[A much-abridged version of this review appears in the Journal of Consciousness Studies, Volume 11, No. 9, September 2004.]

Paul Bloom
Descartes' Baby: How the Science of Child Development Explains What Makes Us Human
New York: Basic Books, 2004, 271 pp., $26, ISBN 0-465-00783-X (hbk).

Paul Bloom covers a lot of ground in this book, though not quite as much as the subtitle would indicate. He draws upon his background in child development to offer a pithy and engaging introduction to the universal human penchant for mindreading (sometimes called ‘theory of mind’). From there, he ventures into the cognitive roots of art, morality, philosophy and religion. The unifying claim which saves the book from miscellany and purports to ‘explain what makes us human’ is that we are natural-born Cartesian dualists.

Bloom deftly wields a breezy and conversational style much like that of colleague and advisor Steven Pinker. He has a knack for lively anecdotes, jokes, quotes and metaphors, delivered in language so admirably jargon-free that he could have called this book Dualism for Dummies. Bloom's book is also provocative, challenging conventional wisdom on a number of points. Not every author would have the nerve to devote a whole chapter to disgust, which he calls ‘the body and soul emotion’, or to link it with social phenomena ranging from slapstick humor to the Holocaust. As an argument concerning the genesis of Cartesian dualism, though, Descartes' Baby is disappointing.

Here is the central claim in Bloom’s own words (p. 191):

The premise of this book is that we are dualists who have two ways of looking at the world: in terms of bodies and in terms of souls. A direct consequence of this dualism is the idea that bodies and souls are separate. And from this follow certain notions that we hold dear, including the concepts of self, identity and life after death.
Although he refers to it here as a ‘premise’, Bloom clearly believes that his claim is supported by the evidence he provides, especially by ‘the science of child development’. Much of the research presented here will be familiar to many JCS readers; concerning ‘the concepts of self and identity’, for instance, Maria Legerstee’s article on ‘Mental and bodily awareness in infancy’ covered similar ground in JCS 5 (5-6). But neither Legerstee nor anyone else, to my knowledge, has used this material to argue for early-onset Cartesian dualism, or even for a predisposition towards it. Most of the novelty in Bloom’s book, then, lies in its interpretation of the evidence. A strong argument for such an interpretation would therefore need to present it as more plausible than competing interpretations. Consider, for example, the possibility that Cartesian dualism is a culturally transmitted conceptual epidemic rather than a genetically transmitted predisposition (a ‘biological accident’, as Bloom puts it). Is one of these hypotheses more plausible than the other? This is the kind of question that Bloom does not address, so allow me to outline it here and show why i consider that omission significant.

We can begin with the well-established observation that infants distinguish between two kinds of things that they perceive: those which are self-moved or self-motivated, and those which remain inert unless moved by external forces. Babies show greater interest in the former kind, i.e. in bodies with properties variously called animacy, agency, intentionality, subjectivity and so on. Exactly which properties are inherent in this primary dualism (my term) is not clear, but researchers tend to lump them together rather than trying to tease them apart when presenting the the data on very young children. Legerstee, for instance, says that ‘by 5 weeks, infants imitate facial expressions modelled by people and not by inanimate objects simulating these movements’; Bloom suggests that babies ‘treat animated objects as beings with psychological states’ (p. 18). Clearly this primary perceptual dualism prepares the ground for mindreading and thus for the infant’s development as a social being.

The question, then, is how we get from the vaguely defined primary dualism to the specific Cartesian variety. In one scenario, some children pick up the notion of a separate and immaterial soul from the cultural surround, and map it onto one side of the primary dualism, concluding that some bodies are animate or aware because they are inhabited by souls. Meanwhile, children with a different cultural background might reach a very different conclusion: that animate and sentient bodies have these properties because they are alive and complex in themselves, and not by virtue of a separable soul. These latter children would thus be intuitively prepared for the more elaborate concepts of self-organization and complex adaptive systems, while the former group might be more inclined toward the theological concepts elaborated by many religions. We might compare these two groups to children who speak different languages, and the ‘primary dualism’ to what Pinker calls ‘the language instinct’. So in this scenario, what is ‘native’ is not any specific theory of human nature, but a predisposition to learn one.

