Theiling Online    Sitemap    Conlang Mailing List HQ   

Re: Stress shift

From:Keith <kam@...>
Date:Wednesday, March 5, 2003, 0:02
Be 3 Mar 2003 wektaba Andreas Johansson <andjo@...>

> the development of Steienzh as currently envisaged requires that the language > within a relatively short period (a few centuries at the maximum) moved the > formerly phonemic stress (which could occur on any stem syllable, as well > as on certain affixes) to the initial syllable of words, whereupon it started > pretty drastic reductions of unstressed vowels. Is there any natlang precedent > > for a such change? I'm able to invoke rather heavy substratum influences, if > > that helps.
Old Irish, or rather the immediate precursor to OI suffered just such a change. Starting with (presumably) a fairly standard IE phonology (pitch accent perhaps no really strong dynamic stress, something like Ancient Greek), a seriously strong initial stress set in and the original 2nd, 4th etc vowels usually dropped completely. Which of course lead to all sorts of 'interesting' accommodations between the consonants that crashed together. To make matters worse, Irish morphology and derivation makes great use of IE prefixes, mostly related to prepositions, ro << pro, for << super, as << eks and so on. A single preposition was allowed before the stress, but when a second was added the stress shifted, which means that the opposite set of vowels (original 1st, 3rd ...) now dropped out. Given that some verbs are compounded with prepositions to-ber- "to bring", eks-ber- "to say" and many more, and that (p)ro- and no- come and go in different parts of the verb paradigm (before you even begin to think about negative and relative particles, infixed object pronouns etc) you'll realise that OI verbal morphology looks like a city about a year after a major earthquake struck. Most of the Middle Irish period was spent sorting out the mess! Something similar seems to have happened in Potawatomi (hence the native name "bodwadme"). We must have got "wigwam" from this or a related lang., in Ojibwe IIRC it has all its vowels intact, something like "wikiwamu". So all too plausible, but the results can be messy! Keith Mylchreest