«Discussion Review Steven Pinker, How the Mind Works, New York: W. W. Norton, 1997, xii + 660 pp., $29.95 (cloth), ISBN 0-393-04535-8; $17.00 (paper), ISBN ...»
Steven Pinker, How the Mind Works, New York: W. W. Norton, 1997, xii + 660
pp., $29.95 (cloth), ISBN 0-393-04535-8; $17.00 (paper), ISBN 0-393-31848-6.1 )
1. Steven Pinker should be well-known to readers of this journal, not only for
his research in linguistics and cognitive science, but for his semi-popular book,
The Language Instinct (1994; for the M & M review, see Kemmerer 1995). For
those of us who liked The Language Instinct (as I did), Pinker now offers us the rest of the story: How the Mind Works, despite its inclusive-sounding title, literally complements its predecessor (cf. p. x), since it omits all discussion of language;
thus, readers hoping for a user manual for the mind (and instructors hoping for a single-volume cognitive-science text) may be disappointed. One might have hoped for at least a chapter-length précis of the earlier book, though that would no doubt have made this 660-page tome even heftier. As it is, the book covers a wide range of topics: vision; computationalism; evolutionary psychology and natural selection; consciousness and qualia; intelligence; concepts and categories; the problem of other minds; reasoning; emotions; the psychology of social relations; and the psychology of the arts, humor, religion, and philosophy.
The book is aimed at "anyone who is curious about how the mind works", including "professors and students... scholars and general readers" (p. x). Researchers in cognitive science will ﬁnd its "bird’s-eye view of the mind" useful and its evolutionary-psychology slant interesting, even if most of the topics covered go over familiar ground; lay readers should ﬁnd it enlightening (if incomplete – see above) but perhaps a bit snide (especially on the topic of religion – see below). The writing is appealing, with many humorous asides – some to the point, others just throwaway lines that are probably funnier if heard in an undergraduate lecture than when read on paper.
2. Two themes run through Pinker’s story: computationalism and evolutionary
The mind is a system of organs of computation, designed by natural selection to solve the kinds of problems our ancestors faced in their foraging way of life, in particular, understanding and outmaneuvering objects, animals, plants, and other people. The summary can be unpacked into several claims. The mind is what the brain does; speciﬁcally, the brain processes information, and thinking is a kind of computation. The mind is organized into modules or mental organs, each with a specialized design that makes it an expert in one arena of interaction with the world. The modules’ basic logic is speciﬁed by our genetic program.
Their operation was shaped by natural selection to solve the problems of the hunting and gathering life led by our ancestors in most of our evolutionary hisMinds and Machines 10: 381
tory. The various problems for our ancestors were subtasks of one big problem for their genes, maximizing the number of copies that made it into the next generation. (p. 21; cf. pp. x, 429.)
2.1. Pinker advocates a strong version of the computational theory of mind:
What "the brain does, which makes us see, think, feel, choose, and act... is information processing, or computation" (p. 24, my emphasis). Thus, mind (= what the brain does) is a species of information processing, which is equated to computation.
This solves part of the mind-body problem, viz., "how to connect the ethereal world of meaning and intention, the stuff of our mental lives, with a physical hunk of matter like the brain" (p. 24). The solution is that... beliefs and desires are information, incarnated as conﬁgurations of symbols... [which] are the physical states of bits of matter, like chips in a computer or neurons in the brain. They symbolize things in the world because they are triggered by those things via our sense organs and because of what they do once they are triggered. If the bits of matter that constitute a symbol are arranged to bump into the bits of matter constituting another symbol in just the right way, the symbols corresponding to one belief can give rise to new symbols corresponding to another belief logically related to it, which can give rise to symbols corresponding to other beliefs, and so on. Eventually the bits of matter constituting a symbol bump into bits of matter connected to the muscles, and behavior happens. The computational theory of mind thus allows us to keep beliefs and desires in our explanations of behavior while planting them squarely in the physical universe. It allows meaning to cause and be caused. (p. 25.) Although Pinker’s computationalism is a strong one, he distinguishes the computational theory of mind from the "computer metaphor"; i.e., brains aren’t PCs.
Rather, "brains and computers embody intelligence for some of the same reasons" (pp. 26–27); e.g., birds and airplanes both ﬂy, for the same physical reasons, but birds don’t have jet engines or "complimentary beverage service" (p. 27).
Moreover, the computational theory of mind is essential for "mak[ing] sense of the evolution of the mind.... [E]volution equipped us... with a neural computer" (p.
