Theiling Online    Sitemap    Conlang Mailing List HQ   

Re: emindahken phoneme list

From:David Peterson <thatbluecat@...>
Date:Monday, March 8, 2004, 8:36
I should be doing my syntax and then sleeping right now, but instead, I'll do
this.

James wrote all the following which is between <<>>:

<<VOWELS
close: [i]; [I]; [u\]*; [u]
close-mid: [e],[2]; [8]**; [o]
open: [A]

* I may mean [y] here, which to me is like an [u] moved to the front.
But then perhaps [u\] is what I had in mind.
** Is this the representation of the German (sorry, no dialectal
knowledge here) o-umlaut?>>

Your high vowel space is very strange...   Throughout your entire inventory,
you have no strictly lax vowels (disregarding [A]) except for [I], assuming
you meant what you wrote ([I] is used for the "i" in "tick".   If you mean the
central vowel, it's either [1], [i-] or [i\]).   If you really had a lax [I] as
a phoneme, you'd expect at very least lax [U], and probably lax [E] and [O]
(and maybe the front rounded versions, but that's a separate issue).   You
might say that /U/ migrated to become [u\], if you mean "barred-u" (high, rounded
central vowel), but that would strike me as odd...   Less odd would be if you
just had a phoneme [y], which is the german "u-umlaut", and would match your
description of "an [u] moved to the front".   If you want barred u, then the
description would be "an [u] moved to the center".   Notice, though, if you
meant barred i and barred u as your two middle high vowels, then you'd have a very
balanced system.

Now for the mid vowels.   You asked if [8] was the representation for German
o umlaut?   Generally not.   Actually, that representation would be [2]
(though it also has a lax variety, which would be [9]--you were off by one).   So
what exactly did you mean by the [2] if you thought [8] was German o umlaut?
The sound you came up with a is a rounded, close-mid central vowel, which
is...French /e/?   Something northeast of schwa.   Anyway, a useful way to do this
would be to look at the vowel space and plunk some points down where you want
your vowels.   As you've written it, your mid-vowel inventory looks rather
strange.

Oh, and low [A] is fine.

<<Long vowels:
[A:], [e:], [o:]
Diphthongs (palatalizations really) allowed:
[A_j], [A:_j], [e_j], [e:_j], [2_j], [o_j], [o:_j], [8_j]
Phonological constraints:
none at the moment, but likely to appear.>>

Why no long other vowels?   Anyway, this looks pretty good.   As for
"phonological constraints", usually a system that's unbalanced like this is an evolved
system, and so there are lots of restrictions on what vowels appear where.
So, for example, if you were to just look at the phonemic inventory of German,
there'd be no reason why you couldn't get a word like [k9n2lYnym9].   This is
an utterly impossible German word, though.   Part of the reason is that only
specific vowels can end a word in German, and lax vowels can only appear in
certain situations.   Also, the front rounded vowels arose, so you wouldn't
expect their distribution to be the same as /a/, /e/, /i/, /o/ and /u/, since they
can appear in any environment (except no longer word-finally, for most of
those).   So if you want to do an unbalanced system, I'd recommend creating a
balanced system for an older version of the language, and then creating some
sound changes that give you the vowels you want.   Once you've got those in place,
you'll only get your unbalanced vowels in very specific environments.

Now, you can further restrict this.   For example, even in the older form (if
you go that rout), you can specify that certain vowels can't end words.   To
avoid this, you could create some rules like: /e/ > [i] / _# and /o/ > [u]
/_#.   You could also do similar things with stressed vs. unstressed syllables,
light vs. heavy syllables, etc.

You'll probably need to come up with a story about why none of the high
vowels can be long.   You can do this one of two ways: (1) High vowels did used to
have long counterparts, but they dropped out for some reason; or (2) there
were no long vowels, but certain processes (which affected only non-high vowels)
gave rise to new long vowels, which were phonemic.   There are a number of
ways you can do this, but I can't think of any plausible ones off-hand (i.e.,
ones that only affect non-high vowels)...   I'll give it some thought.

Related to this, long vowels, even when phonemic, often undergo changes.
So, for example, let's say you had a sound change where short mid-vowels raised
word finally, but also long vowels shortened word finally.   Then you could
have two words: /kano/ and /kano:/ which would become [kanu] and [kano],
respectively.   Now you've got some interesting stuff going on.   Anyway, the point
is you want to avoid words like /kina:lo:ke:/.   Stress languages do stuff to
these long vowels to make it so that the unstressed version aren't as long,
usually (note: *usually*.   I can think of counterexamples).

Now, onto the consonants.

<<CONSONANTS
plosive: [p],[b]; [t],[d]; [k],[g]; [q]; [?]
nasal: [m]; [n]
tap/flap: [4]
fricative: [f],[v]; [T],[D]; [s],[z]; [x]>>

This is a nice inventory.   Good strong inventory.   One question: Where's
[h]?   I only ask because it occurs in the name of the language.