Bloom rejects this scenario—or rather ignores it, since he does not test it against his ‘premise’ that Cartesian dualism is innate, but simply presents all of his wide-ranging evidence in the light of that premise. The problem is that most of this evidence could be equally well read as demonstrating the vague primary dualism, or as showing that the specifically Cartesian spin on it has been learned; and Bloom makes little or no effort to refute (or even mention) these alternate readings. I will present below a few examples of Bloom's use of evidence, and of alternative readings.

At one point, Bloom summarizes a large body of research by narrating a conversation he had with his six-year-old son Max. Asked about what the brain does, Max listed seeing, hearing, smelling and thinking; but he also said that the brain is not responsible for dreaming, feeling sad, or loving his brother: ‘Max said that this is what he does’ (p. 200). Bloom of course interprets this as an expression of Cartesian dualism. But nothing in this anecdote indicates belief in an immaterial soul. To me it seems more likely that Max is simply giving a holistic explanation: his whole self (body, person), and not just a part of it such as the brain, is the subject of at least some experiences. As a budding philosopher, Max seems more like a young Merleau-Ponty than a Descartes. Note also that it is the more emotional experiences that Max refers to in more holistic terms as ‘what he does’: this would make perfect sense in terms of Damasio's ‘somatic marker’ theory about the origin of selfhood in the brain's mapping of the body (Damasio 1994, 1999).

At another point, Bloom remarks that if we are really predisposed to Cartesian dualism, ‘even young children should believe that the soul survives the destruction of the body’ (p. 207). To confirm this, Bloom reports research by Bering and Bjorklund, who told children a story (with pictures) which ended with Brown Mouse getting eaten by Mr. Alligator, and then asked a series of questions about the late Mr. Mouse.

The results were striking. When asked about biological properties, four-to-six-year-olds appreciated the effects of death—no need for bathroom breaks, the ears don't work, and neither does the brain. The mouse's body is gone. But when asked about the psychological properties, over half of the children said that they would continue—the mouse can experience hunger, thoughts, and desires. The soul survives.
In the first place, it is hardly plausible that children in this age group would not be influenced by the beliefs of those around them, so we are not getting purely ‘natural’ ideas here. And for ‘over half’ of these children to affirm the survival of the soul is hardly a ringing endorsement for the idea, in a society where 90% of all people believe in Heaven, and ‘most state that they look forward to meeting their friends and family members in heaven’ (Bloom, p. 190). So what this evidence shows is that children learn easily to imagine a disembodied soul. But Bloom presents no evidence that soul/body dualism is easier to learn than, say, an organic/mechanical dualism.

Sherry Turkle of MIT has studied the impact of computers, toy robots and the like on child development. In her view, the normal development of basic concepts such as animacy has been altered by the advent of what she calls ‘computational objects’:

Piaget, studying children in the world of ‘traditional’—that is, non-computational—objects, found that as children matured, they homed in on a definition of life which centered around ‘moving of one's own accord.’ First, everything that moved was taken to be alive, then only things that moved without an outside push or pull. Gradually, children refined the notion of ‘moving of one's own accord’ to mean the ‘life motions’ of breathing and metabolism. This meant that only those things that breathed and grew were taken to be alive. But from the first generation of children who met computers and electronic toys and games (the children of the late 1970s and early 1980s), there was a disruption in this classical story. Whether or not children thought their computers were alive, they were sure that how the toys moved was not at the heart of the matter. Children's discussions about the computer's aliveness came to center on what the children perceived as the computer's psychological rather than physical properties. To put it too simply, motion gave way to emotion and physics gave way to psychology as criteria for aliveness. (Turkle 1998)

In Turkle's developmental story, children's treatment of ‘animated objects as beings with psychological states’ is a recent historical development, a possibility that Bloom does not consider. The point here is that any of these stories may be ‘putting it too simply’ if they identify the vague primary dualism with a specific variety of it such as the Cartesian. And indeed the Cartesian spin on dualism seems an especially odd one, proposing as it does that animals are not animate: they are merely complicated mechanisms devoid of inner life and agency. A human, in contrast, is a subject of experience and an intentional agent because his material mechanism is animated by a separate and immaterial soul (anima). The Cartesian dualism leaps from the observation that some objects are inanimate to the conclusion that material bodies cannot be agents (subjects, organisms, self-organized systems) except by virtue of a separate agency, generally called the soul.

