4       The roots of scientific reasoning: infancy, modularity and the art of tracking

 

            Peter Carruthers

 

 

This chapter examines the extent to which there are continuities between the cognitive processes and epistemic practices engaged in by human hunter–gatherers, on the one hand, and those which are distinctive of science, on the other. It deploys anthropological evidence against any form of ‘no-continuity’ view, drawing especially on the cognitive skills involved in the art of tracking. It also argues against the ‘child-as-scientist’ accounts put forward by some developmental psychologists, which imply that scientific thinking is present in early infancy and universal amongst humans who have sufficient time and resources to devote to it. In contrast, a modularist kind of ‘continuity’ account is proposed, according to which the innately channelled architecture of human cognition provides all the materials necessary for basic forms of scientific reasoning in older children and adults, needing only the appropriate sorts of external support, social context, and background beliefs and skills in order for science to begin its advance.

 

1          Introduction

It needs no emphasis that there has been a staggering and explosive increase in scientific knowledge, together with associated technological ability, over the last five centuries. But to what extent has this depended upon extrinsic cultural–economic factors, and to what extent upon intrinsic cognitive ones? Undoubtedly changes of both kinds have taken place, and have played a significant role. The invention of the printing press, and the existence of a class of moneyed gentlemen with time to devote to systematic scholarship and scientific enquiry were surely important; as were new inferential practices – both mathematical, and those distinctive of the experimental method. And without doubt changes of both kinds have continued to be important, too – had it not been for the development of new scientific instruments, and without the economic growth necessary for significant resources to be committed to scientific research, we would certainly not be in the epistemic position we are in today; but the development of statistical methods of reasoning, for example, have also been crucially significant.

 

1.1       The questions

The questions I want to address in this chapter are these: just how fundamental were the cognitive changes necessary for science to develop (and continue)? and what is the balance between the respective extrinsic and intrinsic factors underlying science? Although these questions are vague, they are important. To the degree that extrinsic factors dominate, to that extent we will be able to say that there is a fundamental continuity between the cognitive processes of scientists and those living in pre-scientific cultures. (I shall refer to this as a form of ‘continuity hypothesis’.) But to the extent that differences in cognitive processing are important, we may need to say that the cognition of pre-scientific peoples is radically distinct from our own. (Exaggerating a great deal, I shall call this the ‘no-continuity hypothesis’.)

            Just what sort of question is it that I am asking, however? Plainly, it is not intended to be a question about the causally necessary conditions for the development of science. For it may well be the case that both the printing press and beliefs about experimental method (say) were causally necessary factors in the scientific revolution. My question rather concerns the nature and extent of the cognitive changes necessary for science to begin. Continuity views maintain that any changes in mental functioning were relatively peripheral and minor; whereas no-continuity views hold that major cognitive re-structuring and a ‘re-programming’ of the mind were necessary.

            Put somewhat differently (and non-equivalently – see the following paragraph), my question concerns the innate basis of science and scientific reasoning.[1] According to a continuity account, most of the kinds of cognitive processing and reasoning which are necessary for science form part of the innate human cognitive endowment, needing only to be supplemented by some changes in belief or desire, and perhaps also by changes in external social-economic resources, in order for science to become possible. According to no-continuity accounts, in contrast, our innate cognitive endowment is by no means sufficient for scientific reasoning, and most of the necessary cognitive materials (beliefs and/or reasoning processes) need to be socially constructed and learned before science can start its advance.

            Notice that these two different ways of raising our question – was major re-programming of the mind necessary for science to begin? how close does our innate endowment come to being sufficient for science? – are not strictly equivalent. For someone might claim that our innate endowment falls a long way short of what is needed for science, and yet claim that there is a form of culturally-transmitted, but universal and trans-historical, cognition which is close to being sufficient. This would then be a form of continuity thesis without any commitment to the innateness of our scientific abilities. But actually the thesis which interests me, and which I shall defend in this paper, is that our innate endowment does come pretty close to being sufficient for science, and it is because of this that only relatively minor cognitive changes were necessary for science to originate.

            Everyone should allow, of course, that the innate cognitive basis of the mind (whatever it is) is held in common between scientists and hunter–gatherers; for we are all now members of the same contemporary species. (And it is, in any case, very implausible to suppose that substantive genetic change might have taken place – and in parallel, on a global scale – in the mere 10,000 years or so since the beginning of the agricultural revolution.) So it is obvious that the differences between scientists and hunter–gatherers are not innate ones. As a result, my question certainly isn’t whether some change in the human genotype explains the genesis and rise of science in recent centuries. That would be absurd. Rather (to repeat), I am asking just how close our innate cognitive endowment comes to being sufficient for scientific reasoning. My sort of continuity account maintains that it comes pretty close - relatively few changes of belief and/or social circumstances were necessary for science to begin. No-continuity accounts, in contrast, claim that our innate endowment falls a long way short of being sufficient for science, and that they needed to be radically extended or re-programmed.

            Equally, everyone should allow that there are immense cognitive differences between a contemporary scientist (or even a scientist of the sixteenth century) and a hunter–gatherer. The differences in their belief-systems will be vast; and no doubt some of the sorts of reasoning process in which they engage will differ also. (After all, it requires training to become a scientist, and not all of this training is of a purely practical, visuomotor, variety – some is mathematical, and some is methodological.) But it still might be possible to claim that their cognitive processes are essentially similar, if the differences are (almost) all differences of belief, and if (almost) all of the basic (innate) sorts of general cognitive processing in which they engage are both shared and sufficient (or almost sufficient) for science. My sort of continuity theory maintains that this is the case, whereas no-continuity accounts deny it.

 

1.2       The four options

On one view, the innate basis of the mind is mostly domain-general in nature, having to do with general capacities for learning and/or reasoning, though perhaps containing some initial domain-specific information and/or attention-biases (Elman et al., 1996; Gopnik and Melzoff, 1997). On a contrasting view, much of the innate structure of the mind is domain-specific, embodying information about evolutionarily-significant domains, and/or containing learning-principles specific to particular domains (Barkow et al., 1992; Pinker, 1997).

