On Wed, 27 Jul 2011 17:58:22 -0400, neo gu <[log in to unmask]> wrote:

>So, how would you go about creating, in historical fashion, a language
>with an ablaut morphology from a strictly concatenative one?

This happens all the time by sound change (of course analogy to smooth
things over is good too).  The most common pattern goes like this: the main
vowel of a base is fronted by a following [i], or lowered by an [a], or
raised by a high vowel, or so on; then the contrast between these following
vowels is lost (perhaps the vowels are lost entirely), leaving only the main
vowel to carry the distinction.  A standard English example of this are the
following developments:

*mu:s     *mu:s-i   -i is a plural marker
*mu:s     *my:s-i   i-ablaut
*mu:s     *my:s     loss of final vowels
*mu:s     *mi:s     unrounding of front vowels
"mouse"   "mice"    (this after the great vowel shift)

Of course any sort of change whereby a word-internal sound can react to
material present in some affixes will suffice for this.  You could start
instead from metathesis (e.g. with vowels, [manu] > [maun] > whatever
development; this happens in Roger Mills' favourite Austronesian langs), or
stress alternations (e.g. this will have been the ultimate origin of the PIE
ablaut), or syllable-structure sensitive changes (e.g. the Uralic
gradations, originally conditioned as onset weakening in a closed syllable),
or so on.