Something to be aware of is that high vowels do funny things around uvulars,
e.g., [q].   In Arabic, for example, you get two different types of /a/'s.   I
can't think of any real examples with [q], so I'll make one up.   To say "he
was" in Arabic, you say /kaana/, which comes out as ['ka:.na].   If you were
to have a word like that beginning with [q], though, it'd be /qaana/, which
would come out as ['qA:.na].   It may seem slight, but they sound very different.
  Also, in a language like Siglit (shout out to my man Joe, who's learning
Inuktitut!), the high vowels /i/ and /u/ lower to [e] and [o] when occurring
before (not after) uvulars.   So, this is something to keep in mind.  For
example, my newest language, Epiq, has a very simple four vowel system: /i/, /u/,
/a/, /@/.   The language itself, though, has like ten vowels, though (I can't
remember the exact number off-hand), specifically becomes of the uvulars.   So,
you have...

/maku/ ['ma.ku] = one word, but
/maqu/ ['ma.qo] = another word

A similar lowering happens with [i], and so now you have a six vowel system.
 But there's more.   The vowel combinations of /iu/ and /ui/ became [y] and
[M], respectively.   When these occur after uvulars, they also lower:

/makiu/ ['ma.ky] = x
/maqiu/ ['ma.q2] = y

/makui/ ['ma.kM] = z
/maqui/ ['ma.q7] = a

And now the six vowel system has become a ten vowel system.   (Hey, I was
right on the money!   Unless I'm leaving any out...   I don't think so.)   So
these are things to keep in mind for your system, with its uvular stop.

Also, from here, you want to think about possible sound changes.   For
example, voiceless stops often become voiced intervocalically, and voiced stops
often become voiced fricatives intervocalically.   This would produce some curious
ambiguities:

/mado/ > [maDo] = x
/maDo/ > [maDo] = y

Voiced sounds often devoice word finally.   So, let's say you had...

/manat/ > [manat] = x
/manad/ > [manat] = y

And now you've got some more ambiguity.   Though, you could resolve this
ambiguity in an interesting way.   Let's say you lengthen vowels before voiced
consonants (like in English, and other languages), and that this change happens
before the devoicing.   Then you'd get...

/manat/ > [manat] = x
/manad/ > /mana:d/ > [mana:t] = y

And there's your long /a/!   Of course, you'd have to come up with a story as
to why you can't have high vowels in such a position...   One thing that
happens commonly is that long vowels shorten and short vowels become lax.   But
we're off vowels: Consonants.

Something else that's interesting is gemination.   Now maybe you don't want
geminates in your language.   That's fine.   There's nothing that says that
your language couldn't have had them at some point in time in the *past*, though.
  And notice what can happen...

If voiceless stops voice intervocalically, voiced stops fricativize
intervocalically, and all geminates shorten, then you could have a *very* interesting
situation in which...

/mata/ > [mada]
/mada/ > [maDa]
/matta/ > [mata]
/madda/ > [mada]
/maDa/ > [maDa]

And if you want voiceless fricatives to voice intervocalically, then...

/maTa/ > [maDa]

And maybe you want voiced fricatives to disappear, unless they're in a
geminate.   Oooh!   Add geminates to the fricatives!   Then the big list becomes...

/mata/ > [mada]
/mada/ > [maDa]
/matta/ > [mata]
/madda/ > [mada]
/maTa/ > [maDa]
/maDa/ > [ma:]
/maTTa/ > [maTa]
/maDDa/ > [maDa]

And now you have a medial distinction between [t], [d], [T] and [D], *and*
long vowels!   Plus, if you have a conservative orthography (and these are fun),
then all the words on the right will be spelled like the words on the left!

All right, I'll move on.   But also note you didn't list [j] as a consonant,
though you have it in your diphthongs (and how will it work, by the way?   Is
it an allophone of [i], or a consonant in its own right?   Is its distribution
restricted?   Can it occur before high vowels?   If so, does it change or
stay the same?)

<<Additionally, I have the following palatalized consonants:
plosive: [t_j],[d_j]
nasal: [n_j]
fricative: [s_j] (which I think is actually [s`]), [z_j] (probably [Z])>>

This reminds me: The diphthongs you described aren't palatalization: That'd
be the opposite (like Russian /je/ instead of /ej/).   Though you did write
them as if they were...   Can you palatalize a vowel, like [e_j]?   I don't see
how...

Anyway, back to this.   If you have a consonant [j] (which you might not need
to, actually), you'll have to decide whether these are actual phonemes, or
arose from Cj sequences.   Also, if these are palatalized, does this mean that
/t/, for example, won't palatalize before [i]?   Also, if you think /z_j/ is
[Z], then it would stand to reason that /s_j/ is [S].   However, they don't need
to be: You can have a palatalized fricative.   Or even just a palatal
fricative (which is different from the [S] and [Z]).   Anyway, if you think of these
as actual phonemes, and not derived in some way, why did you list them
separately?