27; cf. pp. 36ff).
But the computational theory of mind disappears after the careful and generally good discussion of it in the ﬁrst two chapters, and little is said about the computational implications of the evolutionary analysis. This may, however, merely represent the state of the art; if Pinker (and his primary sources) are right about what the mind is, much work remains to be done on the details of how it works (despite the book’s title).
2.2. The real heart of the book is evolutionary psychology. Pinker does an admirable job clarifying evolutionary theory, though occasionally he slips into teleological or anthropomorphic locutions (as he also did at the end of the "key idea" quotation from p. 21: How can genes have "problems"?). For instance, instead
DISCUSSION REVIEWThe ultimate goal that the mind was designed to attain is maximizing the number of copies of the genes that created it. (p. 43, my emphasis.) he should have said that the mind has evolved in such a way that the number of
genes that created it is maximized. And instead of saying:
Natural selection cares only about the long-term fate of entities that replicate, that is, entities that retain a stable identity across many generations of copying.
(p. 43, my emphasis.) It would have been more accurate to say merely that it is replicators that natural
selection applies to. Or instead of this:
When we ask questions like "Who or what is supposed to beneﬁt from an adaptation?" and "What is a design in living things a design for? the theory of natural selection provides the answer: the long-term stable replicators, the genes. (p. 43; my emphasis.) It would have been better simply to say that genes beneﬁt (not: are "supposed to" beneﬁt) from an adaptation, and that genes give rise to and are the beneﬁciaries of (apparent?) design in living things.
Or consider this passage, which is an interesting mixture of what I would consider correct terminology, as well as bad teleological/anthropomorphic terminology:
But almost everyone misunderstands the theory. Contrary to popular belief, the gene-centered theory of evolution does not imply that the point of all human striving is to spread our genes.... Our goals are subgoals of the ultimate goal of the genes, replicating themselves. (pp. 43–44; my boldface; italics in original.) It is misleading, especially in a book one of whose purposes is to clarify evolution, to talk about goals; I would have said simply that genes replicate themselves (not
that that is their "goal"). And instead of this:
But the two are different. As far as we are concerned, our goals, conscious or unconscious, are not about genes at all, but about health and lovers and children and friends.... Sexual desire is not people’s strategy to propagate their genes.
It’s people’s strategy to attain the pleasures of sex, and the pleasures of sex are the genes’ strategy to propagate themselves. (pp. 43–44; my boldface;
italics in original.) I would have said that genes have evolved in such a way that we humans experience the pleasure of sex and, as a result, genes propagate themselves.
3. In addition to the twin themes of computation and evolution, I detected an implicit third theme: The "mind" that Pinker endeavors to explain is not mind in general or in the abstract, but the human mind in particular. Some cognitive scientists (I count myself among them – see Rapaport 1998) are more interested in the former – in how cognition, mentality, thinking (call it what you will) works, independently of how it is implemented. Granted, other cognitive scientists (e.g., Thagard 1986) have been at pains to point out that the implementing medium may contribute a lot to the mind’s functioning. And some of these cognitive scientW.J. RAPAPORT ists have suggested that (probably) only minds implemented in animal brains are worthy of the name (Searle 1980 can be read this way; for further discussion, see Rapaport, forthcoming).
Although Pinker gives lip service to the more
... the mind is not the brain but what the brain does, and not even everything it does, such as metabolizing fat and giving off heat.... Information and computation reside in patterns of data and in relations of logic that are independent of the physical medium that carries them." (p. 24; also see the quotation from p.
65, below.) He is concerned primarily with the human mind, as evidenced by passages such as one on postpartum depression: "... as with all explanations of complex emotions, one must ask why the brain is wired so as to let hormones have their effects" (p.
444). It seems unlikely that robots will suffer from postpartum depression. Thus, there remains a certain tension between such an implementation-dependent view of (the human) mind and the computational view, for if mind is computational, then it can be implemented in other media besides the (human) brain, in which case, the effect of hormones is peculiar to the human mind. (Unless, of course, it turns out that mind in the abstract requires some (computational?) analogue of hormones in order to function. Maybe robots would suffer from postpartum depression?)