This was apparently Descartes' own version of the dualism Bloom calls ‘Cartesian’, though Bloom does not emphasize this mechanistic aspect of it, and in fact told me (in response to an e-mail question) that he does not consider the human/animal contrast to be an essential part of it. (We have already discussed above children's belief in the soul of a mouse, though as a character in a story Mr. Mouse may fit better into the person category than the animal.) Yet Bloom clearly takes a mechanistic view to be intuitively obvious. On p. 190, for instance, Bloom poses a multiple-choice question:

Do you believe that you are (A) a machine or (B) an immaterial soul? (B) is the aesthetically appealing choice…. Some might wish to answer ‘all of the above,’ self-identifying as both a body and a soul. But only a small minority would choose just (A).
Bloom omits the possibility that someone might answer ‘none of the above’ and self-identify as a living body, an animal of the human kind—which would be the natural choice for any biologist. As Mayr (1982, p. 52) puts it,
All biologists are thorough-going ‘materialists’ in the sense that they recognize no supernatural or immaterial forces, but only such as are physico-chemical. But they do not accept the naive mechanistic explanation of the seventeenth century and disagree with the statement that animals are ‘nothing but’ machines. Organismic biologists stress the fact that organisms have many characteristics that are without parallel in the world of inanimate objects. The explanatory equipment of the physical sciences is insufficient to explain complex living systems and, in particular, the interplay between historically acquired information and the responses of these genetic programs to the physical world.
Clearly Descartes did accept the ‘mechanistic explanation’ of animal bodies, which helps to explain why the animating soul needed to be immaterial in his conception. Later on, some scientists confused the body/machine dualism (outlined above by Mayr) with ‘vitalism’, a cousin of Cartesian dualism. According to Mayr, mainstream biology got over that confusion decades ago, but theoretical biologists such as Robert Rosen (1991, 2000) appear to disagree: Rosen in particular was vociferous, sometimes taking on the tone of a voice crying in the wilderness, in denouncing a persistent addiction the mechanistic model. Meanwhile, the physical sciences seem to have made some progress in dealing with complex adaptive systems (see for instance Gell-Mann 1994), but one could argue that such concepts remain as counter-intuitive to ordinary folk as quantum mechanics and string theory. The point here is that Bloom has not made such an argument, because that would require comparing the organismic concept with the Cartesian in this respect. Bloom tends to assume implicitly that the concept of an immaterial soul is ‘common sense’ while the concept of an organism is not.

For a final example of a source which differs from Bloom in interpretation despite great similarity of data, consider Boyer (2001). This book overlaps with Bloom's to a surprising degree, even in their material on child development (such as the discussion of ‘imaginary friends’)—surprising considering that this is Bloom's home field while Boyer is an anthropologist. Bloom also follows Boyer in many specific examples of religious belief, in the general principle that ‘religious concepts are parasitic upon moral intuitions’, in the emphasis on unconscious ‘inference systems’ (Boyer's term), and in an overall approach usually called ‘evolutionary psychology’ (Bloom calls it ‘cognitive psychology’). Bloom's linking of disgust to religious concepts is much indebted to Boyer's argument (Chapter 6) that human thinking about souls persisting after death is rooted in the deeply unsettling and unavoidable problem of what to do with dead bodies.

But Boyer makes a number of distinctions that Bloom does not, and sometimes rejects as too facile explanations which Bloom readily accepts. For instance, on the question of the soul's survival of death, Boyer says (p. 224):

Philosophers and anthropologists often assume that death poses a special conceptual problem to humans mainly because humans are incorrigible dualists. That is, we all intuitively feel that body and mind are things of a different nature. This would make it difficult to understand how a mind can disappear as a result of the body's destruction. But the cognitive puzzle created by corpses is in fact much more specific than that. It does not result from abstract conceptions of the body and the mind but from our intuitions and from the particular way in which some of our inference systems work. You could very well be a dualist and accept that minds are extinguished when bodies cease breathing. What creates a special problem is not the notion that the person goes on but the conflicting intuitions delivered by two systems, both of which are focused on persons, one dealing with animacy and the other with person identification.

