A catalogue of the variation across different languages

The infant has to start combing through the variation from day one to work out the particular instantiation of the this variation which is the language being spoken between his or her parents and others. Isn’t there something simpler than this?

Typology studies the variation across different languages. In a sense, the learner has only to take account of what he or she observes. But these observations are useless until they are correctly analysed against the background of other possibilities. Every language has words for things and ways of describing states and events. Both of these things vary widely, and interact in different ways. Not infinitely, but widely enough to require a careful evaluation of the evidence.

The Indo-European group of languages are the largest group of languages with around 400 currently spoken by around half the world’s population. Turkic, Semitic, Chinese, the 2,000 odd languages of India, a similar number in Africa, 1,400 across the Pacific ocean and Australia, 1,000 across the Americas, are all unrelated to IndoEuropean. The similarities between Sanskrit, the ancient language of India, and Latin and Greek, had been observed a number of times. But it was William Jones who systematised, catalogued and crucially popularised the observation. As well as words for body parts, members of the family, and numbers, Jones noted that grammatical forms like ways of expressing plurality and tense were also highly diagnostic. So IndoEuropean languages tend to make tense a key feature of the sentence as an inflection on the verb, similar patterns of agreement between the subjects of sentences and verbs and between nouns and adjectives. Of these characters, only tense is strongly reflected in English. Some languages inflect the verb, but not with respect to tense. In a more extreme way, the Maya languages spoken in Mexico have between 250 and 500 words with some of the properties of verbs, adjectives, prepositions, and adverbs, with meanings equivalent to lying curled up the ground. These languages have very few words like IndoEuropean prepositions. Some have only one. How such portmanteau words might be fitted into the framework here is not obvious. Some typologists, Robin Dixon springs to mind, would say that they can’t be fitted into the framework, and adduce that as a supposedly demolition argument against the framework. I personally prefer to allow the possibility of more work on the framework on the one hand and a deeper study of recalcitrant languages on the other.

Suppose a three year old knows that a neighbour is not trusted by the adults in the family and hears it said about a newly purchased luxury car “He allegedly made a very lucky bet.” What sort of word is allegedly? It would standardly be categorised as an adverb with the special property that it can be ordered anywhere amongst the various constituents of the sentence. But this ordering freedom is a tell tale. Allegedly casts doubt about the character of the neighbour and or the propiety of the purchase. But there is a large degree of uncertainty,  a negative evidentiality. Allegedly is effectively a portmanteau word, though one of a different sort from those in Maya languages. Unfortunately has the same ordering freedom, though with quite different meaning. But these words are marginal in English. They do not define in any fundamental way how the language works.

Hearing allegedly, what does the three year old conclude? Possibly that it compounds a negative estimation of the neighbour. If so it falls outside the standard role of adverbs to modify the verb. It is a hint of something like a portmanteau category, an instance of discourse impinging on both the lexicon and syntax.

For the child learning a Maya language there are of course immeasurably more instances of the truly portmanteau words. They are not hints. They are unmistakable. But in order for the child learning a Maya language to progress reliably on the pathway to a grammar where such words are a core aspect of the target language and for the child learning English to realise that the cases of allegedly and unfortunately are just marginal extrusions of discourse, the learning system has to be sensitive to the possibility of throwing the relevant switch either way,

This is the relevance of typology to the study of acquisition and pathology. From the perspective of speech and language therapy, it gives content to the notion of the learnability space.

A number of languages impossible to count exactly

The very scholarly Southern Institute of Linguistics now estimates that there are just over seven thousand languages spoken in the world today.

Very sadly, a language is lost forever roughly every two weeks. For a time there are only a few older speakers. Then there is only one in the sad situation of having no one to talk to in his or her first language. Then, as the last living speaker dies, the language dies with them . But new languages are being formed in large cities as people come together speaking what are plainly and obviously different languages, learn the language of the host community, but adapt it in subtle and not so subtle ways, until within two generations at most, there is a new language which older speakers of the host community language find difficult or impossible to understand.

This process of new language formation is impossible to count. But it may be proceeding as fast as language death.