            The domain-general account of the innate basis of cognition is one or another version of the general-purpose computer model of the mind. In some versions (e.g. Dennett, 1991, 1995) what is given are a suite of massively parallel and distributed processors which nevertheless have the power to support a serial, language-involving, digital processor running on linguistic structures. (Dennett dubs this the ‘Joycean machine’ after the stream-of-consciousness writing of James Joyce’s Ulysses.) This latter system is almost entirely programmed by enculturated language-use, acquiring both its contents and general patterns of processing through the acquisition of both information and habits of thought from other people, via linguistic communication and language-based instruction and imitation. On this view, the basic cognitive differences between ourselves and hunter–gatherers will be very large; and there will be a great deal of cognitive–linguistic programming required before the human mind becomes capable of anything remotely resembling scientific reasoning.

            Quite a different sort of domain-general view is entailed by ‘theorising theory’ accounts of the nature of human development through infancy and childhood (e.g. Gopnik and Melzoff, 1997; Gopnik and Glymour, this volume).[2] On this view, all human children are already little scientists, in advance of any exposure to scientific cultures – gathering data, framing hypotheses, and altering their theories in the light of recalcitrant data in essentially the same sort of way that scientists do. So on this view, the cognitive continuities between scientific and pre-scientific cultures will be very great, and almost all the emphasis in an explanation of the rise of science over the last five hundred years will have to be on extrinsic factors. On Gopnik and Melzoff’s account, most adult humans (including hunter–gatherers) are scientists who have ceased to exercise their capacity for science, largely through lack of time and attention. But this account can at the same time emphasise the need for extrinsic support for scientific cognition (particularly that provided by written language, especially after the invention of the printing press) once theories achieve a certain level of complexity in relation to the data.

            Domain-specific, more-or-less modular, views of cognition also admit of a similar divide between no-continuity and continuity accounts of science. On one approach, the modular structure of our cognition which those of us in scientific societies share with hunter–gatherers is by no means sufficient to underpin science, even when supported by the appropriate extrinsic factors. Rather, that structure needs to be heavily supplemented by culturally-developed and culturally-transmitted beliefs and reasoning practices. In effect, this account can share with Dennett the view that a great deal of the organisation inherent in the scientific mind is culturally acquired – differing only in the amount of innate underlying modular structure which is postulated, and in its answer to the question whether intra-modular processing is connectionist in nature (as Dennett, 1991, seems to believe), or whether it rather involves classical transformations of sentence-like structures (as most modularists and evolutionary psychologists think: see, e.g., Fodor, 1983, 2000; Tooby and Cosmides, 1992; Pinker, 1997).[3]

            Alternatively, a modularist may emphasise the cognitive continuities between scientific and pre-scientific cultures, sharing with the theorising theory the view that the basic ingredients of scientific reasoning are present as part of the normal human cognitive endowment. However, such an account will also differ significantly from theorising theory by postulating a domain-specific cognitive architecture, and in claiming that capacities for scientific reasoning are not a cognitive given at the start of development, but rather emerge at some point along its normal course. This is the view which I shall ultimately be proposing to defend.

            The chapter will proceed as follows. In section 2 I shall argue against either form of no-continuity account of science. That is to say, I shall argue against the claim that the human mind needs to be radically re-programmed through immersion in a suitable linguistic-cultural environment in order for science to become possible, whether this programming takes place within a broadly modularist, or rather in a general-learning, framework. Then in section 3 I shall argue against the sort of domain-general continuity account espoused by those who adopt a theorising-theory view of normal human cognitive development. Finally, in section 4 I shall sketch my favoured alternative, which is a modularist form of continuity account. But in the limited space available I shall be able do little more than comment on the main components of the story.

 

2          Against no-continuity views

According to no-continuity accounts, the human mind needs to be radically re-programmed by immersion in an appropriate language-community and culture - acquiring a whole suite of new cognitive processes - in order for anything resembling science to become possible (Dennett, 1991). But against this view it can be argued that hunter–gatherers actually engage in extended processes of reasoning – both public and private – which look a lot like science. Developing this point will require us to say something about the cognitive processes which are actually characteristic of science, however.

 

2.1       Scientific reasoning

On one view, the goal of science is to discover the causal laws which govern the natural world; and the essential activity of scientists consists in the postulation and testing of theories, and then applying those theories to the phenomena in question (Nagel, 1961; Hempel, 1966). On a contrasting view, science constructs and elaborates a set of models of a range of phenomena in the natural world, and then attempts the develop and apply those models with increasing accuracy (Cartright, 1983; Giere, 1992). But either way science generates principles which are nomic, in the sense of characterising how things have to happen, and in supporting subjunctives and counterfactuals about what would happen, or would not have happened, if certain other things were to happen, or hadn’t happened.

            Crucial to the activity of science, then, is the provision of theories and/or models to explain the events, processes, and regularities observed in nature. Often these explanations are couched in terms of underlying mechanisms which have not been observed and may be difficult to observe; and sometimes they are given in terms of mechanisms which are unobservable. More generally, a scientific explanation will usually postulate entities and/or properties which are not manifest in the data being explained, and which may be unfamiliar – where perhaps the only reason for believing in those things is that if they did exist, then they would explain what needs explaining.

            Science also employs a set of tacit principles for choosing between competing theories or models – that is, for making an inference to the best explanation of the data to be explained. The most plausible way of picturing this, is that contained within the principles employed for good explanation are enough constraints to allow one to rank more than one explanation in terms of goodness. While no one any longer thinks that it is possible to codify these principles, it is generally agreed that the good-making features of a theory include such features as; accuracy (predicting all or most of the data to be explained, and explaining away the rest); simplicity (being expressible as economically as possible, with the fewest commitments to distinct kinds of fact and process); consistency (internal to the theory or model); coherence (with surrounding beliefs and theories, meshing together with those surroundings, or at least being consistent with them); fruitfulness (making new predictions and suggesting new lines of enquiry); and explanatory scope (unifying together a diverse range of data).

