed to machines, which are controlled by man. Furthermore, even when the child denies that a body has consciousness and life, he still credits it with a property that strangely resembles life without consciousness, and which is nothing other than a force, conceived of as a substantial force, inherent in the bodies themselves.

In these two cases, the one where the bodies continue to be held as alive and the other where they are simply invested with substantial forces – and these two cases almost merge with one another – bodies continue to enjoy, during this stage, a kind of spontaneity. This spontaneity is interesting in terms of cross-cultural psychology, because, if the explanations offered in the previous stages are reminiscent of thinking in primitive people; the dynamism of this third stage is strikingly reminiscent of some aspects of Aristotle's physics.

The principal characteristic of this childlike physics, and a very peripatetic one, is that all movement supposes the competing involvement of two distinct forces, an internal force that is the moving thing's own activity, and an external force, which is the physical action responsible for the movement of the moving thing. Note immediately that this feature is explained simply as a legacy from the previous stages. Indeed, up to now, any movement was explained with reference to two forces, one internal, and the other external in relation to the moving thing. The external force, until now, amounted to nothing more than the product of human volition, and the internal force was seen as the duty of obedience and realisation of the moving thing. From now on, the external cause of movement is no longer moral and becomes physical: it is no longer the action of people, but an external factor that exerts pressure on the moving thing, which explains its movement. On the other hand, the internal cause of movement is no longer conceived of in a principally moral way: the submission of the moving thing to the external cause is no longer conceived of as obedience, but as a kind of animal suggestibility. Thus, the wind makes the clouds move forward because there exists inside the cloud a tendency to move forward. The bodies have a certain number of specific tendencies, and the external causes that act on them are channelled in a way to trigger these tendencies.

One can see that the theory of the two forces seems to derive directly from this mix of artificialism and animism, or even of magic and animism, particularly with respect to the early stages. Furthermore, this conception will lead the child step by step to mechanism as a causal principle. We have thus to expect that, thanks to the traces of animism which we have just mentioned, there are all the intermediaries between the explanation of movement that is literally animist, the explanation by the two forces, and finally the mechanical explanation. We can indeed conceive of the following four periods of transition, and we will see that we indeed encounter them in the child's mind:

  • 1Bodies move forward all alone to accomplish their function. For example, clouds ‘fly’ like birds.
  • 2They produce, in order to move forward, an external body that pushes or drags them. For example, clouds produce wind, and this wind pushes them.
  • 3They are moved by an external body but, while moving, they reinforce the action of this external body. For example, clouds are pushed by the wind but, while moving, they strengthen the wind.
  • 4They are moved by an external body that is sufficient to explain their movement. For example, the wind pushes the clouds, and so produces their movement.

As we see, the first of these four types of explanation is typical of the early stages, and the last is typical of the mechanical stage. The two others suppose the scheme of the dual force. Let us further study types 2 and 3, which characterise the present stage and which thus form the transition between the animism of the previous stage and the mechanism of the next stage.

Type 2 seems extraordinary. One hardly understands, at first sight, how the child will come to think that bodies, in order to move, produce other bodies which push or drag them. In reality, interestingly, this type of explanation is extremely frequent at the beginning of the stage that we study now. The existence of bodies such as those that the child conceives under the names of ‘wind’, ‘heat’, and ‘current’ certainly facilitates the formation of this explanatory scheme.

We must say here a few words about children's ideas about air and wind. Recall the experiment where we wave a fan in front of a child, in a closed room, and ask the child where the air produced in this way comes from. We have seen the responses of the first stage. What we need to specify now is that, up until the age of roughly nine years, that is, up until around the middle of the stage that we are studying now, either the child does not know that there is air in the room and believes that the air of the fan is created by the fan itself, or else the child declares that even in a room without air the fan will produce wind. So, things in movement create air. Curiously, it is precisely in this way that the child explains the formation of wind: the wind is the result of the movement of clouds, of trees, of waves and even of dust! The clouds, and sometimes even the stars, create wind by moving. The trees swing by themselves and make wind. The waves and the dust are able to lift themselves, which in turn produces wind.

