One system I had, in a vague experimental conlang, and this kind of summed it up:
Code: Select all
1a: lakki > lakki > lakki: > lakki: > lakk@j > lakka
1b: lakkin > lakin > la:kin > la:ki > lauki > luki
2a: laki > laki > la:ki: > la:ki: > lauk@j > luka
2b: lakin > lagin > la:gin > la:gi > laugi > lugi
3a: laggi > laggi > laggi: > laggi: > lagg@j > lagga
3b: laggin > lagin > la:gin > la:gi > laugi > lugi
4a: lagi > lagi > la:gi: > la:gi: > laug@j > luga
4b: lagin > lajin > la:jin > la:ji > lauji > luji
Basically
1) Consonant gradation occurred in phonologically determined environments (unstressed syllables, lenition, etc.)
2) Vowels lengthened in open syllables
3) Certain syllable coda consonants were lost, causing both vowel length and consonant gradation to become phonemic and grammatically determined
4) Long vowels became diphthongs
5) Those diphthongs became short vowels
The result was that words that had qualitative consonant gradation also underwent vowel changes while words with qualitative consonant gradation did not.
How stable such a system is, I don't know, but I assume, given Germanic languages and Semitic languages, it might be able to stick around for a fair while, and might even be able to handle umlaut and palatalisation:
Code: Select all
1a: lakka > lakk@ > lakka (nothing)
1b: luki > lytS@ > ljutSa (i-umlaut, palatalisation, vowel breaking)
2a: luka > lok@ > loka (a-umlaut)
2b: lugi > lydZ@ > ljudZa (i-umlaut, palatalisation, vowel breaking)
3a: lagga > lagg@ > lagga (nothing)
3b: lugi > lydZ@ > ljudZa (i-umlaut, palatalisation, vowel breaking)
4a: luga > log@ > loga (a-umlaut)
4b: luji > lyj@ > ljuja (i-umlaut, vowel breaking)
Analogy might kick in somewhere along the line, especially where sound shifts become more and more determined by grammatical conditions than the phonological conditions that originally triggered them, but, assuming all of these sound changes are plausible, it kind of gives you an idea of what kinds of changes can occur in related forms of the same word, e.g.
lakki ~
lakkin >
lakka ~
ljutSa.
I think other users might be able to get a handle on how stable or unstable this system might be. I just threw together a series of sound changes to see where it could go.
(getting late, so I decided to switch to X-SAMPA)
DesEsseintes wrote:sangi, I would also be interested to hear what you think of them, since you gave such excellent advice to Trebor here.
I'll definitely have a look over them when I get a chance
EDIT: Not bad, and not too difficult to follow. I think my main problem with it is the attempt to make each pattern four stages long, which doesn't sit well with me. The way consonant gradation is normally presented is basically from a strong grade to a weak grade, so if no change occurs, phonetically, then there has been no change in the grade. So "pp ppʷ → pp ppʷ → p pʷ" doesn't quite make all that much sense since you'd still have to specify under what conditions /pp ppʷ/ switch from the strong grade to the weak grade and where the grade does not change.
Given that consonant gradation, for all sounds which undergo it as determined by the original trigger, undergo similar changes under similar circumstances, where /pp ppʷ → pp ppʷ/, you'd also expect /mb mbʷ/ → /mb mbʷ/ rather than /mb mbʷ/ → /mm mmʷ/... If that makes sense.
Now, I suppose you could come up with some previous state where a given sound, let's say, /hp/ shifted to /pp/ in the weak grade (alongside original /pp/ > weak grade /p/) and then the original strong grade /hp/ merged into /pp/, which would then cause a level of irregularity in the system, since some instances of /pp/ would remain the same in the weak grade while others would shift to /p/ (this might be overridden by analogy, with all instances of /pp/ in the same grammatical environment shifting to weak grade /p/).
Similarly, you'd have to explain "m mʷ → m mʷ".
If some chains are three stages long while others are four stages, I don't see that being much of a problem. The same thing happens, IIRC, in Finnish and in the example I provided, some chains are four stages long while others are just two stages, because each step represents a shift from a strong grade to a weak grade. In Finnish, for example, /t/ is the weak grade of /tt/ but also the strong grade counterpart of /d/, but (to my knowledge, someone might know better), there aren't (many?) instances of /t/ remaining /t/ where, say, /mp/ shifts to /mm/.
I really hope that made sense, and that it's based on the correct reading of your presentation.