 

2.2       Hunter–gatherer science

Is there any evidence that hunter–gatherer communities engage in activities which resemble science? It is a now familiar and well-established fact that hunter–gatherers have an immense and sophisticated (if relatively superficial) understanding of the natural world around them. They have extensive knowledge of the plant and animal species in their environments – their kinds, life-cycles, and characteristic behaviours – which goes well beyond what is necessary for survival (Atran, 1990; Mithen, 1990, 1996). But it might be claimed that the cognitive basis for acquiring this sort of knowledge is mere inductive generalisation from observed facts. It might appear that hunter–gatherers don’t really have to engage in anything like genuine theorising or model-building in order to gain such knowledge. Nor (except in their magic) do they seem, on the face of it, to rely on inferences concerning the unobserved. On the contrary, it can easily appear as if mere careful observation of the environment, combined with enumerative induction, is sufficient to explain everything that they know.[4]

            In fact this appearance is deceptive, and at least some of the knowledge possessed by hunter–gatherers concerns facts which they have not directly observed, but which they know by means of inference to the best explanation of signs which they can see and interpret (Mithen, this volume). For example, the !Xõ hunter–gatherers of the Kalahari are able to understand some of the nocturnal calls of jackals as a result of studying their spoor the next day and deducing their likely activities; and they have extensive knowledge of the lives of nocturnal animals derived from study of their tracks, some of which has only recently been confirmed by orthodox science (Liebenberg, 1990). But it is the reasoning in which hunters will engage when tracking an animal which displays the clearest parallels with reasoning in science, as Liebenberg (1990) argues at length in his wonderful but little-noticed study in anthropology and philosophy of science. (Anyone who was ever tempted to think that hunter–gatherers must be cognitively less sophisticated than ourselves should read this book.)

 

2.3       The art of tracking

It is true, but by no means obvious at first glance, that tracking will always have played a vital role in most human hunter–gatherer communities. This is not especially because tracking is necessary to locate a quarry. For while this is important in many contexts and for many types of game, it is not nearly so significant when hunting herd animals such as wildebeest. It is rather because, until the invention of the modern rifle, it would always have been rare for a hunter to bring an animal down immediately with an arrow or a spear. (And while hunting in large groups might have made it more likely that the target animal could be brought down under a volley of missiles, it would have made it much less likely that the hunters would ever have got close enough to launch them in the first place.)

            In consequence, much of the skill involved in hunting consists in tracking a wounded animal, sometimes for a period of days. (Even the very simplest form of hunting – namely, running an animal down – requires tracking.[5] For almost all kinds of prey animal can swiftly sprint out of easy sight, except in the most open country, and need to be tracked rapidly by a runner before they have the opportunity to rest.) For example, the !Xõ will generally hunt in groups of between two and four, using barbed arrows which have been treated with a poison obtained from the larvae of a particular species of beetle. An initial shot will rarely prove immediately fatal, and the poison can take between 6 and 48 hours to take effect, depending on the nature of the wound and the size of the animal. So a wounded animal may need to be tracked for considerable periods of time before it can be killed.

            As Liebenberg (1990) remarks, it is difficult for a city-dweller to appreciate the subtlety of the signs which can be seen and interpreted by an experienced tracker. Except in ideal conditions (e.g. firm sand or a thin layer of soft snow) a mere capacity to recognise and follow an animal’s spoor will be by no means sufficient to find it. Rather, a tracker will need to draw inferences from the precise manner in which a pebble has been disturbed, say, or from the way a blade of grass has been bent or broken; and in doing so he will have to utilise his knowledge of the anatomy and detailed behaviours and patterns of movement of a wide variety of animals.[6] Moreover, in particularly difficult and stony conditions (or in order to save time during a pursuit) trackers will need to draw on all their background knowledge of the circumstances, the geography of the area, and the normal behaviour and likely needs of the animal in question to make educated guesses concerning its likely path of travel.

            Most strikingly for our purposes, successful hunters will often need to develop speculative hypotheses concerning the likely causes of the few signs available to them, and concerning the likely future behaviour of the animal; and these hypotheses are subjected to extensive debate and further empirical testing by the hunters concerned. When examined in detail these activities look a great deal like science, as Liebenberg (1990) argues. First, there is the invention of one or more hypotheses (often requiring considerable imagination) concerning the unobserved (and now unobservable) causes of the observed signs, and the circumstances in which they may have been made. These hypotheses are then examined and discussed for their accuracy, coherence with background knowledge, and explanatory and predictive power.[7] One of them may emerge out of this debate as the most plausible, and this can then be acted upon by the hunters, while at the same time searching for further signs which might confirm or count against it. In the course of a single hunt one can see the birth, development, and death of a number of different ‘research programmes’ in a manner which is at least partly reminiscent of theory-change in science (Lakatos, 1970).

 

2.3       Tracking: art or science?

How powerful are these analogies between the cognitive processes involved in tracking, on the one hand, and those which underlie science, on the other? First of all we should note one very significant disanalogy between the two. This is that the primary goal of tracking is not an understanding of some general set of processes or mechanisms in nature, but rather the killing and eating of a particular animal. And although knowledge and understanding may be sought in pursuit of this goal, it is knowledge of the past and future movements of a particular prey animal, and an understanding of the causal mechanisms which produced a particular set of natural signs which is sought, in the first instance. (Of course the hunters may also hope to obtain knowledge which will be of relevance to future hunts.)

            This disanalogy is sufficient, in my view, to undermine any claim to the effect that tracking is a science. Since it doesn’t share the same universal and epistemic aims of science, it shouldn’t be classed as one. But although it isn’t a science, it is perfectly possible that the cognitive processes which are involved in tracking and in science are broadly speaking the same. So we can still claim that the basic cognitive processes involved in each are roughly identical, albeit deployed in the service of different kinds of end. Indeed, it is just such a claim which is supported by the anthropological data.

            First of all, it is plain that tracking, like science, frequently involves inferences from the observed to the unobserved, and often from the observed to the unobservable as well. Thus a tracker may draw inferences concerning the effects which a certain sort of movement by a particular kind of animal would have on a certain kind of terrain (which may never actually have been observed previously, and may be practically unobservable to hunter–gatherers if the animal in question is nocturnal). Or a tracker may draw inferences concerning the movements of a particular animal (namely the one he had previously shot and is now tracking) which are now unobservable even in principle, since those movements are in the past. Compare the way in which a scientist may draw inferences concerning the nature of a previously unobserved and practically unobservable entity (e.g. the structure of DNA molecules before the invention of the electron microscope), or concerning the nature of particles which are too small to be observable, by means of an inference to the best explanation from a set of broadly-observational data.

            Second, it is plausible that the patterns of inference engaged in within the two domains of tracking and science are isomorphic with one another. In each case inferences to the best explanation of the observed data will be made, where the investigators are looking for explanations which will be simple, consistent, explanatory of the observational data, coherent with background beliefs, maximal in explanatory scope (relevant to the aims of the enquiry, at least), as well as fruitful in guiding future patterns of investigation. And in each case, too, imaginative–creative thinking has a significant – nay, crucial – role to play in the generation of novel hypotheses.

