Christine Kenneally’s “The First Word”

The horizontal rules below represent breaks in thought and thus need of organization.

The book catches me up with the saga of the influence of Chomsky’s ideas over the last 30 years. On page 38 I must pause to comment on the notion that the Universal Grammar is some pristine mathematically perfect structure that biology (in the form of us) approximates. I see no fact at issue here. There may indeed be a mathematically elegant structure that we can indeed elucidate. It may even generate good predictions about how humans talk. As to the remark that biology builds kludges on kludges, I agree that it usually does. The biological eye with a lens, however, is a good enough model of a mathematically elegant structure for the simple reason that if the model is not closely approximated in several critical ways, the eye simply does not work well. Degrees of approximation provide corresponding degrees of visual competency; this drives gradual evolution. If the ideal is very specific, as the lens is, evolution may well find it and has done so several times. The nature of this ideal has to do with physics.

There may as well be an ideal UG, perhaps indeed along the lines Chomsky proposes. It may be that like the lens, the UG like things that ‘work’ are largely isomorphic to each other. Here the ideal is of mathematical nature rather than physical. Biology on the Earth has so far produced only one sample of linguistic capability, unlike the eye.

Another force for consistent grammars is our appreciation for patterns. We notice patterns in the way we speak and extend those to other situations. The patterns are specific to our local language but at least two things push towards isomorphism between languages:

I look to logic for the ideal. The proposition that ‘you should not drink the water in the stream near the big rock’ is a logical proposition which can be mixed with other propositions to help you act to survive. Our language is good at registering and conveying these propositions to others.


Mathematicians have tried variations on formal logic and the differences are minor. Perhaps this is because of lack of imagination, or merely that we model our formal logic on our own thought processes. Or perhaps there is not much latitude as in the case of the lens.

I skipped a big step above: what is it that propositions are to convey? Generalizations about the world. Propositions require concepts and corresponding logical predicates. Things need names. Logic does not normally include speech acts, imperatives, and other utterances.


Our good but imperfect model of that structure may well be like our model of the pristine UG. Sometimes there is just one way to go—sort of like convergent evolution. We do the elegant thing because there is no alternative—and there is a gradual uphill evolutionary path to it. We know such paths for the eye; I suspect we will discover such paths for language. Kanzi may help us choose among such plausible paths.

There is a brief quotation by Chomsky about the mutation that allowed for the production of a discrete infinity of utterances. I quibble here. There is famous evidence that we are limited to about seven levels of recursion in our ability to generate and parse sentences. This is not a serious objection since 57 is an adequate approximation to infinity in this context. It is a divergence of our language ability from the pristine UG.


I have given my insights better in this context.
I really have to object to the idea expressed on page 172 that music mirrors speech and that is why we like it. This is taken to be an alternative to the Pythagorean observation of integral ratios of frequencies that we find harmonious. Schwarz, Howe and Purvis claim to have observed a tendency towards a the 12 tone scale in a corpus of speech. fragmentary notes.
While I am enjoying the book I must remark that the chapter on ‘Structure’ left me dispirited, at least until it got up to discussing music. I found the second hand reports of structure in animal utterances vague in the extreme. I don’t know whether this is Kenneally or the authors whose work she reports. Few signals in nature are without pattern, even non-biological ones.

Regarding music I think that the noteworthy aspect of music is its connection to emotion. I want badly to know whether the map between music styles and emotion is cultural or genetic.


I think that language and logic (how we think) co-evolved and that we have not done a good job of delineating the relationship. Perhaps language is merely the nexus of a cluster of traits including logic, that co-evolved.
I am reading the chapter “Culture Evolves’. The following question just occurred to me: “Is evolution of the species and evolution of culture an example of co-evolution?”. Not quite! More accurately: “The aptitude for culture is an adaptive trait which evolves as a trait of the species”. If ‘memes’ are the unit of culture then memes can evolve during periods where the aptitude for culture remains constant. Memes can spread while genes are constant. (Thus not co-evolution) Further memes can spread thru a species with a constant diverse gene pool, and even a gene pool with constant statistical distribution. Genetic drift within a constant gene pool is orthogonal to cultural evolution; neither requires the other. Memes can even spread across species; monkeys have learned to point. I think this is mostly spelled out in Blackmore’s book.
Another thought just occurred to me. We have seen the expansion of concepts such as culture and language. It is not so much that we we have learned that animals have culture, but that we have expanded our concept of culture to include what animals do. We have indeed learned that animals do things that are best described by modifying our concepts; it is not entirely avoidance of species bias.
I am enjoying the section “Culture Evolves” where computer modeling of the evolution of language is discussed. Such models are just the sort of “just-so” stories that Chomsky and Gould feared. I think they are fine for they rule out many stories thru shear compute cycles. Computers can carry thru a detailed analysis that is inaccessible to the unaided human intellect. When you set up such a simulation you manifest your assumptions more clearly than in most scholarly papers, at least if you publish your code! One must, however, consider these stories as hypotheses and look for evidence that supports or refutes them.

It sounds like it is time for a book on the results of computer modeling of the evolution of languages.


I am amused that there seems to be a nearly unquestioned notion that there is an original proto-language—sort of a tower of Babel myth. The only likely reason for this is a period, sometimes postulated, when the human population was very small and perhaps ‘degree of separation’ was just a few people. There might also be some sort of analog to the single matrilineal ancestor whose mitochondria design we share. I don’t know how to extend that argument, however.
Language is at once a meme and also a substrate of most other memes. It is also an assemblage of memes which are individual language features. Pointing is a meme that is not part of language proper.
To write of: Logic, Eye, Evolutionary Epistemology. Also language is tied up with speech, understanding speech, logic as a formal language, thinking, convincing, mating, and a good many more. Drawing the line between language and each of these is probably somewhat arbitrary. Thinking of epistemology as a logical activity, as in Kant is a radically different from thinking of it as a biological activity meant to keep an animal alive. Once you have viewed, as an epistemological feat, the swimming of a bacterium in the direction of the sucrose gradient, epistemology, even as people do it, will never be the same again. This perspective is apropos to the question of language. How does communication aid survival?