related texts

Deacon: The symbolic Species

Deacon's book, The Symbolic Species, got many good reviews. I was specially intrigued by the fact that some reviewers claim that it 'destroys Chomsky'. The book itself (The symbolic species, Terence Deacon (1997), W.W. Norton & Company, 0393317544), however, is a serious disappointment.

The main deficiency of this book is that it does not argue for his main hypothesis at all. The main tool for convincing the reader that the hypothesis is correct is a book-length discussion which is based on this hypothesis, so the reader cannot make sense of the book without accepting the hypothesis.

I thought this book, with such a glaring hole in its argument, would be a good test for critical readers. However, browsing through the many reviews of the book that I could find on the net, it seems that no other reviewer paid attention to this hole. That is true even for those reviews which were critical of the main hypothesis of the book.

Also disappointing is the general praise to the presentation of material in the book. Every reader with a little knowledge in the relevant fields can easily see that the presentation is extremely biased and mixes up speculations, overinterpretations and established facts as if they are all established facts. Yet most reviewers did not bother to point this out.

Maybe even more alarming, apparently this book is used in graduate studies as a textbook. I guess it is useful as a way of killing critical thinking.

The central claim

The central claim of this book is that the evolution of the human brain was driven by language, mainly through the development of symbolic references. The obvious null hypothesis is that human brains evolved for greater intelligence, which gives them advantages in all areas, including handling language. Other possibilities are that the brain evolved mainly for other purposes (e.g. making tools). Thus to support his claim Deacon needs, at least, to argue against the possibility of the brain evolving to support greater intelligence, and better still, to argue against other possibilities as well.

So how does Deacon argue against the 'greater intelligence' hypothesis? He doesn't. In chapter 1, we find this (p. 44):

"If neither greater intelligence, facile articulatory abilities, nor prescient grammatical predisposition of children were the keys to cracking this symbolic barrier, then the evolution of these supports for language complexity must have been consequences rather then causes or prerequisites of language evolution".
However, contrary to the implication of this sentence, the question of greater intelligence was not discussed before this at all, and it is not discussed afterwards either. This kind of argument (by implication of a non-existent preceding argument) is a hallmark of a lack of real argument.

Later, inside his argument against the idea of Language Acquisition Device, there is a slightly longer argument (p. 105):

"We cannot discount the obvious and immense gap that separates what children accomplish easily from what otherwise highly intelligent species and even sophisticated inductive learning algorithms cannot."
The reference to the 'sophisticated inductive learning algorithms' is simply ridiculous, because none of these has anything like intelligence. The comparison to 'highly intelligent species' is also not real, as none of these has anything approaching the intelligence of a child, and more importantly, they don't have the child's learning potential, as measured by the number of new patterns that can be introduced into the child's brain without destroying existing patterns. Thus it is not surprising at all that the child learns far better.

Deacon continues the argument (p.105):

"This is especially compelling since young children are quite limited in other aspects of their learning abilities".
Really? Maybe Deacon means that children may be limited compared to adult humans, but for the argument they need to be limited compared to animals and learning algorithms. Clearly, children have much better learning abilities than these, and even Deacon himself brings an example of this later (select smaller pile to get larger, p. 414).

Apart from the 'arguments' quoted above, Deacon does not bother to argue at all against the 'greater intelligence' hypothesis. He argues against the assumption that simple languages are simple to learn, because they need symbolic references, and from this jumps directly to the conclusion that the brain evolved mainly to support symbolic references.

The rest of the book is based on this hypothesis, and thus argues for it by implication, because to make sense of the book, the reader must accept the hypothesis. Thus the main persuasive force of the book is based on a demagogical trick rather than on sound arguments.

Indexical and symbolic references

Deacon's ideas about the relationship between `symbolic reference' and `indexical reference' are very confused, as this quote from an otherwise quite positive review (James R Hurford, The times Literary supplement, 23Oct1998,p.34) demonstrates:
"As a measure of the difficulty (but also of the seriousness with which Deacon's book is being taken) I can report that I have participated in several reading groups, of university lecturers and postgraduate students in Linguistics and Cognitive Science, devoted to Deacon's chapter entitled `Symbols aren't simple'. In all cases, the discussion groups have ended in disarray, with no one person able to convince the others of any particular interpretation of Deacon's complex three-layered diagram with six different kinds of arrows in it (p.87), illustrating his vision of the relationship between `symbolic reference' and `indexical reference'."
Obviously, if the experts cannot make sense of a text, then it doesn't have a (scientific) sense. It can be interpreted by each reader the way they feel like, which normally means matching their own prejudices. This way Deacon makes his ideas look plausible to many readers, even though it is essentialy content-free.

Discussion of the brain

Deacon is regarded as an expert on the brain, and therefore it may be expected that his discussion of the brain would be better. However, Deacon tries to convince the reader to accept his ideas, rather than give the reader a good picture of the brain. As a result, the book is not giving the reader a balanced picture of the known facts, and thus is not even useful as an introduction to the subject for naive readers. For example, Deacon does not discuss at all the overall structure of the brain, and the basal ganglia are metioned only in passing. The structure and internal connectivity of the cortex itself is not seriously discussed, while connectivity from the cortex to other parts of the brain is mentioned many times (but never discussed systematically). Deacon takes this point to extreme in the only mention of Pyramidal cells in the book (p.247):

In all mammals, the cortex includes neurons (called "pyramidal cells" because of their shape) that send output axons to deeper structures of the brain.
The following discussion doesn't mention intra-cortical connectivity. Note that it is not only that the reader (who is assumed not to know what pyramidal cells are) gets only partial picture of the pyramidal cells. The way the text is written, the reader is most likely to get a strong impression that pyramidal cells don't have any role in intra-cortical connectivity, and also a weaker impression that pyramidal cells are only a small part of the cortex. This is a very misleading picture, as pyramidal cells are the majority of neurons in the cortex, and because they have more connections than other cells, their connections constitute the vast majority of connections in the cortex.

The text is also full of smaller errors, e.g. when Deacon relies on Darwinian evolution in the brain (chapter seven, p.193). He writes in p.195:

"The logic of this process is essentially a darwinian logic : overproduction of random variants followed by selective support of some and elimination."
This, however, is clearly not a Darwinian logic: Darwinian logic requires replication of the more successful 'replicators' (genes in biological evolution). Since nothing reproduces inside the (individual) brain, there are no darwinian processes in the brain.


Yehouda Harpaz