As we have seen, Bloom tries to conflate the products of these two ‘inference systems’ into a single native intuition: ‘Common sense tells us that our mental life is the product of an immaterial soul, and this intuition gives rise to the deeply reassuring idea that the soul can survive the destruction of the body and brain’ (p. 226). Boyer, in contrast, is broadly skeptical about the ‘reassurance’ function ascribed to religious ideas (pp. 19-22), even though ‘religious’ for him is virtually synonymous with ‘supernatural’ or even ‘superstitious’. For one thing, Boyer's idea of a typical religion is a local indigenous belief system (rather than one contaminated by imperialistic literacy such as Christianity), and the beliefs about the afterlife common in these tribal traditions are often not at all reassuring. But beyond that, Boyer's idea of the intuitive roots of religous concepts is quite different from Bloom's. For instance, according to Boyer (p. 144), ‘anthropologists know that the only feature of humans that is always projected onto supernatural beings is the mind’—but this ‘mind’ is not to be confused with the disembodied entity of Cartesian dualism:

Intuitive psychological inferences are applied to intentional agents in general, not just to persons. So it is quite likely that concepts of gods and spirits are mostly organized by our intuitive notions of agency in general (the abstract quality that is present in animals, persons, and anything that appears to move of its own accord, in pursuance of its own goals) rather than just human agency.
This ‘abstract quality’ seems to be nothing more or less than what i have called the primary dualism, and the ‘immaterial soul’ does not seem to be a necessary element in Boyer's version of it.

On his final page, Bloom remarks that ‘the notion that our souls are flesh is profoundly troubling’ (227)—in other words, we intuitively resist thinking of ourselves as organisms. Indeed his whole chapter on disgust would be irrelevant to his argument if we did not take it as supporting this point. But if it is true that our embodiment can evoke disgust on occasion, the same is true of our entire repertoire of emotions. Bloom singles out disgust as ‘the body-and-soul emotion’ simply because it is the only emotion that motivates belief in a disembodied soul.

Perhaps Bloom is simply reading his own intuitions into the evidence presented in the book. If so, this hermeneutic blindness is doubly disappointing, because Bloom is not oblivious to problems of interpretation when he sees them in the work of others. One of the most original points he raises in connection with mindreading is a thoughtful challenge to the usual interpretation of false-belief tests, which are central to the currently prevailing view that children normally begin to attribute mental states to others at age three or four (see for instance Baron-Cohen 1995). So it is rather baffling that Bloom does not consider alternative interpretations of his evidence for children's belief in an immaterial soul.

On the other hand, perhaps the blindness is mine. Bloom (p. 22) says that children ‘suffer from an exaggerated version of a bias that adults possess, sometimes called the curse of knowledge: a pervasive assumption that others have the same knowledge that they do.’ Perhaps Bloom knows that Cartesian dualism is so obvious and intuitive to everyone that there is no need to prove the point, no competing interpretations to defend it against; and his assumption that the reader has this same knowledge is ‘cursed’ for at least this one reader. In that case my whole review is perversely misguided. Well, let other readers decide.

Gary Fuhrman

[For more on Boyer's Religion Explained, see my companion piece on science, religion and explanation.]


Baron-Cohen, Simon (1995), Mindblindness (Cambridge, MA: MIT Press).

Boyer, Pascal (2001), Religion Explained: The Evolutionary Origins of Religious Thought (New York: Basic).

Damasio, Antonio (1994), Descartes' Error (New York: Avon).

Damasio, Antonio (1999), The Feeling of What Happens (New York: Harcourt).

Gell-Mann, Murray (1994), The Quark and the Jaguar: Adventures in the Simple and the Complex (New York: W.H. Freeman).

Mayr, Ernst (1982), The Growth of Biological Thought (Cambridge, MA: Harvard University Press).

Rosen, Robert (1991), Life Itself: A Comprehensive Inquiry Into the Nature, Origin and Fabrication of Life (New York: Columbia University Press).

Rosen, Robert (2000), Essays on Life Itself (New York: Columbia University Press).

Turkle, Sherry (1998), ‘Cyborg Babies and Cy-Dough-Plasm: Ideas about Self and Life in the Culture of Simulation’, in Robbie Davis-Floyd and Joseph Dumit (eds.), Cyborg Babies: From Technosex to Technotots (New York: Routledge). http://web.mit.edu/sturkle/www/cyborg_babies.html.

[ gnoxic studies (home) ]