Re: LONG: Analogy in Conlangs (please help!)

On Sun, Feb 10, 2002 at 09:33:55PM -0500, Elliott Lash wrote:
> Hello Everyone!! I am currently working on the 4th incarnation of Nindic,
> a cousin language of Silindion. Nindic is very influenced by Welsh, and
> like Welsh has undergone lenitions and apocope throughout it's history.
> This has produced a wide variety of irregularities and general
> strangenesses, [or familiaritiesses, depends where you're coming from]
> in both the forms of the stem, and suffixes.
In general terms the developments you list are remarkably parallel to those
found in Welsh, Cornish and Breton.

> A few examples at this point are probably needed:
>
> The past participle is now wholly formed by adding -edh/-es to the root:
>
> meredh <shaded>
> merchedh <hidden>
> tedhes <ruled> (-edh dissimilated to -es)
> pedes <eaten/failed> (homophonous verbs, infinitive: pedi, peidi)
> lhychedh <fallen>
> miredh <loved>
>
> All of these however seem to be formed by analogy with liquid stems (-r/-l)
>
> the *regular* forms would be:
> meredh  < merdh < mor-dai
> merched < merchd < mersk-dai
> teid < teedd < tad-dai
> pest/peist (eaten/failed)<pethd/peethd < pet-dai/pat-dai
> lhyched < lhychd < lusk-dai
> miredh < mirdh < mir-dai
This reminds me of the forms of the 3s past in Middle Welsh. The commonest
form is -as, generalised in Cornish and Breton, but a whole rag-bag of
historically regular endings are found in -ws, -wys, -is, also kemmerth
'found, got' root kemmer, cant 'sang' root can- etc. Mod. Welsh rather than
generalising -as like C. and B. against all the odds went for -odd /-oD/
from MW -awd /-aUD/.

> How realistic is this?  I always worry that i'm introducing far too
> many analogy forms.
>
> Elliott Lash
Very realistic IMHO, a nat lang can only sustain so much complexity (although
written langs and langs with a strong aural literary tradition can hold on to
otherwise obsolete forms). When you get a variety of inhereted forms and
the conditioning factors are lost or obscured, then the alternatives can
be selected on a kind of pick-and-mix basis with the most distinctive forms
usually coming to the fore. A good example are the various plural endings
in WCB, in origin stem formatives in British. They are redistributed on the
basis of semantic and phonological analogies, with different choices often
being made in different daughter languages and their dialects.

Have you posted your sound changes anywhere, I'd be interested in looking
at them?

Keith Mylchreest