Unnatural clauses

Jacques Guy is not bamboozled by claims that babies lack enough exposure to language to make then fluent as adults

Both these books deal with first-language acquisition. Both address the question of the "poverty of the stimulus". Both argue that there is no such thing. There ends the similarity.

First Language Acquisition is a book that is, for once, accurately described by its publisher: "(It) presents the major findings and debates in highly readable form." In terse sentences, Eve Clark summarises so-and-so's observations and claims about this or that, and closes with a reference. The book is certainly readable, if not highly readable.

Much of it is deeply interesting, even when one begs to disagree. Take the "poverty of the stimulus". This fine piece of jargon means that as babies we get far too low an exposure to our native language to be able to work it out and become truly fluent as adults. Noam Chomsky concluded that we are born with a universal grammar prewired, and that even such notions as "carburettor" are likely to be innate.

In pursuit of evidence against the poverty of the stimulus, Clark reports Bambi Schieffelin's observations of how adults coach young children in Kailuli (New Guinea). They "typically held the child up, facing the prospective addressee, said what the child should say on that occasion, and ended the utterance with... 'say it like that'". That reminded me of when my Lolovuevue informant, his grandchild on his lap, noticed the infant gazing at my Swiss Army knife. "Red, this is red," said old Tarilamalama grabbing the knife. Then he pointed here and there until he had exhausted everything red within sight. So much for the poverty of the stimulus? Not yet.

Four hundred pages later, Clark considers another argument. This time, that "children cannot learn certain structures because they are never exposed to them or because they are not exposed to them in a form that allows the analysis necessary for acquisition". One piece of evidence presented is that "the position of the inverted auxiliary verb depends on the structure of the sentence being used". The following are examples from Clark. A sentence containing a relative clause sees its auxiliary moved to initial position as in " Could the girl who lost her ticket come to the desk?", but this is not so in a sentence containing a conditional clause: "If you don't need this, can I have it?" We don't hear " Can if you don't need this I have it?" Now, anyone in their right mind can easily counter that the rule is mis-stated: the auxiliary is not moved to the beginning of the sentence, but to the beginning of the clause.

So all is not well in Clark's book, but it seems that most often this is because the fault lies in her references. She quotes a five-year-old French child's utterance, la voiture que le monsieur met une roue , being put in a supposedly correct adult form, la voiture à laquelle le monsieur met une roue . But if I said the latter in France, people would say: "You have lived in Australia for far too long." Le monsieur is not adult speech; it is baby talk.

Even then, the construction sounds awfully unnatural, one of those specials fabricated by linguists to make a grammatical point without regard for what a native speaker would say, such as "the car with a flat tyre, which is being replaced now" ( la voiture avec le pneu crevé qu'on est en train de changer ).

Clark is aware of the fabrication problem, for she writes: "The problem with many of these debates lies in the virtual absence of empirical findings and testable hypotheses. The premises have all too often been regarded as facts, and the arguments have raged from there on in."

An index is sorely lacking. Having read the book, if one wishes to return to a specific topic, one cannot locate it. References occupy 160 pages; a modest 80-page index would have been welcome.

And so to Michael Tomasello's Constructing a Language . Coming after First Language Acquisition , what strikes one most is the change in style. Where Clark is terse and dry, Tomasello is florid and verbose. The idea seems to be: think of an obscure, learned-sounding word and use it. And if one can use a word out of context, do so, especially if it might confuse the reader.

For example: "As adumbrated in Chapter 1, the generative grammar hypothesis focuses only on grammar and claims that the human species has evolved during its phylogeny a genetically based universal grammar." In the context of language and grammar you expect "phylogeny" to refer to language trees or syntax trees, but here it can mean only "evolution". So humans have evolved during their evolution. Seminal stuff. "Adumbrated" leaves one pondering whether this idea was outlined, announced or partially disclosed in the chapter, and, since it too is written in the same opaque style, the reader is forgiven for being mystified.

Does this matter? Reading Tomasello's book is like unpacking a Christmas present wrapped in many layers of gaudy paper each swathed in curly ribbons. The unpacking becomes more engrossing than what's inside. Alas, you have to pick through the litter to find the pieces needed to make sense of the puzzle.

There are strange surprises. "Unlike other animal species, the human species does not have a single system of communication." Which animal has a single system of communication? Dogs bark and whine, mark their territories and wag their tails.

Continue unwrapping. "Different groups of human beings have conventionalised different systems of communication (there are more than 6,000 of them)I" The accompanying figure gives away that "systems of communication" are really languages. As so often in this book, three words are used where one would have sufficed.

Now and then you stumble across a potentially interesting piece of data, such as that from Annette Karmiloff-Smith. In a 1986 publication, she claimed that children under five do not master the use of pronouns and definite articles. One example was cited: "There's a little boy in red.

He's walking along and he sees a balloon man and he gives him a green one and he walks off home and it flies away into the sky so he cries."

Interesting, because that is exactly how my informants spoke, to my despair. It is also how Father Sebastian Englert's informants spoke, judging from the aural traditions he collected on Easter Island. My wife does it all the time too. When theorising about children's speech, one should not forget to observe how adults speak.

Eventually, after carefully picking through the wrapping, you find there is in fact no shortage of stimulus, no evidence for an inborn language instinct à la Steven Pinker and no evidence for innate notions and flickable language switches à la Chomsky - the much-touted complexities of grammar, syntax and so on are mainly artefacts of linguists' models. That about unwraps it.

Jacques B. M. Guy is a computer scientist interested in natural language understanding. He holds a PhD in linguistics from the Australian National University.