Hawaii
The Symbolic Species by Terrence Deacon, Penguin/W. W. Norton,
拢20/$29.95, ISBN 0713991887
TERRENCE DEACON brings formidable credentials to his task of explaining how
human brains and language coevolved. A renowned researcher in neuroscience, he
joins a familiarity with the structure, functions and evolution of the brain to
a thorough knowledge of biology and anthropology. Skilfully integrating content
from these and numerous other fields, he makes complex ideas readily accessible
to the general reader without either patronising or pontificating.The dust
jacket carries encomiums seldom seen even in these days of praise-inflation. One
predicts that it will 鈥渢ransform the foundations of the human sciences鈥�.
Advertisement
Time after time, in sorting through the countless proposals put forward by
language evolutionists, Deacon makes the right choices. Could language have come
directly out of some prehuman trait? No. Does it resemble forms of animal
communication? No. Are vervet alarm calls embryo words? No. Are indexical
associations (simple associative links between a word and an object) truly
symbolic? No. Deacon is often first-rate when he鈥檚 discussing other people鈥檚
ideas. It鈥檚 when he鈥檚 on his own that the trouble starts.
What makes humans different from other species? It鈥檚 not syntax, the rules
which govern the way in which words and elements are linked to produce meaning,
he tells us without discussing that possibility. It鈥檚 symbolic reference. Never
mind that, as he admits, chimpanzees can be taught symbolic reference, and that
at least one bonobo, Kanzi, acquired it so effortlessly that a possible damning
counterexample (that bonobos may refer symbolically in the wild) cannot be ruled
out. And never mind that children鈥檚 reference is indexical, rather than
symbolic, on his own criteria.
鈥淪ymbolic reference derives from combinatorial possibilities,鈥� says Deacon.
But a child鈥檚 first words neither combine into longer structures nor are
definable in terms of other words. But no ape, despite intensive training, has
yet acquired even the rudiments of syntax, and many language acquisitionists
insist that syntax is there even at infants鈥� one-word stage, and fails to be
expressed only because of some 鈥減roduction bottleneck鈥�.
Why is syntax so cavalierly dismissed? Perhaps it is because Deacon is a
neuroscientist, rather than a linguist. Syntax is the core of language, and
those who don鈥檛 thoroughly understand it can never fully explain language
evolution. Any adequate theory should be able to show not merely how syntax
developed but why it has the particular properties it has, and no others.
But recent work in syntax seems hard for the uninitiated to grasp. Deacon
is far from alone in trying to dodge it, but the consequences are dire.
Take his approach to linguistic universals. Do these imply innatism, he asks?
No: 鈥渃hildren鈥檚 minds need not innately embody language structures, if languages
embody the predispositions of children鈥檚 minds鈥�.
Wait a minute鈥攊sn鈥檛 that exactly the same thing in different words?
Perhaps aware that something is wrong, he tries to extricate himself, only to
land in a morass. Languages, he claims, evolve spontaneously, colonise human
brains, and reproduce themselves by passing from generation to generation.
Taking metaphor as bald fact, he then declares them to be living organisms.
Because he wisely eschews the notion of a single ancestral language,
similarities between languages must be analogies鈥攗niversals like the
dorsal fins of sharks, ichthyosaurs and dolphins鈥攔ather than homologies.
With aquatic creatures we know what caused analogous developments: the mechanics
of manoeuvring a largish body at speed through a resistant, three-dimensional
element. What causes language universals? Answer: among other,
unstated factors, 鈥渋mmature learning biases, human mnemonic and perceptual
biases, the constraints of human vocal articulation and hearing, and the
requirements of symbolic reference鈥�.
And that鈥檚 it. Deacon makes no attempt to spell this out, to show how these
factors, singly or in combination, could have yielded a single syntactic
universal. Come to that, he doesn鈥檛 even mention one specific
universal鈥攅xcept colour terms鈥攚hich are not syntactic, and not even
universal, even if their patterning is sharply constrained by perceptual
factors.
Later he claims that innate, brain-instantiated universals cannot exist
because what brains actually process are surface structures. These, he correctly
notes, are often remote from the highly abstract universals that underlie them,
and vary markedly from language to language. Consequently, brains that processed
them could never converge on the deeper structures and thus could never
incorporate them as universals. But those who believe in an innate language
faculty have never proposed a universal grammar established by Baldwinian
evolution (which suggests that both learning and behaviour act as selective
pressures); how could universals have been hard-wired through the brain鈥檚
tracking of behaviour if that behaviour is already driven by universals?
One of the most frustrating things about this book is the frequency with
which Deacon follows a fair, insightful and scholarly discussion of some issue
with a wildly speculative proposal that lacks empirical support, even
plausibility. Faced with the eerie contrast between the syntactic perfection and
low IQ of people with Williams syndrome, the best he can claim is that poor
indexical learning is compensated for by enhanced ability to 鈥渄iscover the
higher-order combinatorial connections among words鈥�. Even granting him both
halves of this equation, what reason have we for thinking that one would have
any connection with the other?
The origins of symbolic reference get similar treatment. With many pages of
densely textured and knowledgeable argument, Deacon shows convincingly that
primates who practised group living, reciprocal altruism, extractive foraging
and female-infant provisioning by males had to have some means of preventing
widespread sexual cheating and consequent social breakdown. It鈥檚 one thing to
make the speculative, but not unreasonable, claim that the means was a primitive
marriage ritual. It鈥檚 quite another to speculate further鈥攖hat this ritual
seeded symbolism in the hominid mind, that it was also the sole source of
language. How on earth was all this done, and what was this earliest language
like? Again, Deacon skates smoothly round the problem area, moving on.
Alas, these lapses into solution by fiat are not the book鈥檚 only flaws.
Another is Deacon鈥檚 tendency not to explain, but to explain away. The most
stunning paradox in the fossil record is that hominid brains tripled in size
with negligible behavioural payoff, then stayed the same size or even shrank,
with revolutionary results. Instead of tackling this paradox head-on, Deacon
chips away at it: earlier hominids may have had sophisticated wood or leather
tools (or they may not); the tools of early sapiens weren鈥檛 so hot; the
explosion of human creativity 50 000 years ago may have been merely 鈥渁
reflection of incidental ecological changes and access to more durable media鈥�.
So in 1.5 million years, Homo erectusnever found the right ecology or
the right medium? What rotten luck.
If I have unduly emphasised the flaws in this book at the expense of its
virtues, that鈥檚 for good reason. The sections where Deacon sticks to what he
knows and understands are so sound, well-crafted and convincing that readers who
have not thought long and hard about language may be swept over the yawning gaps
in fact and logic that lie between without even noticing them. Deacon does not
begin to grapple with the really difficult problems鈥攈ow words emerged, how
syntax emerged. But these problems lie at the heart of language evolution.