4. As the context of the previous quotation about hormones may have suggested, there are some topics in the book that are of questionable relevance: Postpartum depression, parent-offspring conﬂict, the nature of sex differences, the human mating system, fashion, and war are not likely to be high on any cognitive scientist’s list of mental functions. Besides the implicit theme of human-brain dependency that I
noted, a clue to the relevance of these topics may be found in a passage on war:
War is, to put it mildly, a major selection pressure, and since it appears to have been a recurring event in our evolutionary history, it must have shaped parts of the human psyche. (p. 510.) So, the theme of evolutionary psychology leads Pinker to be very inclusive about the (human) mind’s functions. But lots of things must have been major selectional pressures – e.g., climate and natural disasters – yet these would clearly be irrelevant topics to include in a book called "How the Mind Works". I hasten to add that the very inclusion in this book of topics such as postpartum depression, war, etc., may well cause them to become mainstream cognitive science, and certainly represents the fact that some cognitive scientists already consider them thus.
5. One disappointment is the overly cute, but opaque, chapter titles. Someone skimming the table of contents would be hard pressed to know what "Revenge of the Nerds" or "Hotheads" were about. So let me tell you.
Chapter 1, "Standard Equipment", describes the design of a robot with vision, motion control, reasoning and common sense, and motives. Evolutionary psychology (relying heavily on the work of John Tooby and Leda Cosmides)2 and computational cognitive science play the central roles in this chapter, though Pinker
DISCUSSION REVIEWsees the latter as a component of the former. Pinker spends the rest of the chapter clarifying evolutionary theory, and concludes with a discussion of some moral implications of evolutionary psychology.
Chapter 2, "Thinking Machines", deﬁnes ‘intelligence’ as "the ability to attain goals in the face of obstacles by means of decisions based on rational (truthpreserving) rules" (p. 62; Pinker’s discussion is based on Dennett 1971, Newell and Simon 1972, and Pollock 19933 ). And, following Dretske 1981, ‘information’ is deﬁned as "a correlation between two things that is produced by a lawful process (as opposed to coming about by sheer chance).... Correlation is a mathematical and logical concept; it is not deﬁned in terms of the stuff that the correlated entities are made of" (p. 65). Pinker expands on the computational theory of mind, giving an extended and good description of one computational model of thinking (production systems), and he discusses the McCulloch-Pitts (1943) theory of mental computation. He takes an internalist, methodologically solipsistic stance (without mentioning Fodor 1980): "The information in an internal representation is all that we can know about the world" (p. 84), and he presents and tries to rebut the Chinese-Room Argument (Searle 1980) and (more brieﬂy) the arguments of Penrose 1989, 1994.
He discusses connectionism, correctly noting that "Connectionism is not an alternative to the computational theory of mind, but a variety of it" (p. 114). As might be expected from those familiar with Pinker and Prince 1988, he takes a moderately, but detailed, anticonnectionist stand: Without modiﬁcation, pure connectionism cannot handle: (1) "the concept of an individual" (p. 114), (2) compositionality, (3) variable binding and quantiﬁcation, (4) recursive thoughts, or (5) rule systems. The chapter concludes with a discussion of consciousness; eventually, Pinker goes along with Colin McGinn’s (1993) view that we probably are incapable of understanding it.4 Chapter 3, unhelpfully titled "Revenge of the Nerds", elaborates on the theory of natural selection and the origin and evolution of life, intelligence, and the brain.
It contains crucial passages on a proper understanding of natural selection:
The eye... appears to have been designed... with the goal of putting together something that sees.... What else but the plan of God could effect the teleology... of life on earth?
Darwin showed what else. He identiﬁed a forward-causation physical process that mimics the paradoxical appearance of backward causation or teleology.
The trick is replication.... [State of affairs] B can’t cause A if A comes ﬁrst.
... But let’s say that A causes B, and B in turn causes the protagonist of A to make a copy of itself[:]... AA. AA looks just like A, so it appears as if B had caused A. But it hasn’t; it has only caused AA, the copy of A. Suppose there are three animals, two with a cloudy lens, one with a clear lens. Having a clear lens (A) causes an eye to see well (B); seeing well causes the animal to reproduce by helping it avoid predators and ﬁnd mates. The offspring (AA) have clear lenses and can see well, too. It looks as if the offspring have eyes 386 W.J. RAPAPORT so that they can see well (bad, teleological, backward causation), but that’s an illusion. The offspring have eyes because their parents’ eyes did see well (good, ordinary, forward causation). Their eyes look like their parents’ eyes, so it’s easy to mistake what happened for backward causation. (pp. 156–57.) Chapter 4, "The Mind’s Eye", concerns the evolution and nature of vision, and contains a fascinating and detailed explanation of "magic eye" autostereograms and Bela Julesz’s (1971) theory of cyclopean perception, as well as a discussion of mental imagery (although he spends only minimal time on the pictorial-vs.propositional debate).