We saw last week and in other recent posts the process of language change, as sounds get shifted, simplified or just dropped over time. There is an element of randomness to this – sometimes, a change just becomes “cool” with a particular community or generation. If there were no randomness, Latin would have developed in more or less the same way everywhere. However, thankfully, there are also some rules or, at least, identifiable tendencies.
One principle of language change is simply that of “least effort” – people do not want to make too much effort to recreate a sound! Put more formally, this can be demonstrated as a tendency, known as the “sonority hierarchy” (this is not a religious group, but a phonological trait!)
Essentially, this hierarchy requires syllables to conform to an order of sounds (at least as a preference) from more sonorous to less, starting from the middle out. Making this slightly more complex, the preference is for vowels; then glides (like the /w/ we heard about last week; also /y/ in words like ‘cube’); then liquids (/l/ or /r/ predominantly); then nasals (/n/ and /m/) and then obstruents (fricatives such as /f/ or plosives such as /p/).
The most obvious use for this hierarchy is in fact to assess the volume of each sound – vowels are much louder, in practice, than obstruents (consider /a/ versus /v/).
The relevance to language change, however, is fundamentally linked to phonotactics. Here, the tendency will be for the middle of the syllable to take the loudest sound (most obviously a vowel) and the extremes to take the quietest: the most commonly given example is ‘plant’, with two obstruents on the outside, a vowel in the middle and a liquid and a nasal towards the middle. This will be allowed in most languages which allow consonants together at all (though many, even major ones such as Japanese and Indonesian, do not), but it will be specific to this order: *lpatn will not generally be allowed, nor even in all probability will *palnt or *plnat.
This is contested and languages do of course have exceptions, hence the use of the word “tendency” rather than “rule” in this post. Nevertheless, it can be used to predict changes or identify dialect variations. For example Latin tabula was fine (three syllables centred on a vowel), as is modern Italian tavola and even Catalan taula (two syllables), but modern French table is tricky, because in practice it has become a single syllable where the obstruent /b/ comes before the liquid /l/ (thus contrary to the hierarchy). Although it is pronounced this way in Standard French (/tabl/), we can predict that this will become unstable. In fact we know this, because in Quebec it is already generally pronounced simply /tab/, with the final /l/ dropped, which thus brings it back into line with the hierarchy.
This does not mean that sounds will not otherwise be lost: English ‘talk’ does conform to the hierarchy (the final obstruent comes after a liquid) but, nevertheless, the /l/ has been lost. The hierarchy gives us a guide as to what to expect, but the “principle of least effort” can become all-conquering…
Ian James Parsley 