            Let me say something more about the role of creative thinking in tracking and in science. It is now widely accepted that inductivist methodologies have only a limited part to play in science. Noticing and generalising from observed patterns cannot carry you very far in the understanding of nature. Rather, scientists need to propose hypotheses (whether involving theories or models) concerning the underlying processes which produce those patterns. And generating such hypotheses cannot be routinised. Rather, it will involve the imaginative postulation of a possible mechanism – guided and constrained by background knowledge, perhaps, but not determined by it. Similarly, a hunter may face a range of novel observational data which need interpreting. He has to propose a hypothesis concerning the likely causes of that data – where again, the hypothesis will be guided and constrained by knowledge of the circumstances, the season, the behaviour of different species of animal and so on, but is not by any means determined by such knowledge. Rather, generating such hypotheses requires creative imagination, just as in science.

            These various continuities between tracking and science seem to me sufficient to warrant the following claim: that anyone having a capacity for sophisticated tracking will also have the basic cognitive wherewithal to engage in science. The differences will be merely these - differences in overall aim (to understand the world in general, as opposed to the movements of a given animal in particular); differences of belief (including methodological beliefs about appropriate experimental methods, say); as well as some relatively trivial differences in inferential practices (such as some of the dispositions involved in doing long-division sums, or in solving differential equations). In which case some version of the continuity hypothesis can be regarded as established. We can assert that the cognitive processes of hunter–gatherers and modern scientists are broadly continuous with one another, and that what was required for the initiation of the scientific revolution were mostly extrinsic changes, or changes relating to peripheral (non-basic) aspects of cognition, such as mere changes of belief.

            It remains to enquire, though, roughly when and how the capacity for scientific reasoning emerges in the course of human development, and how it relates to other aspects of our cognition. In the next section I shall consider and criticise theorising-theory accounts of development, according to which scientific abilities are innate and put to work very shortly after birth. Then in section 4 I shall sketch a view which also allows that scientific abilities are innately channelled, but which sees their development as dependent upon language – perhaps only emerging in later childhood or early adolescence (I shall remain open as to the exact timing).

 

3          Against the ‘theorising theory’ continuity view

The orthodox position in cognitive science over the last two decades has been that many of our cognitive capacities – including the capacity to attribute mental states to oneself and others (i.e. the capacity to ‘mind-read’), and the capacity to explain and predict many of the behaviours of middle-sized physical objects – are subserved by bodies of belief or knowledge, constituting so-called ‘folk theories’ of the domains in question. But a number of developmental psychologists have gone further in suggesting that these theories are developed in childhood through a process of theorising analogous to scientific theorising (e.g. Carey, 1985; Wellman, 1990). This theorising-theory approach has been developed and defended most explicitly by Gopnik and Melzoff (1997), and it is on their account that I shall focus my critique.

 

3.1       The absence of external supports

In my view, the main objection to a theorising-theory account is that it ignores the extent to which scientific activity needs to be supported by external resources.[8] Scientists do not, and never have worked alone, but constantly engage in discussion, co-operation and mutual criticism with peers. If there is one thing which we have learned over the last thirty years of historically-oriented studies of science, it is that the positivist–empiricist image of the lone investigator, gathering all data and constructing and testing hypotheses by him- or her-self, is a highly misleading abstraction.

Scientists such as Galileo and Newton engaged in extensive correspondence and discussion with other investigators at the time when they were developing their theories; and scientists in the 20th century, of course, have generally worked as members of research teams. Moreover, scientists cannot operate without the external prop of the written word (including written records of data, annotated diagrams and graphs, written calculations, written accounts of reasoning, and so on). Why should it be so different in childhood, if the cognitive processes involved are essentially the same?

            I should emphasise that this point doesn’t at all depend upon any sort of ‘social constructivist’ account of science, of the sort that Gopnik and Melzoff find so rebarbative (and rightly so, in my view). It is highly controversial that scientific change is to be explained, to any significant extent, by the operation of wider social and political forces, in the way that the social constructivists claim (Bloor, 1976; Rorty, 1979; Latour and Woolgar, 1986; Shapin, 1994). But it is, now, utterly truistic that science is a social process in at least the minimal sense that it progresses through the varied social interactions of scientists themselves – co-operating, communicating, criticising – and through their reliance on a variety of external socially-provided props and aids, such as books, paper, writing instruments, a variety of different kinds of scientific apparatus, and (now) calculators, computers and the internet (Giere, this volume). And this truism is sufficient to cause a real problem for the ‘child as scientist’ account of development.

            One sort of response would be to claim that children do not need external props, because they have vastly better memories than adult scientists. But this is simply not credible, of course; for it is not true that children’s event-memories are better than those of adults.[9] (Nor do young children have any memory for their own previous mental processes. And in particular, lacking any concept of belief as a representational state, children under four cannot remember their own previous false beliefs – Gopnik, 1993.)[10] And in any case it isn’t true that science depends upon external factors only because of limitations of memory. On the contrary, individual limitations of rationality, insight and creativity all play an equally important part. For scientific discussion is often needed to point out the fallacies in an individual scientist’s thinking; to show how well-known data can be explained by a familiar theory in ways that the originators hadn’t realised; and to generate new theoretical ideas and proposals.[11]

            Nor can these differences between children and adults plausibly be explained in terms of differences of attention, motivation, or time, in the way that Gopnik and Melzoff try to argue. For adult scientists certainly attend very closely to the relevant phenomena, they may be highly motivated to succeed in developing a successful theory, and they may be able to devote themselves full-time to doing so. But still they cannot manage without a whole variety of external resources, both social and non-social. And still radical (conceptually-innovative) theory change in science (of the sort which Gopnik and Melzoff acknowledge occurs a number of times over within the first few years of a child’s life) is generally spread out over a very lengthy time-scale (often as much as a hundred years or more; Nersessian, 1992 and this volume).

 

3.2       The extent of the data

According to Gopnik and Melzoff (1997), the main difference between childhood theorisers and adult scientists lies in the extent and ease of availability of relevant data. In scientific enquiry, the relevant data are often hard to come by, and elaborate and expensive experiments and other information-gathering exercises may be needed to get it; perforce making scientific enquiry essentially social. But in childhood there are ample quantities of data easily available. Young children have plenty of opportunities to experiment with physical substances and their properties – knocking objects together, dropping or throwing them, pouring liquids, and mixing materials – when developing their naïve physics; and they have plenty of opportunity to observe and probe other agents when developing their naïve psychology as well. Moreover, since infants come into the world with a set of innate domain-specific theories, on Gopnik and Melzoff’s account, they already possess theoretical frameworks which constrain possible hypotheses, determine relevant evidence, and so on.

