Proto-Languages Part 2
▲
1▲ 1 ▼ 0
How to use diachronics to create languages
This public article was written by [Deactivated User], and last updated on 17 Jan 2015, 13:42.
[comments] untagged
?FYI...
This article is a work in progress! Check back later in case any changes have occurred.
This article is a work in progress! Check back later in case any changes have occurred.
Once you've got a proto-language, how do you go on to develop a language descended from it? Might you just randomly change each word slightly into something that looks like it could be a descendant? Not at all. The process of sound change can be seen to be totally regular (though this is in fact a gross simplification, which we will return to), affecting all words that contain the target phone(s) in the applicable environment(s) without exception. The phonological development of a language can thus be expressed as an ordered series of such regular sound changes. The challenge, then, is to decide on this set of sound changes.
Not all sound changes are created equal. Some are bizarrely improbable, found only in one isolated and poorly documented language, while some are very natural and attested the world over. Naturally, this latter class of sound change is much more likely to occur in the development of your language, but even then, some are more likely than others. The determining factor is the phonology. A language containing, say, one voiced fricative with no other voiced obstruents is likely to either create more voiced obstruents, or remove it completely. If there are voiceless fricatives, it may merge with them, or it could lenite to an approximant. Or, it may spontaneously nasalise, or simply be elided in all environments. Anything is possible.
Most people visualise sound change as a series of spontaneous events, point-like, ocurring for all speakers of a language simultaneously. This isn't true in any respect. Sound change is motivated by learner error, with some people making the same production mistake, and believing it to be correct. This mistake then spreads like a wave to other members of the community, and indeed other communities, until it becomes established. Speakers at the focal point of a sound change will be the most affected, with it applying to virtually every applicable case, and exceptions are few, far-between, and in extremely rare words. Towards the peripheries, the sound change is more sporadic, applying in a more and more restricted set of circumstances, until, right on the very fringes, it only affects a small handful of words, and beyond these speakers, the change doesn't apply at all.
The post-PIE change of satemisation is a perfect example of this phenomenon. The 'Satem' IE branches - Indo-Iranian, Balto-Slavic, Albanian and Armenian all display the fronting of the PIE front-velars (traditionally named palato-velars), typically with subsequent spirantisation. The core of satemisation is likely to be found in Indo-Iranian, where the change was virtually exceptionless. Balto-Slavic, however, frequently blocked the palatalisation in various environments. The change is therefore best represented by a wave, originating in Indo-Iranian, and spreading across the dialects, petering out as it spread, until the change stopped before it reached what are known as the 'Centum' branches.
This wave model of sound change also helps explain how a unitary language fragments. Sound changes affect some speakers but not all, differentiating the language into dialects. Further sound changes create more dialects, and force the older ones apart, creating a dialect continuum. Eventually, the dialects are so distant, they are no longer dialects of a single language, but languages in their own right. This is the best model for creating a naturalistic language family. When a sound change occurs, think about how far it will reach. For languages stemming from dialects in the centre of this region, apply the change without exception. For those on the edges, restrict the change more and more the closer to the edge you get. And remember that dialects move around and spread out. Some dialects die out, and new ones arise. Dialects also borrow words from each other, so they can appear to show evidence for a sound change uncharacteristic of their region.
Comments
