Menu Close

Cats, dogs, people

And  compositionality

What makes the faculty of human language different from any non-human system of communication is that it is ‘compositional’ – in almost all respects composed of parts which can be put together in organised, ordered ways, like the parts of a face.

The only exceptions are laughter, gestures like waving goodbye, swearing and cursing, gestures made with the mouth like tut tut, and a small handful of expressions like shh, argh, wow, expressing reactions to and feelings about people and events, exclusively part of discourse. Even these forms are almost all built from elements of the sound structure of the language, and often integrated into overly grammatical structures in significant ways.

Other than these special cases, the words, the way they are put together in what is known as syntax, the sounds they are built from with an ordering in time between the gestures, their meanings, are all structured. Crucially, the syntax of human language involves recurring contrasts between ‘content words’ like mother and house and ‘functors’ like there, where, is and the. These contrasts between different sorts of words make it possible to contrast sentences like “There is the house” and “Where is the house?” differing by only the first sound, but very different in their effective force. There are corresponding physical and semantic features in the sign languages of the deaf. This makes the faculty of human language quite different from the barks of dogs and the miaows of cats, or the songs, squawks, and tweets of birds, or the clicks and songs of dolphins and whales, or even the elaborate systems of chimpanzees.

This claim, broadly accepted by linguists over the past 200 years, is disputed by some. For instance, on the strength of observations, Con Slobodchikoff claims that prairie dogs cross-multiply numerous categories of what they correctly see as their main predator, humans, and even create new terms based on these cross-multiplications. So they are said to categorise people by their size, the colour of their clothes, whether they are carrying guns, and even their individuality. It is a huge and astonishing claim. Prairie rats are in the same group of animals as rats and squirrels, both highly intelligent as those who have cared for them know. What Slobodchikoff claims is a significant step towards compositionality in a non-human species. But this does not seem to me a serious challenge to the notion of human exceptionality in speech and language. For one thing, Slobodchikoff has not carried out all the necessary tests, varying all the elements systematically, one by one. But more importantly, Slobodchikoff’s observations all relate to categorisations of potential threat, in an obvious way important to many non-human species, but quite unlike the free compositionality of human language.

The structure of language allows humans to use a finite number of symbols, the sounds or ‘phonemes‘ of the language or the signs of a signed language, to say an infinite number of things and be understand by any other native speaker or signer. This is known as ‘discrete infinity’.

No non-human has ever shown evidence of anything like discrete infinity. To this extent, humans are both unique and exceptional in the animal kingdom. But for some reason, the exceptionality of human language is strongly resisted by some.

Some geneticists take this line because of the difficulty of identifying the corresponding sequence in the DNA. But to me, the onus is on those who deny any sort of genomic explanation to provide a more plausible explanation of discrete infinity.

This is not to suggest that children are born knowing how to talk, a self-evidently absurd aunt sally sometimes peddled by those opposed to any idea of human exceptionality or specific genetic endowment. My claim here, is just that children come to the task of learning language as though expecting to find structures which make discrete infinity definable.  Exactly how this happens has been at the centre of linguistic research for the past 70 plus years by a project which originates in the work of Noam Chomsky. It is the main focus of the proposal here.

The linguist, Stephen Anderson, sets out the issues here in his excellent and famous 2004 book, Doctor Dolittle’s delusion.


When babies babble, they sometimes seem to be experimenting with the phonotactics and phonetic features of a given target language. But there is no evidence of meaning here.

Do you have an enquiry?