            This point, although valid as far as it goes, does not begin to address the real issue. If anything, the extent of the data available to the child is a further problem for the ‘child as scientist’ view, given the lack of external aids to memory, and the lack of any public process for sorting through and discussing the significance of the evidence. That plenty of data is available to an enquirer is irrelevant, unless that data can be recalled, organised and surveyed at the moment of need – namely, during theory-testing or theory-development.

            Gopnik and Melzoff also make the point that much relevant data is actually presented to children by adults, in the form of linguistic utterances of one sort or another. Now, their idea is not that adults teach the theories in question to children, thus putting the latter into the position of little science students rather than little scientists. For such a claim would be highly implausible – there is no real evidence that any such teaching actually takes place (and quite a bit of evidence that it does not). Their point is rather that adults make a range of new sorts of evidence available to the child in the form of their linguistic utterances, since those utterances are embedded in semantic frameworks which contain the target theories. Adult utterances may then provide crucial data for children, as they simultaneously elaborate their theories and struggle to learn the language of their parents.

            This proposal does at least have the virtue of providing a social dimension to childhood development, hence in one sense narrowing the gap between children and adult scientists. But actually this social process is quite unlike any that an adult scientist will normally engage in, since scientists (as opposed to science students) are rarely in the position of hearing or engaging in discussions with those who have already mastered the theories which they themselves are still trying to develop. And in any case the proposal does nothing to address the fundamental problems of insufficient memory, bounded rationality, and limited creativity which children and adults both face, and which the social and technological dimensions of science are largely designed to overcome.

 

3.3       Simple theories?

Might it be that the scientific problems facing adults are very much more complex than those facing young children, and that this is the reason why children, but not adult scientists, can operate without much in the way of external support? This suggestion is hardly very plausible, either. Consider folk psychology, for example, of the sort attained by most normal four or five year-old children. This has a deep structure rivalling that of many scientific theories, involving the postulation of a range of different kinds of causally-effective internal states, together with a set of nomic principles (or ‘laws’) concerning the complex patterns of causation in which those states figure. There is no reason at all to think that this theory should be easily arrived at. Indeed, it is a theory which many adult scientific psychologists have denied (especially around the middle part of the 20th century) – viz. those who were behaviourists. It took a variety of sophisticated arguments, and the provision of a range of different kinds of data, to convince most people that behaviourism should be rejected in favour of cognitivism.

            So why is it that no children (excepting perhaps those who are autistic) ever pass through a stage in which they endorse some form of behaviourism? If cognitivism were really such an easy theory to frame and establish, then the puzzle would be that adult scientific psychologists had such difficulty in converging on it. And given that it is not so easy, the puzzle is that all normal children do converge on it in the first four or five years of development (at least if they are supposed to get there by means of processes structurally similar to those which operate in science).[12] The natural conclusion to draw, at this point, is that folk psychology is both theoretically difficult and not arrived at by a process of theorising, but rather through some sort of modular and innately channelled development.

            I conclude this section, then, by claiming that the child-as-scientist account is highly implausible. Although this account (contra the no-continuity views discussed in section 2) at least has the virtue of emphasising the continuities between childhood development, hunter–gatherer thinking and reasoning, and adult scientific cognition, it suffers from a number of fatal weaknesses. Chief amongst these is that it ignores the extent to which scientific enquiry is a social and externally-supported process.

 

4          A modularist continuity view

The anthropological data reviewed in section 2 give us reason to think that some version of the continuity thesis is correct – that is, for thinking that the basic cognitive processes of ourselves and hunter–gatherers are fundamentally the same. But now, in section 3, we have seen that there is good reason to reject the domain-general form of continuity thesis defended by theorising-theorists such as Gopnik and Melzoff. So it looks as if a domain-specific form of continuity view must be established by default. For this is the only alternative now remaining. Plainly, however, we cannot let matters rest here. We need to show how a modularist form of continuity account can be true, and how it can explain the data (both anthropological and developmental). All I can really hope to do here, however, is assemble the various materials which should go into a properly developed form of modularist continuity-view of science. I cannot at this stage pretend to be offering such an account in any detail.

 

4.1       Developing modules

The form of continuity view which I favour helps itself to a version of what might be called ‘central-process modularism’ (as distinct from Fodor’s ‘peripheral-process modularism’; 1983). On this account, besides a variety of input and output modules (including early vision, face-recognition, and language, for example), the mind also contains a number of innately channelled conceptual modules, designed to process conceptual information concerning particular domains. While these systems might not be modular in Fodor’s classic sense – they would not have proprietary inputs, and might not be fully encapsulated, for example – they would conform to at least some of the main elements of Fodorian modularity. They would be innate (in some sense or other) and subject to characteristic patterns of breakdown; their operations might be mandatory and relatively fast; and they would process information relating to their distinctive domains according to their own specific algorithms.

            Plausible candidates for such conceptual modules might include a naive physics system (Leslie, 1994; Spelke, 1994; Spelke et al., 1995; Baillargeon, 1995), a naive psychology or ‘mind-reading’ system (Carey, 1985; Leslie, 1994; Baron-Cohen, 1995), a folk-biology system (Atran, 1990, 1999 and this volume), an intuitive number system (Wynn, 1990, 1995; Gallistel and Gelman, 1992; Dehaerne, 1997), and a system for processing and keeping track of social contracts (Cosmides and Tooby, 1992). And evidence supporting the existence of at least the first two of these systems (folk-physics and folk-psychology) is now pretty robust. Very young infants already have a set of expectations concerning the behaviours and movements of physical objects, and their understanding of this form of causality develops very rapidly over the first year or two of life. And folk-psychological concepts and expectations also develop very early, and follow a characteristic developmental profile. Indeed, recent evidence from the study of twins suggests that three-quarters of the variance in mind-reading abilities amongst three year olds is both genetic in origin and largely independent of the genes responsible for verbal intelligence, with only one quarter being contributed by the environment (Hughes and Plomin, 2000).

