Re: [PEER REVIEW] Mutations and sound changes (longish)

En réponse à Peter Clark <peter-clark@...>:

>         I think I'm coming down with a case of Maggelity. I had a moment
> of free time
> when I was thinking about how I could have a fairly unique mutational
> system
> when I remembered with Christophe said: "Let's say that if the
> mutation
> system was consistent 1000 years ago, and then evolved like mad, you'd
> have a
> pretty good idea of what it's about." I whipped out my ideas for
> Proto-Enamyn
> and after a little data massaging, got some interesting possibilities.
> Only
> thing is, I need a little help filling in some of the gaps...
Since I gave you the first inspiration, let's carry on ;)) .

>         First, the phonology of Proto-Enamyn, as hypothetically
> reconstructed by some
> academician in an ivory tower somewhere:
> Vowels:
>
>         i 1 u
>         e @ o
>         & a A
>
> Consonants:
>
> p_h p b t_h t d c_h c J\ k_h k g
> m       n       J       N
> f v s z C j\ x G
I suppose you mean p\ and B for the first two of this line, since your
mutations below refer to p\ and B but not to f and v.

> K l
>
> Notes: you will quickly observe that this is a sickeningly orderly set
> of
> phones, but remember that this is the work of an academician who is
> toying
> with some non-existant data in order to explain an unusual system.
Well, Sanskrit is as orderly as this, and nobody ever said it was the product
of an academician in an ivory tower. I think the consonant system is quite
plausible (the absence of /r/ may not be that much of a problem, since /l/ is
there. At worst, you could say that it's an allophone of /l/ in certain
positions)

>         The mutation system is (was?) something as follows ("?" marks
> the ones I
> would like suggestions for, or am extremely doubtful about):
> System 1:
>
> p_h -> p
> p -> b
> b -> B
> t_h -> t
> t -> d
> d -> z
> c_h -> c
> c -> J\
> J\ -> j\
> k_h -> k
> k -> g
> g -> G
So far we could call that 'lenition'.

> m -> p\ ?
> n -> s ?
> J -> C ?
> N -> x ?
Now that's strange. The denasalisation is not a problem, but the devoicing is,
since it kind of implies hardening. Doesn't fit well with lenition. If you want
Proto-Enamyn to have regular mutations, I'd say:
m -> B
n -> z
J -> j\
N -> G
would be a bit more plausible. Nothing prevents you afterwards to devoice them
(to explain why those would be devoiced and not other occurences of the voiced
fricatives, you could hypothesize that they were actually not exactly
fricatives but affricates bB, dz, J\j\ and gG, which got devoiced because of
the absence of their voiceless counterpart, and then were fricativised. So
actually your original mutation would be:
m -> bB
n -> dz
J -> J\j\
N -> gG.
This change still would fit my idea of lenition, since affricates are not as
hard as stops and nasals are nasal *stops*. What do you think of this idea?).

> p\ -> B
> B -> m
> s -> z
> z -> n
> C -> j\
> j\ -> J
> x -> G
> G -> N
I have nothing against voicing voiceless fricatives, it goes well with the idea
of lenition. But nasalising the voiced fricatives doesn't fit well the system,
unless the trigger for lenition was a previous nasal (but then why would the
nasals lose their nasalisation?). On the other hand, simply losing the voiced
fricatives would be possible. Or you could introduce approximants:
B -> w
z -> r\ (or h\ - yes, the voiced glottal fricative :)) -)
j\ -> j
G -> M\ (or G -> 0 would be nice too, it's nice to have at least one consonant
completely disappearing through mutation :))) )
which would still keep your system orderly, or add a very small bit of disorder
by adopting the z -> h\ and G -> 0 changes. Of course, you'd need to define
what those approximants become when mutated too. If we accept the idea of
lenition, then they would simply disappear entirely (or wouldn't be influenced
at all).

> K -> l
> l -> ?
>
If we keep the idea of lenition, l would probably behave like those
approximants I introduced above. So if you decide that they disappear, you'd
have l -> 0 too. But if you decide that they just stay the way they are, then
you'd have l -> l too. One point in mutations is that the resulting system
doesn't need to be unambiguous. On the contrary, look at the different mutation
systems of Celtic languages. There are usually losses of oppositions due to
mutations (like Irish /d/ and /g/ aspirate to the same /G/-/j/, and /s/ and /t/
aspirate both to /h/). You could do the same. No need for your mutation system
to keep things unambiguous. Or what's the fun of it then? ;))))

Of course, this advice is valid only if you want your Proto-Enamyn mutation
system to be regular. But if you want it to be already irregular at Proto-
Enamyn times, then you can discard all my advice :(( .