These can do interesting things.   For example, a palatalized consonant can
raise a low vowel to a mid front vowel.   Palatalization can also spread to
other consonants.   Would you allow palatalized consonants word-finally, like in
Russian?   If so, they'd be true phonemes.   And, if so, you might want to
look at what Russian and other Slavic languages do with their palatalized
consonants.   If they were true phonemes, they'd probably obey whatever rules you
decided on for your other consonants.   But also, if this is actually
palatalization, and not just a palatal series, you might expect palatalization on
*everything*, not just coronal/palatal sounds.   (Oh, hey: Does your language not
have an /l/?   Interesting.   I usually try to avoid /r/'s--they drive me nuts!)
 So, in Russian, for example, you have, for example, [m_jel] which means
something, and you might have [mEl] as well (plus maybe even [m_jel_j] and
[mEl_j]).   Everything gets palatalized.   I think this is why they orthographically
mark it on the vowel, except word-finally (or, at least, nowadays).   So those
are some thoughts on the palatals.

<<Phonological constraints:
No clusters allowed except for /st/ at the moment.>>

This is clearly not true.   Look at the name of the language: Emi*nd*a*hk*en.
  You got two clusters right there!   They're not onset clusters (which is
probably what you meant), but they're clusters nonetheless.   With what you've
written above, you'd predict a language that looked exactly like Hawaiian
except it allowed /st/ to begin a word (hey, this actually matches up with the
Hawaiian in the song "Henehene kou 'aka" as sung by Israel Kamakawiwo'ole!   All
the Hawaiian is (C)V(:), except for one borrowed word from English, which is
pronounced /stu/!).   I've mentioned some rules already, but here you may want
to consider syllables.   What can and can't end a word?   What other onset
clusters can you have?   How about medial?   So, you can have /ste4i/ as it stands
now, but what about /stest4i/ or /stenst4i/?   On the subject of nasals, will
you allow such minimal pairs as...

/stenpi/ vs. /stempi/?

(It's obvious that you want nasal+homorganic stop sequences, at the very
least since there's one in the name of the language.)

Most languages don't allow this.   Usually that /n/'ll become [m] in this
situation.   The same goes for when /n/ comes next to a palatal, or a velar, or a
uvular...   Oh, what's the distribution of [?] going to be?   Will it be a
true phoneme like all the rest, or will it be treated specially?   In Hawaiian
you'd want to say it's a true phoneme, but not in English, where it appears at
the beginning of (some) V-initial words.   However, if it is a true phoneme,
will it be able to appear in cluster?   Word-finally?   Will there be a
distinction (as there is in Hawaiian) between words that begin with and without a
glottal stop?   If it does appear in clusters, a common side-effect is that, say,
/p?/ becomes [p'], an ejective.   Your language could have ejectives!   Isn't
that exciting?!

Back to nasal+consonant sequences, what about nasals+fricatives?   Some
languages allow this (like English: "insist"), but some languages don't.   The
Bantu languages are a good example.   They show what's commonly referred to as
fortition, or strengthening.   So, if you have words such as those listed below,
and you add a general nasal prefix...

/N + fatu/ > [mpatu]
/N + seni/ > [nteni]
/N + Sona/ > [ntSona]
/N + xalu/ > [Nkalu]

Some Bantu languages take this even further and voice all those stops
(voicing assimilation), and some also do it with /l/ and /r/.   So, if in your
language you had two words...

/kinta/
/kinsa/

The might both become [kinta].   A common way to fix this is simply to put a
stop in between, producing [kintsa].   And, hey, maybe the previous vowel
becomes nasalized, and maybe the nasalization drops out, and then you have another
medial distinction [kitsa] (vs. [kida], [kita], [kisa], [kiza], [kiTa], and
[kiDa]).

Another issue is what consonants can end medial syllables, and what
consonants can end words.   These are often not the same.   So, if I remember right, in
Ancient Greek, a word could only end in [n] or [s], but you only have to go
as far as "alpha" to see that a medial syllable could end in more than just
those two consonants.   The reverse can also apply, though.   In Epiq, a medial
syllable can end in [s], [n] or [l].   A word can end in these, as well, but
can also end in [T], [k], [q], and [X] (everything but labials [there's a rule
that affects only coronal consonants, where final [-cont, -nasal] coronals
fricativize.   This changes /t/ > [T], and also /l/ > [L] (? voiceless lateral
fricative), but this sound dropped out word-finally]).

Then after this you want to consider what size of words you want, but that's
a different matter.   This is just to give you some ideas on what you can do.
 Now onto syntax.

-David