            Some of the disputes which are internal to central-process modularism – in particular, concerning which conceptual modules are or are not developmentally primitive – are highly relevant to our present concerns. Thus Carey (1985) thinks that children’s folk biology is built from an initial given folk psychology by a process of theorising, whereas Atran (1990 and this volume) conceives of our folk-biology system as a distinct module in its own right. If Carey is right, then plainly quasi-scientific theorising abilities will need to be in place through the early school years, since children’s grasp of the main principles of biological life-cycles, for example, is well established by about the age of seven. Whereas if folk biology emerges at about this time through a process of maturation rather than of theorising (albeit maturation in the context of a set of normal as well as environment-specific social inputs), then scientific abilities might not emerge until much later. I propose simply to set this dispute to one side for present purposes, on the simple but compelling grounds that I have nothing to contribute to it.

            One important point to notice in this context is that many of these conceptual modules provide us with deep theories of the domains which they concern. Thus the folk-physics system involves a commitment to a number of unobservable forces or principles, such as the impetus of moving objects, or the solidity of impenetrable ones. And the folk-psychology module involves commitment to a variety of inner psychological states, together with the causal principles governing their production, interaction, and effects on behaviour. These modules might therefore very plausibly be thought to provide us with a set of initial theoretical contents, on which our explanatory and abductive abilities can later be honed.

 

4.2       Cross-modular thinking and language

There is some tentative reason to think that language may have an important role to play in the genesis of scientific reasoning. For notice that the reasoning processes involved in tracking an animal will often require bringing together information from a variety of different intuitive domains, including folk-physics, folk-biology and folk-psychology. For example, in interpreting a particular scuff-mark in the dust a hunter may have to bring information concerning the anatomy and behaviours of a number of different animal species (folk-biology) together with inferences concerning the physical effects of certain kinds of impact between hoof and ground, say (folk-physics). And in predicting what an animal will do in given circumstances a hunter will rely, in part, on his folk-psychology – reasoning out what someone with a given set of needs and attitudes would be likely to do in those circumstances (Liebenberg, 1990).

            On one sort of view of the role of natural language in cognition, then – the view, namely, that language serves as the link between a number of distinct cognitive systems or modules (Mithen, 1996; Carruthers, 1996a, 1998; Hermer-Vazquez et al., 1999) – speculative intelligent tracking would have depended on the evolution of the language-faculty. So it may be that hunting-by-tracking only fully entered the repertoire of the hominid lineage some 100,000 years ago, with the first appearance of anatomically modern, language-using, humans in Southern Africa, as Liebenberg (1990) suggests. And it may then be that the cognitive adaptations necessary to support scientific thinking and reasoning were selected for precisely because of their important role in hunting.[13] But it might also be that these abilities were selected for on other grounds, only later finding application in tracking and hunting. It would go well beyond the scope I have set myself in this chapter to try to resolve this issue here. For present purposes the point is just that sophisticated cross-modular abductive reasoning may crucially implicate the language faculty (as also might reasoning which extends or corrects our naïve modular theories, as frequently happens in science).

 

4.3       Developing imagination

In addition, a capacity for imaginative or creative thinking is vital in both tracking and science, as we have seen. Novel explanatory hypotheses need to be generated in both domains – hypotheses which may be unprecedented in previous experience, and may go well beyond the observational data. There is no doubt that such a capacity is a distinctive element in the normal human cognitive endowment. And on one natural view, it is the developmental function of the sort of pretend play which is such a striking and ubiquitous part of human childhood to enhance its growth, through practice in suppositional or pretend thinking (Carruthers, 1996b, 1998, and forthcoming).

            What cognitive changes were necessary for imaginative thinking and/or pretend play to make its appearance in the hominid lineage? Nichols and Stich (2000) argue that it required the creation of a whole new type of propositional attitude, which they represent as a separate ‘box’ in the standard flow-chart of cognition involving boxes for belief, desire, practical reasoning, theoretical inference, and so on – namely a ‘possible worlds box’. And this might be taken to suggest that it would have required some powerful selectional pressure in order for imaginative thinking to become possible. But actually there is some reason to think that imagination comes in at least two forms, each of which would be provided ‘for free’ by the evolution of other faculties. Let me briefly elaborate.

            First, there is experiential imagination – namely, the capacity to form and to manipulate images relating to a given sense modality (visual, auditory, etc.). There is some reason to think that a basic capacity for this sort of imagination is a by-product of the conceptualising processes inherent in the various perceptual input-systems (Kosslyn, 1994). There are extensive feed-back neural pathways in the visual system, for example, which are used in object-recognition when ‘asking questions of’ ambiguous or degraded input. And these very pathways are then deployed in visual imagination so as to generate quasi-perceptual inputs to the visual system. Evidence from cognitive archaeology (concerning the imposition of sophisticated symmetries on stone tools, which would have required a capacity to visualise and manipulate an image of the desired product) suggests that this capacity would have been present at about 400,000 years before the present (Wynn, 2000) – i.e. considerably before the evolution of full-blown language, if the latter only appeared some 100,000 years ago, as many believe.

            Second, there is propositional imagination – the capacity to form and consider a propositional representation without commitment to its truth or desirability. There is some reason to think that this capacity comes to us ‘for free’ with language. For a productive language system will involve a capacity to construct new sentences, whose contents are as yet neither believed nor desired, which can then serve as objects of reflective consideration of various sorts. In which case a capacity for propositional imagination is likely to have formed a part of the normal human cognitive endowment for about the last 100,000 years.[14]

            Of course a mere capacity for creative generation of new sentences (or images) will not be sufficient for imaginative thinking as we normally understand it. For there is nothing especially imaginative about generating any-old new sentence or image. Rather, we think imagination consists in the generation of relevant and/or fruitful and interesting new ideas. And such a capacity will not come to us ‘for free’ with anything. But then this may be precisely the developmental function of pretend play, if through frequent practice such play helps to build a consistent disposition to generate novel suppositions which will be both relevant and interesting (Carruthers, forthcoming).

            Both forms of imagination play a role in science (Giere, 1992, 1996). And both forms are important in tracking as well. Thus by visualising the distinctive gait of a wildebeest a tracker may come upon the hypothesis that some faint marks in the dust on a windy day were produced by that animal. And by recalling that a zebra was seen in the vicinity the previous afternoon, he may frame the alternative hypothesis that the marks were produced by that animal instead.

 

4.4       Developing abductive principles

How are the principles involved in inference to the best explanation (or ‘abductive inference’) acquired? Are they in some sense innate (emerging without learning in normal humans in the course of development in normal environments)? Or are they learned? Plainly, the mere fact that such principles are employed by adult humans in both hunter–gatherer and scientific societies is not sufficient to establish their innateness. For it may be that broadly-scientific patterns of thinking and reasoning were an early cultural invention, passed on through the generations by imitation and overt teaching, and surviving in almost all extant human societies because of their evident utility.