> System 2:
>
> p_h -> b
> p -> p\
> b -> p
> t_h -> d
> t -> s
> d -> t
> c_h -> J\
> c -> C
> J\ -> c
> k_h -> g
> k -> x
> g -> k
Strange system. It look a bit like an "inversion", with aspirates becoming
voiced and voiced becoming unvoiced, unvoiced fricativising for some reason.

> m -> B
Not m -> b to compare with other nasals?

> n -> d
> J -> J\
> N -> g
> p\ -> p
> B -> b
> s -> t
> z -> d
> C -> c
> j\ -> J\
> x -> k
> G -> g
It's a kind of hardening then? IMHO it fits well with the behaviour of the
stops, even if it would be difficult to explain it :))) .

> K -> ?
What can be harder than a voiceless lateral fricative... Maybe just leave it
unchanged. Mutations don't always influence all phonemes of a language.

> l -> K
>
As for the approximants I introduced earlier, if you decide to keep them, I'd
follow the example of l and mutate them into voiceless fricatives. So:
w -> p\
r\ -> s (or h\ -> s if you take this one. It looks a bit strange, but would be
natural if speakers connect h\ and s together, what they would certainly do
with the s -> h\ change)
j -> C
M\ -> x (if you take this approximant).

Of course, all this is just a starting point. You can always modify it a bit so
that you don't have to introduce new phonemes like the approximants I gave you.
But if you want Proto-Enamyn to have really regular mutations, it's a good
start I think.

> Got all that straight? Suggestions, comments, sharp pins to balloons
> welcome.
Well, how do you categorise what I've just done? ;)))

> The academician has not yet theorized on what triggered the mutations,
> mind
> you...
I thought so ;))) . With the little advice I gave you, the answer for system 1
is easy: plain ol' lenition. For system 2 it's a bit stranger, and I wouldn't
be surprised if it was the result of the merge of two separate mutations in
Proto-Proto-Enamyn ;))) . It's quite possible, once the phonological
environment triggering the change has disappeared and the change is seen only
as grammatical.

>         Ok, next up: sound changes: The vowels are pretty straight
> forward:
>
> i -> i
> 1 -> M
> u -> M
> e -> E
> @ -> _
Schwa disappears?

> o -> O
> & -> E
> a -> a
> A -> O
>
New system a E i O M. Pretty natural...

>         For the consonants, we have several series. First, aspirates
> become
> fricatives:
>
> p_h -> p\
> t_h -> T *new phoneme to the language
> c_h -> C
> k_h -> x
>
> Then, the palatal series is lost:
>
> c -> k_j -> kj (before a vowel); c-> k_j -> k (after a vowel)
> J\ -> g_j -> gj (before a vowel); J\ -> g_j -> g (after a vowel)
> J -> n_j -> nj (before a vowel); J -> n_j -> n (after a vowel)
> C -> S *new phoneme to the language
> j\ -> G
>
So C advanced but j\ retracted... Maybe some dissimilation or something was in
play here...

> Finally, we have a couple of more random sound changes at various
> times:
>
> z -> ts
How did this happen? I could see a z -> dz -> ts chain, but I have difficulties
on having the fricative getting spontaneously a stop onset. Generally the order
goes stop -> affricate -> fricative. I don't remember ever seen the opposite
order...

> g -> j
> p\ -> p
> B -> v
So you would get v without f?

> G -> -> R -> R\ -> r (trilled)
> N -> n (possibly in allophonic variation?)
>
>         Does this seem plausible?
Apart from my two remarks, the rest looks plausible to me.

 The parts that concern me are the
> complete loss of
> the palatal series and the /G/ -> /r/ transformation.
The /G/ -> /r/ transformation looks quite plausible to me. I do /G/ -> /R/ all
the time ;))) . The /R/ -> /R\/ is not impossible (since /s/ -> /r/ is, why not
a nearly parallel change there?) and the /R\/ -> /r/ is even well represented
if we agree that the rhotic in Ancient Greek was /R\/ and became /r/ in Modern
Greek (which is what the common reconstruction of Ancient Greek points at).

 Also, /j\/ by
> all
> rights ought follow /C/ and become /Z/, but since Enamyn doesn't have
> /Z/,
> that just leaves another phoneme to get rid of.
It could merge with /S/, or become /j/... But if your data points out that
Proto-Enamyn /C/ became /S/ and /j\/ became /G/, who are we to discuss
facts? ;))))

>         Also, is it likely that a language would have /K/ and /l/
> (treated by the
> language as a unvoiced/voiced pair) and not /r/?
I'd guess it's OK. If you are really worried, allow [r] to be an allophone
of /l/ in some situations (that could even lead to the separation of /l/ into
two phonemes, both having a relationship with /K/ :)) ).

 I suppose I could
> explain
> the late presence of /r/ as a foreign intrusion...
Not so much if you allow allophonic representation.

Christophe.

http://rainbow.conlang.free.fr

Take your life as a movie: do not let anybody else play the leading role.

Replies