In these conditions, the explanation of movement becomes relatively simple. As far as clouds are concerned, things are easy to understand. The cloud sets itself in motion. It produces in this way a light draft, which then circles around to the back of it to blow from behind which, in turn, pushes it forward. Moved in this way by the combined action of the air and its own force, the cloud moves faster and provokes the formation of a stronger wind. As we see, the child imputes to the cloud an experience that he had himself: it is indeed the case that, when we run, we produce a draft. But the child has the impression of being pushed by this draft and to have created himself this flow of air that pushes him in this way, when on the contrary it is, as we know, the momentary vacuum producing this reflux of air hindering the journey.

As far as the stars are concerned, the child sometimes repeats this explanation, or he imagines that a warm current produced by the sun makes them move. This is what he means by claiming that ‘the heat’ is the cause of the movement of the sun.

Rivers conform to a similar scheme. They firstly move forward thanks to their ‘current’, in other words, on their own. But then the child reifies the current, and transforms it into a kind of substance that pushes the river. So, in most cases, and probably helped by language, he identifies this ‘current’ as an ‘air current’ and comes to believe that the river is also pushed by the air which the river produces itself. Without always going that far, the child very often claims that waves produce wind, which then generates other waves; so, these waves help the flow of the river.

It can be seen that such explanations derive directly from the animism of the previous stage since, at the starting point, things are supposed to set themselves in motion. However, there is great progress, since, for the child, this actual movement is not enough: there also needs to be a material mechanism that takes the form of an external cause given to the moving thing. However, in these first attempts at a physical explanation the external force is conceived of as produced by the moving thing itself. We are then at a critical point, the junction between the animist and the dynamist schemes of the dual force scheme.

At this point it suffices that a flow of air is no longer conceived of as created by the moving thing, in order for the child to begin to understand the third type of explanation mentioned before: the moving thing is moved by an external body, but while moving it reinforces the action of this external body. In other words, for example, the cloud is pushed by the wind, conceived of as coming from somewhere else. But, when moving forward, the cloud increases the current of wind, which has the effect of increasing the movement of the cloud itself.

In practice and with regard to the movements of nature itself, as with the movements of clouds, stars or rivers, we will have great difficulty in distinguishing this type of explanation from the previous one. This goes without saying, therefore it could appear absurd, and overly interpreted, to make this distinction. However it is not just a matter of preference for symmetry or orderly logic that we do this. Indeed, as far as some artificial movements are concerned, such as the movement of a ball in a room, or of a spinning top, or of a body hanging from a string, the child makes the distinction by himself. In others words, even when around the age of nine, ten and eleven he conceives of the room as filled with air, and even when he realizes that bodies do not create air while moving, he continues to explain the movement by the backflow, or reflux, of air.

This point has a certain historical interest because, as one can immediately see, this explanation strikingly resembles one that Aristotle gave for the movement of projectiles. We know, indeed, that the movement of projectiles was greatly embarrassing for Aristotle's physics. On the one hand, by virtue of the theory of the two forces it was necessary to explain why the arrow, for example, continued to move forward once parted from the bow despite having left behind its external force. Indeed, the bow cannot act on the arrow at a distance: thus, according to the theory of dual forces, the thing cannot move forward if its external source of movement does not accompany it! On the other hand, for Aristotle, the natural site for heavy bodies is the ground and therefore the arrow should have fallen down as soon as it left the bow. To this double question, Aristotle answers in the following way: on the one hand, the projector continues to act on the projectile, even from a distance, because it delegates its movement to the ambient air. In other words, it shakes the air and then the air pushes the projectile. On the other hand, the projectile itself shifts air when moving forward, and the air shifted in this way comes to flow back behind the projectile, which thus continues to force it forward. This last process, or the process of air flowing back (reflux), was called by Aristotle the αντιπεριστασιζ.