            In fact, however, it is hard to see how the principles of inference to the best explanation could be other than substantially innate (Carruthers, 1992). For they are certainly not explicitly taught, at least in hunter–gatherer societies.[15] While nascent trackers may acquire much of their background knowledge of animals and animal behaviour by hearsay from adults and peers, very little overt teaching of tracking itself takes place. Rather, young boys will practice their observational and reasoning skills for themselves, first by following and interpreting the tracks of insects, lizards, small rodents, and birds around the vicinity of the camp-site, and then in tracking and catching small animals for the pot (Liebenberg, 1990). They will, it is true, have many opportunities to listen to accounts of hunts undertaken by the adult members of the group, since these are often reported in detail around the camp fire. So there are, in principle, opportunities for learning by imitation. But in fact, without at least a reasonably secure grasp of the principles of inference to the best explanation, it is hard to see how such stories could even be so much as understood. For of course those principles are never explicitly articulated; yet they will be needed to make sense of the decisions reported in the stories; and any attempt to uncover them by inference would need to rely upon the very principles of abductive inference which are in question.

            The question of how a set of abductive principles came to be part of the normal human cognitive endowment is entirely moot, at this point. One possibility is that they were directly selected for because of their crucial role in one or more fitness-enhancing activities, such as hunting and tracking. It may be, for example, that successful hunting (and hence tracking) assumed an increasing importance for our ancestors, either during the transition to anatomically modern humans some 100,000 years ago or shortly thereafter. Note, however, that evidence from contemporary hunter–gatherers suggests that the selective force at work would have been sexual rather than ‘natural’. For almost all hunter–gatherer tribes which have been studied by anthropologists are highly egalitarian in organisation, with meat from successful hunts being shared equally amongst all family groups. In which case successful hunting would not have improved the survival-chances of the hunter or his off-spring directly. Nevertheless, successful hunters are highly respected; and there is evidence, both that they father significantly more children than do other men, and (somewhat more puzzling) that their children have a greater chance of surviving to reproductive age (Boyd and Silk, 1997).[16]

Another possibility (although not one which I really understand, I confess) is that the principles of abductive inference originally evolved to govern the intra-modular processing of our naïve physics or naïve psychology systems (or both), and that they somehow became available to domain-general cognition through some sort of ‘representational re-description’ (Karmiloff-Smith, 1992). For notice that the contents of these modules are at least partially accessible to consciousness (they are not fully encapsulated), since we can articulate some of the principles which they employ. In which case it is possible that whatever developmental and/or evolutionary processes resulted in such accessibility, might also have operated in such a way as to make accessible the abductive inference-principles which they use. (I suppose it is some support for this proposal that mechanics and folk-psychology continue to form our main paradigms of causal explanation, to which we are inclined to assimilate all others.)

The evolutionary explanation of abductive inference-principles falls outside the remit of this chapter, fortunately. For present purposes what matters is that there is good reason to think that they are a part of our natural cognitive endowment.

 

4.5       How to build a tracker or a scientist

I am suggesting, then, that a number of distinct cognitive elements had to come together in just the right way – by accident or natural selection – in order for both sophisticated tracking and scientific reasoning to become possible. The ingredients are as follows.

(1) A variety of innately channelled conceptual modules, including folk physics, folk psychology, folk biology, and perhaps some sort of ‘number sense’.

(2) An innately channelled language-system which can take inputs from any of these modular systems, as well as providing outputs to them in turn. This system thus has the power to link together and combine the outputs of the others.

(3) An innate capacity for imagination, enabling the generation of new ideas and new hypotheses.

(4) An innately channelled set of constraints on theory choice, amounting to a distinct non-modular faculty of inference to the best explanation.

With all these ingredients in place, humans had the capacity to become trackers or scientists, depending on their motives, interests, and the use to which those ingredients were put.

 

4.6       And is that all?

If this sketch is on the right lines, then the changes which had to take place in facilitating the scientific revolution were mostly extrinsic ones, or involved only minor cognitive changes. People needed to begin taking as an explicit goal the systematic investigation and general understanding of the natural world. (It may be that the pursuit of knowledge, as such, was already an innately-given goal – see Papineau, 2000.) They needed to have the leisure to pursue that goal. They needed a variety of external props and aids – the printing press, scientific instruments, and a culture of free debate and exchange of information and ideas – in order to help them to do it. And on the cognitive side, they needed the crucial belief that in evaluating theories it is a good idea to seek for experimental tests. But they did not need any fundamentally new cognitive re-programming.

            Of course, much depends on what one counts as ‘fundamental’, here. In particular, a powerful case can be made for the significance of mathematical knowledge and mathematical skills and practices in making the scientific revolution possible. And it might be claimed that a mind which reasons in accordance with such knowledge is fundamentally different from one which does not. In which case the claim that our innate cognitive endowment is nearly sufficient for science would turn out to be a considerable exaggeration.

            Here I am perfectly prepared to be concessive. For in the end it really doesn’t matter whether or not the label ‘continuity thesis’ is applicable. What matters is the detailed account of the cognitive capacities and social practices underpinning science, and the facts concerning which of these are part of our innate endowment, and which are culturally acquired. If the views I have been sketching and defending in this chapter are correct, then we can conclude that the former (innately endowed) set is much more extensive than is generally recognised, and in a way which is quite distinct from that proposed by developmental theorising-theorists.

 

Conclusion

I have argued that the cognitive processes of hunter–gatherers and scientists are broadly continuous with one another, and that it did not require any extensive cognitive re-programming of the human mind to make the scientific revolution possible. But I have also argued that it is unlikely that a capacity for scientific reasoning is present in normal human infants soon after birth, and is then put to work in developing a series of naïve theories. Rather, the capacity for scientific reasoning will emerge at some later point in childhood, through the linking together of a number of innately-given cognitive faculties, both modular and non-modular.

 

Acknowledgements

I wish to thank Alex Barber, Nick Chater, Alison Gopnik, Michael Siegal, and Stephen Laurence for their helpful comments on earlier drafts; with thanks also to all those who participated in the discussion of this material at the Cognitive Basis of Science conference held in Sheffield in June/July 2000.

 

 

References

Atran, S. 1990. Cognitive Foundations of Natural History: Towards an anthropology of science. Cambridge University Press.

