In the linguistic crystal ball

One of the language topics in news chat this last week has been the latest 'coming of age' of voice recognition technology. There is, of course, a long history in this field, and there has been software on the shop shelves for a while, but it is in the headlines now because it is gaining a higher profile in mobile devices.

So what could be the linguistic consequences of an increasingly widespread use of this technology, by which a computer or device is able to process spoken input?

It is always difficult to look into the future where mass communication trends are concerned. Take the impact of text messaging, for example. The need to communicate complex messages with a limited number of characters sparked a whole set of linguistic habits among the texting population. This has led some linguists into a knee-jerk reaction about what they see as a destructive effect on language use.

However, a good starting point on this debate from an analytical point of view would be David Crystal's book Txtng: The gr8 db8. In this, Crystal argues that the language of texting is neither as deviant nor as disruptive as some linguistic purists might have us believe.

It is certainly fascinating to see a new linguistic context developing its own grammatical behaviour. Take 'you' and 'your', for example. One texting principle is that words can be reduced to a phonetic representation, so 'you' becomes the single character 'u'. This is preferable due to its more economic use of 'UI real estate'. But what happens to 'your'? Many years ago, if I needed to shorten 'your' in a scribbled classroom note, it became 'yr'. This follows another texting principle of shortening words by removing vowels, and has been used in texting. However, 'u' seems to have grown beyond being a simple phonetic representation. As one might expect in a language like English, which has a famously eccentric link between graphology and phonology, 'u' has become a morpheme as well as a phoneme. Therefore, 'your' can also become 'ur', with the 'u' being given a possessive inflection and losing its purely phonetic character. The phonetic change is made clearer by a comparison with 'u r' ('you are').

But I digress.

What about the mass use of voice recognition as a potential influence on language? For one thing, voice recognition technology starts with a limited vocabulary of commands, so that communication is more reliable within a given 'scripted' framework. Anyone who has some knowledge of the aviation industry may recognise a principle of 'controlled language' here. One trouble is that, when used by the population at large, language rarely remains 'controlled' for long.

So does the phonological similarity between, say, 'recent files' and 're-send files' mean that we will start to pronounce words differently, or in a more exaggerated manner, or with a preference for a particular regional pronunciation? Or will the technology truly develop so that it follows natural speech, rather than vice versa?

Only time will tell.