We see that this process is exactly the same as that described for our children, except that the moving thing does not create air while moving forward, but simply produces a current in the surrounding air. However, as we have just seen, the explanation of air flowing back persists in the child even after having understood this last point. To be sure, we wanted to confront the child with the problem suggested by Aristotle: why, when we throw a ball a great distance, does it not fall to the ground immediately after leaving the hand that has thrown it? Most children aged nine and ten told me that the ball continued its trajectory because of the draft. Some attributed the draft to the hand that throws, others (more numerous) to the ball itself. Nevertheless, these children insisted that the ball did not ‘create’ air. For them, in a room empty of air, the ball would have fallen to the floor as soon as thrown! This is precisely the explanation offered by Aristotle, which we can thus consider not as a knowledgeable explanation but a simple coding of the commonsense ideas of his time.

Finally, in order to finish the study of this stage, we come to certain explanations from which burst a dynamism specific to the child, where we might have expected mechanical explanations.

Let us first say, when the child simply attributes to the wind the movement of clouds, stars and rivers, without admitting that this wind is due to the moving things themselves, the explanation is not at first mechanical. Often the child still accepts the idea that, without any wind, these bodies could move weakly. Thus, one has often the impression that, for the child, the wind triggers the internal force of these moving things rather than conveying this moving force to them from outside. However, the child rapidly gives up his reservations, and so we move to the fourth stage that we will tackle in a moment.

We then asked the child how clouds can stay suspended in the air, or how boats manage to stay afloat. Interestingly, between the age of eight and ten, that is, at the age of explanations based on the dual force that we have just studied, the child invokes more dynamist causes than mechanical reasons. Indeed, suspension or flotation are not yet considered to be due to the relative body weight of the body considered relative to the density of the fluid in which it is immersed. The weight does feature a little but as a symbol of force. The true explanation, for the child, is akin to gliding. The clouds stay in the air because they go fast enough to be supported by their own movement, or by the draft that drags them. In the same way, the boats float because they go fast. One then naturally points out to the child that there are some motionless clouds and boats. ‘That's true’, answers the child, ‘but that is only for a while. That's the run up. If they stop for too long, they will fall’. This actually never happens, but one remembers that everything is for the best in the harmony of things for the child.

Finally, we asked the child why the water in a glass changes level and rises when one immerses a big stone. Until the age of nine, the child does not think of the simple reason that the stone occupies a certain volume and moves the water for an equivalent amount. The answer is only: ‘the water rises because the stone is heavy’. Superficially, this seems to be the same and it seems that the child simply wants to say that the stone, being heavy, sinks to the bottom and so occupies a space in the water. But, in fact, the child's conception is again more dynamic than static. He thinks that the stone weighs on the water and creates a current that makes the water move up. According to the child, a small bag of lead shot that is heavier but smaller than a piece of clay will make the water rise higher than the clay; and yet both of them will go to the bottom. Also, the child considers that a stone maintained at middle height of a column of water, would not make the water rise because it is suspended by a string and does not put weight on the water.

In short, it is noticeable that, in every domain, the responses of this stage show of a dynamism in full, such that bodies are considered as endowed with strengths, activities, and spontaneity. This is important progress, compared to the previous stage, because the explanation becomes physical. But the physical explanation, which appears for the first time in its generality, is still impregnated with animism.

IV. The stage of mechanical explanations

During the fourth stage the physical explanation refines itself. It is difficult to locate the beginning of this stage precisely. Certain signs announce the stage from the age of seven to eight years. Mechanical causality replaces the dynamism of some domains by the age of nine to ten. Nevertheless, we can give the average age of ten to eleven as the age for the eradication of animistic dynamism, thus marking the real starting point of the fourth stage.

We will be brief about this stage because it is the one during which the child starts to think like an adult. However, it is useful to note some convergences between the points in development at which the principal characteristics of mechanical causality appear.

It is around the age of nine to ten years that the explanation of natural movements becomes mechanical. About the age of ten, the child refuses to explain, as he had previously done, the formation of wind by the movement of clouds or trees: also around about nine to ten, the child explains the movement of clouds by simply saying that it is the wind that pushes them. There is no longer any trace of the spontaneity given to the cloud or of the scheme of air movement feeding back into itself. At around the same age, the child attributes the movement of the stars also to the wind, (providing the school has not imposed the Copernican explanation). From the age of ten to eleven, the child understands that the movement of rivers is due to the weight of the water, which drags the river down, and not a mysterious ‘force’ or a ‘current’. It is also around the age of ten to eleven that the child explains the suspension of clouds and the floatation of boats by the respective density of the bodies.