Atran, S. 1999. Folk biology and the anthropology of science: cognitive universals and cultural particulars. Behavioral and Brain Sciences, 21.

Baillargeon, R. 1995. Physical reasoning in infancy. In M.Gazzaniga, ed., The Cognitive Neurosciences, MIT Press.

Barkow, J., Cosmides, L. and Tooby, J. (eds.) 1992. The Adapted Mind. Oxford University Press.

Baron-Cohen, S. 1995. Mind-blindness. MIT Press.

Bloor, D. 1976. Knowledge and Social Imagery. Routledge.

Boyd, R. and Silk, J. 1997. How Humans Evolved. Norton.

Carey, S. 1985. Conceptual Change in Childhood. MIT Press.

Carruthers, P. 1992. Human Knowledge and Human Nature. Oxford University Press.

Carruthers, P. 1996a. Language, Thought and Consciousness. Cambridge University Press.

Carruthers, P. 1996b. Autism as mind-blindness: an elaboration and partial defence. In P. Carruthers and P.K. Smith (eds.), Theories of Theories of Mind. Cambridge University Press.

Carruthers, P. 1998. Thinking in language: evolution and a modularist possibility. In P. Carruthers and J. Boucher (eds.), Language and Thought. Cambridge University Press.

Carruthers, P. forthcoming. Human creativity: its cognitive basis, its evolution, and its connections with childhood pretence.

Carruthers, P. and Chamberlain, A. (eds.) 2000. Evolution and the Human Mind. Cambridge University Press.

Cartright, N. 1983. How the Laws of Physics Lie. Oxford University Press.

Chater, N. 1999. The search for simplicity: A fundamental cognitive principle? Quarterly Journal of Experimental Psychology, 52A.

Cosmides, L. and Tooby, J. 1992. Cognitive adaptations for social exchange. In J. Barkow, L. Cosmides and J. Tooby (eds.), The Adapted Mind, Oxford University Press.

Dehaerne, S. 1997. The Number Sense. Penguin Press.

Dennett, D. 1991. Consciousness Explained. Penguin Press.

Dennett, D. 1995. Darwin’s Dangerous Idea. Penguin Press.

Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi, D., and Plunkett, K. 1998. Rethinking Innateness. MIT Press.

Evans, J. and Over, D. 1996. Rationality and Reasoning. Psychology Press.

Fodor, J. 1983. The Modularity of Mind. MIT Press.

Fodor, J. 2000. The Mind Doesn’t Work that Way. MIT Press.

Gallistel, R. and Gellman, R. 1992. Preverbal and verbal counting and computation. Cognition, 44.

Giere, R. 1992. Cognitive Models of Science. Minnesota Studies in the Philosophy of Science, 15. University of Minnesota Press.

Giere, R. 1996. Visual models and scientific judgement. In B. Baigrie (ed.), Picturing Knowledge, University of Toronto Press.

Gopnik, A. 1993. The illusion of first-person knowledge of intentionality. Behavioral and Brain Sciences, 16.

Gopnik, A. and Melzoff, A. 1997. Words, Thoughts and Theories. MIT Press.

Hempel, C. 1966. The Philosophy of Natural Science. Prentice Hall.

Hermer-Vazquez, L., Spelke, E. and Katsnelson, A. 1999. Sources of flexibility in human cognition: dual-task studies of space and language. Cognitive Psychology, 39.

Hrdy, S. 1999. Mother Nature: A history of mothers, infants and natural selection. Pantheon.

Hughes, C. and Plomin, R. 2000. Individual differences in early understanding of mind: genes, non-shared environment and modularity. In Carruthers and Chamberlain.

Karmiloff-Smith, A. 1992. Beyond Modularity. MIT Press.

Koslowski, B. 1996. Theory and Evidence: The development of scientific reasoning. MIT Press.

Lakatos, I. 1970. The methodology of scientific research programmes. In I. Lakatos and A. Musgrave (eds.), Criticism and the Growth of Knowledge. Cambridge University Press.

Latour, B. and Woolgar, S. 1986. Laboratory Life: The construction of scientific facts. Princeton University Press.

Leslie, A. 1994. ToMM, ToBY and Agency: Core architecture and domain specificity. In L. Hirschfeld and S. Gelman (eds.), Mapping the Mind. Cambridge University Press.

Liebenberg, L. 1990. The Art of Tracking: the origin of science. Cape Town: David Philip Publishers.

Miller, G. 2000. The Mating Mind: How sexual choice shaped the evolution of human nature. Pantheon.

Mithen, S. 1990. Thoughtful Foragers: A study of prehistoric decision making. Cambridge University Press.

Mithen, S. 1996. The Prehistory of the Mind. Thames Hudson.

Nagel, E. 1961. The Structure of Science. Harcout and Brace.

Nersessian, N. 1992. How do scientists think? Capturing the dynamics of conceptual change in science. In R. Giere (ed.), Minnesota Studies in the Philosophy of Science, University of Minnesota Press.

Nichols, S. and Stich, S. 1998. Theory-theory to the max. Mind and Language, 13.

Nichols, S. and Stich, S. 2000. A cognitive theory of pretence. Cognition, 74.

Papineau, D. 2000. The evolution of knowledge. In Carruthers and Chamberlain.

Pinker, S. 1997. How the Mind Works. Allen Lane.

Rorty, R. 1979. Philosophy and the Mirror of Nature. Princeton University Press.

Samuels, R. forthcoming. Innateness in cognitive science.

Shapin, S. 1994. The Social History of Truth. University of Chicago Press.

Spelke, E. 1994. Initial knowledge: six suggestions. Cognition, 50.

Spelke, E., Vishton, P. and von Hofsten, C. 1995. Object perception, object-directed action, and physical knowledge in infancy. In M.Gazzaniga, ed., The Cognitive Neurosciences, MIT Press.

Stein, E. 1996. Without Good Reason. Oxford University Press.

Tooby, J. and Cosmides, L. 1992. The psychological foundations of culture. In J. Barkow, L. Cosmides and J. Tooby (eds.), The Adapted Mind, Oxford University Press.

Wellman, H. 1990. The Child’s Theory of the Mind. MIT Press.

Wynn, K. 1990. Children’s understanding of counting. Cognition, 36.

Wynn, K. 1995. Origins of mathematical knowledge. Mathematical Cognition, 1.

Wynn, T. 2000. Symmetry and the evolution of the modular linguistic mind. In Carruthers and Chamberlain.