Re: Plausible Sound-shifts

>I need to know if its plausible to have had sound shifts that would
>yeild the following kinds of words:
>
>Initial fricatives (v'ed or v'less), but no initial affricates (other than
>the s-clusters).
>However, word-final, if there is nothing between the vowel and the
>consonant
>(no syllabant) then you may also have a fricative, but not an affricate. If
>there is a syllabant between the vowel and the consonant, then the
>fricative
>must affricate: f > pf, and so on.
>
>So you can have:
>
>fet, but not pfet.
>tef, but not tepf
>tenpf, but not tenf
>terpf, but not terf
>teypf, but not teyf
Fric > affric before nasals or /r/ makes plenty of sense, but IMO the same
before /j/ seems a bit surprizing.


>I've decided to not allow double syllabants surrounding a vowel, so you
>can't have an initial AND final syllabant, and you cannot have a vowel by
>itself without a syllabant, so your only options are *we*, *le*, *en*,
>*er*,
>and *ey*.
I like basic idea, but what raises my eyebrow is how your the approximants
are split into the seemingly arbitrary groups /w l/ and /j r/. On first
thought, /l/ actually being a velarized [5] could be a possible
explanation... but AIUI velarization is more likely for *coda* liquids? (CF
English.)

It would seem better to me if you had /j/ in the onset, too; then you'd only
need to explain /r/ being aberrant. The yV / wV variation would then be a
bit like the palatalized / non-palatalized distinctions of Irish or Russian,
only vowel-based, and neutralized after /l/. You still wouldn't NEED to have
palatalization occuring later, however; the nuclei might be better
interpreted as difthongs /ie ue/ that could then evolve as vowels, not
consonants.

A problem in this variation is that I can't offhand see a reason to consider
neutralization _before_ /n r/, however, unless some sort of freak metathesis
had occured there. No, wait, if the /l/ were from a non-liquid sorce - a
central semivowel as just discussed maybe??? - then it wouldn't really be
that freaky. Say:
kefi > kjefi > kjef
kefy > kGefy > kGef ... > klef
kefu > kwefu > kwef
krefu > krwefu > kref > kerf > kerpf
kmefu > kmwefu > kmef > kemf > kempf

....Or do you want some sort of alternation between the coda and onset
approximants pairs and need a 2:2 distribution for that? I'm afraid this
won't help in that case.

John Vertical

_________________________________________________________________
Tiesitkö, että varomaton viestintä voi maksaa ihmishenkiä?
http://www.communicationevolved.com/fi-fi/

Reply