So, if our working hypotheses are correct, the attaining of mechanical explanations, around the age of ten to eleven, is not due simply to the elimination of subjectivism and to a growing empiricism, but indeed to the development of reason that allows the child to shed both the subjective adherences of his thought and the empirical appearances of things. We have seen, indeed, that the child's causality starts with the maximum of both subjectivism and phenomenism. We have seen, on the other hand, by examining successively our four stages that the subjective adherences of causality disappear little by little, from personal efficacy to animism and artificialism, from animism to dynamism and from dynamism to mechanism. We need now to see if empiricism disappears in a similar way.

With respect to the above, the transition from the stage of moral explanation to the stage of dynamist causality already provides some lessons. It is during this transition that the child stops believing in the apparent movement of the stars that follow him, and formulates, due to the development of his reasoning, the hypothesis of an astronomical movement independent of us. The explanations of movement by the air feeding back on itself or of the suspension of clouds by gliding are also rational constructions succeeding the naïve empiricism of the previous stage. However it is clear that these constructions are still singularly close to empirical appearances. So it is the transition from dynamism to mechanism that will be the most instructive for the point of view that interests us now.

Two facts are particularly suggestive in this matter. Namely, there are two essential transformations in scientific reasoning of the child, and two transformations, which are contemporaneous with the advent of mechanical reasoning in childlike thinking.

The first of these facts is the development of the principles of conservation. As we have noted, the idea of weight plays an important role in the child's physics but in a very special way: the weight is the symbol of force and of a force that is almost alive. On the other hand, weight without the measurement of weight is something that depends singularly on appearances, and the child is not very exacting or precise in his estimations of the weight of things. A large hollow thing seems to him much lighter than he thought it would weigh. Demoor has created an interesting test based on this illusion of weight, and Sanctis has shown that, up to the age of eight, the weight appears to the child proportional to the volume. Finally, it is clear that, until really late, and certainly until the first experiments in physics and chemistry, the child will not have any concept of the exact measurement of weight. We ask ourselves whether he will spontaneously proceed as a pure empiricist that is, by confining himself to appearances only, or as a rationalist, that is, by suggesting principles of conservation, unverifiable by experience, but simplifying the interpretation of that experience?

In order to do this test, take two balls of clay or plasticine, and make these balls identical in weight and volume. We help the child to notice this identity as he lifts the balls and declares that indeed they have the same weight. Then, we leave one of the balls with the same shape but we roll the other in our hands until it has the shape of a long sausage. We then ask the child if both pieces have still the same weight, and to explain his answer.

Results of this experiment are really interesting, because children's explanations change a great deal in the course of our stages, and because, at each stage, the child is really sure of himself, to the point that he considers the answers from a previous stage as absurd. Thus the changes in the responses progress clearly from empiricism to rationalism.

Until the age of seven to eight, the child slavishly submits himself to salient appearances. The sausage appears to him lighter than the ball because it is thinner and more slender. Therefore the child claims that both things do not have the same weight any more and that the sausage is less heavy. We remind the child that previously the two pieces had the same weight. Indeed he remembers it well, but that fact does not seem to imply at all that the things have retained this weight. The child explains that the sausage has become ‘thinner’ and therefore, consequently, that it is ‘lighter’. Moreover, some children claim that the sausage has become ‘smaller’. In trying to understand these statements, we noticed that, for children of this age, the matter (the stuff of which the sausage is made) itself does not stay the same. We ask if these two things, which previously had exactly the same volume, have kept the same amount of clay or plasticine. The child of this stage claims no: the ball, for them, is ‘bigger’ than the sausage, and has more clay. Then, the child is asked where the disappeared clay has gone: he looks at you, taken aback by this bizarre question, but does not modify his answer.

During the dynamist stage, that is, from the age of seven to ten, things change slightly. The child understands the conservation of matter (stuff). He claims thus that there is as much clay in the sausage as in the ball, and, when one asks him why, he adds laughing: ‘But there is the same as before!’ Thus the principle of conservation seems obvious to him, and he laughs at the thought that one could have any doubt. One year before, maybe, he would have claimed the opposite. Furthermore, this conservation of matter (stuff) does not seem to him to carry over to the conservation of weight, and he continues, until about the age of ten on average, to claim that the sausage is less heavy than the ball!

At around the age of ten, on the contrary, the child claims that both weights have stayed the same, and he smiles at the question, it seeming so obvious to him that the weight is conserved.

These findings are interesting. Such conservation principles certainly mark great progress in the child's mind, although, according to physics, only mass is conserved, and not the weight; and mass itself, as one knows since Einstein, varies with acceleration … The progress that the child accomplishes is certainly linked with the development of the mechanical explanation, since it is the same effort of rationality, which brings the mind to deny, despite appearances, the spontaneous life of bodies and their qualitative changes to the benefit of a universe that is determined and quantifiable. So, how did the child acquire these principles of conservation? By experience? Certainly not: that would need very fine scales and an extraordinarily precocious care for precision. Rather, the child understands this law in the same manner as he comprehends a physical law such as: ice melts with heat. The child acquires an understanding of the principles of the physical world through reasoning and intelligence, not by empirical discovery. The only question is to know why this understanding appears so late? Child psychology teaches us that reason only develops little by little, with exposure to facts, but not under their sole pressure.

The second phenomenon about which we want to talk is the emergence of relative notions, that is, judgements about physical relation, in contrast to absolute categorical judgements. We can again take the concept of weight as an example.

When one asks the child why boats float, the explanation quickly invokes weight. However, as we have already seen, the child starts (before seven to eight) by claiming both that small boats float because they are light and that big boats float because they are heavy. This is because the child does not yet consider comparing the weight of the boats to the weight of water itself. For him, heaviness or lightness are absolute qualities that one either does or does not possess and that bring their effects by the mere fact of their presence: thus the big boats, being heavy, have the strength to support themselves, whereas the small ones, not being heavy, are supported by the water, which is strong. During the dynamist stage, and from the age of seven to eight the child begins to make comparisons. All boats float because they are light, even the big ones, where a big boat is lighter compared to all the water in a lake or in the sea. However on a pond, a steamboat would sink. This explanation is more or less mixed with explanations of gliding, which we talked about previously. It certainly marks a progress in the direction of relativity but, as we can see, it remains rudimentary, because it does not consider the fact that a stone sinks whereas a boat floats. The child is aware of this issue, and that is why he brings in gliding as an additional cause of flotation.

Around the age of ten and eleven, at about the time when the principle of weight conservation emerges, the child starts to conceive of weight as something that is relative to a given volume. He will say, for example, that a rowing boat is lighter than a stone, because it is empty, and with two things, wood and a stone of an equal volume, that the wood is lighter than the stone. Things are, in fact, quite complicated because before he arrives at this point, the child starts to admit that a body decreases in weight absolutely and not only relatively when its shape stretches (the quantity of matter staying the same). But the outcome of these reflections is that, at around the age of ten to eleven, the concept of density is acquired, that is, the weight of the floating thing is always estimated according to the corresponding volume of water that is displaced.

In short, the concept of weight starts as an absolute concept in the manner of Aristotle's absolute ‘up’ and ‘down’, then it transforms into a pure relation. Now it goes without saying that this relational logic is only appropriate to physical reality, as opposed to the verbal and conceptual realism that characterises the beginnings of this development. However it is clear that this elaboration of relations is not due to simple empirical verifications since the child's reasoning results in the construction of a quantitative and rational world that comes to replace the universe of superficial appearances. It is even more interesting to notice that this transformation happens at the time when the principle of weight conservation develops and also at the time when mechanical explanations supplant the anthropomorphical dynamism of the previous stages.

V. Conclusion

It seems that we have verified – from the purely psychological point of view that we have taken – the hypotheses that we have suggested while studying the beginnings of the child's concept of causality.

At the starting point, the self is entirely confused with things. The relations that establish themselves between the organism and its environment then give birth to a primitive form of causality in such a way that purely phenomenist and empirical links are necessarily accompanied by the child's feeling of efficacy of efforts and of personal gestures. Subsequently, the child liberates himself both from these subjective adherences and from superficial appearances. Therefore, on the one hand, causality proceeds from magic to animism and artificialism, from animism to dynamism and from dynamism to mechanism, with a concomitant elimination of the self to the benefit of objectivity. At the same time the crude empiricism of the beginning is succeeded by the construction of a world of relations such that the child discards explanations based on immediate experiences and embraces explanations that use rational principles of conservation coupled with a kind of implicit determinism built on quantitative comparison. In this respect, causality is much more the product of an intellectual construction than the immediate product of an internal or external experience.

One cannot help being struck by the parallelism between this development of a concept of causality, and the development of the concept of cause in the history of human societies. This is what we would like to underline in concluding this article. However, the greatest caution is imperative for two reasons.

Firstly, if there is truly some relation between young children's understanding of causality and the causality in the primitive mind, let us take care in believing, as did Stanley Hall, that this mind-set could have been transmitted to the contemporary civilised child, by the way of heredity. This hypothesis is, from a biological point of view, certainly very weak.

Secondly, we should be cautious in suggesting that if the child arrives at certain rational principles he acquires them entirely by himself. Most certainly, these principles have not been dictated to him. However, in the rational environment in which the civilised child lives, it is obvious that the child's advancement to the final stages, just described, is singularly facilitated by the continuous influence of the adult perspective.

Having made these reservations, it seems that we can nevertheless make some proposals. As we know, Durkheim and the sociologists of his school have accepted as true that causality owes its origin and development to society. It is the idea of social constraint that might have suggested to primitive people the belief in occult forces: the magic and animist beliefs in general. It would be the progress of societies that would then explain the elimination of these forms of anthropomorphic or socio-morphic causality, and the advent of a rationalist mechanism. So, as far as we are concerned, the comparison between the mental development of the child and the historic development of causality seems to speak much more in favour of the second of these theses than in favour of the first.

As far as the genesis of causality is concerned, it seems difficult to attribute to society the formation of this fundamental notion of the mind since the baby already constructs some magico-phenomenist relations when he starts imitating others. For us it seems inappropriate to deny any analogy between the social magic of the primitives and the child's beliefs in his personal efficacy, or between primitive and childlike participations. There is, whatever Durkheim said, a childlike animism. Also the childlike artificialism certainly assumes a society because it stemmed from the representations that children make of the adult's authority, but it also assumes a way of thinking proper to the child himself. In short, it is really difficult not to see in the primitive mind some traces of childlike mentality. As was well said by M. F. C. Bartlett, the fault of Durkheim's sociological school is not to have wrongly analysed the primitive, rather it fails in not giving sufficient scrutiny to the mind of the civilised, and above all, we will add, to the child's mind. With regard to causality in particular, primitive societies seem to have developed, consolidated and stylised the different forms of the childlike causality rather than to have created something absolutely new.

However if primitive causality is childlike in its origin, it appears to us undeniable that rational causality owes its blossoming to social influences. It is not enough to say along with the philosophers that causality is a natural product of the mind. It still needs to be shown under what conditions this form could have become aware of itself. So, in this respect, child psychology gives us a useful lesson: a child can think about things by freeing himself from his self. Also, sociologists teach us that societies become rational by freeing themselves of social pressures and obligatory conformism. These two theses, so different from one another, are less contradictory than it seems because, for the child, the egocentrism of thought and the submission to adult authority go curiously hand in hand. Thus, it is free discussion that creates reason: logical reflexion is only an interiorized discussion. Indeed, free discussion eliminates both the exterior authority and intellectual egocentrism. It is in this sense that, in as much as scientific causality is a product of reason, it is also a product of society.


The editor warmly thanks the following people who assisted in translation: Daniel Acquah, Joan Bliss, Sarah Loher and Leslie Smith. The editor takes full responsibility for the final translation.


  1. Conference given at Cambridge on the 4th of March 1927, under the auspices of Cambridge Education Society.