Re: [tied] FW: Re: Latin viridian
- Wordingham, Richard wrote:
> Dear Alex (and anyone else who's interested),thank you Richard.
> If you have a PC, I
> think it would be well worth your while getting acquainted with Mark
> Rosenfelder's 'sound change applier' at
> http://www.zompist.com/sounds.htm . The -p modifier is useful for
> debugging your sound changes. If you've got a compiler, it's well
> worth increasing the number of rules; a limit of 200 is too low for
> serious work. I've put together Miguel's rules (and a few others) in
> the attached file romanian.sc and some specimen Latin words in
> latin.lex. Note that vowel length needs to be marked with a colon
> It took quite a while to sort out the ordering of the rules, and I'm
> still not happy that I get $epte, not $apte, from septem 'seven'.
> I've cheated with a couple of the Latin words - 'dre:ctus' and
> 'dre:cta', not 'di:rectus' and 'di:re:cta', and 'virdia', not
> 'viridia'. The latter is interesting; I don't remember anyone
> commenting that when 'viridis syncopated to *'virdis, the neuter
> plural changed from vi'ridia to *'virdia
- Miguel Carrasquer wrote:
> On Tue, 28 Jan 2003 02:10:19 +0000, "Wordingham, Richard"Excuse me Miguel, where should the "e" after palatal not be pronounced?
> <richard.wordingham@...> wrote:
>> It took quite a while to sort out the ordering of the rules, and I'm
>> still not happy that I get $epte, not $apte, from septem 'seven'
> The s->$/_!iV rule comes before the iea > ia and ea > e..e# rules, and
> you need a special rule (also before ea > e..e#) that e after palatal
> sound is not pronounced (ceapã = /c^ap&/, sãgeatã = /s&dZat&/, $eapte
> = /$apte/). So sEpte > siepte > sieapte > $eapte > $apte
In PRB or where?
Actually the sound is an "e" in the diphthong "ea". And in which dialect
should be "sãdzatã" for "sãgeatã"? I am not aware of it but you got me
curious. Or maybe /dz/ here is just the way to write for affricated /g/
like in "joke"?
Normally the /e/ how you said in the previous rules, diphtongued to
"ie". The "i" from /ie/ became a part of the consonant before it , so
/t/+/i/=/ts/, /s/+/i/= S and there should have remained the /e/ in the
word , it was not elided or "mute".
Now it seems that this "e" which remained "free" from /ie/ diphtongued
once again to "ea" and monophtongued once again to "a".
So the chain transformations for vocalism of Latin "septem" are
/e/>/ie/>/e/>/ea/>/a/ for giving Rom. "Sapte".
Theoretically very probable. The question is which was the need for
The answer: there was no need. The actual form in Rom. Lang. should be
explained just trough Latin. Therefore it _must have been_ this chain of
Interesting should be in the actual language that from an /a/ you can
get an /e/ when deriving: Sapte but înSepti= to make it seven times
more. But how you showed these rules should work and there should be
more examples for /e/ which became an /a/ after s,d,t,k,g having now the
form /Sa/, /Za/, /Tsa/, /tSa/,/Ga/. We can try to find some with this
form at the begin of the word.
saddle = Sa < sella
seven = Sapte < septem
snake = Sarpe < serpens, serpentis ( see Note 1)
six = Sase < sex
Counterexamples: (see please Note 1)
to sit = a Sedea < sedere
slave = Serb < servus
plain = Ses < sessum
Sarpe= there is too the forms with "Serpe" and all derivatives are with
Serpoaica, SerpiSor, Serpui, Serpoi, Serpuitor
Sedea= in conjugation there is too the form with "a":
eu Sed (Sãd), tu Sezi(Sãzi), el Shede ( el Sade), noi Sedem, voi Sedeti,
ei Sed ( ei Sãd)
larch tree = zardã < daeda (=taeda)
Counterexamples ( see please Note 2)
sauce = zeamã < zema ( is here short e?)
ten = zetSe < decem
dowry = zestre < dexteram
zeama= derivates are with "e". zemos, zemui
country = tsarã < terra
Counter example (see please Note 3)
to weave = tsese < texere
texture = tsesãturã < textura
texture= tsesãturã is not given as inherited from latin . DEx gives it
as a properly Romanian construction from the verb "tsese" and suffix
"-tura". Myself, I am not very clarified about the "-tura" & "-toare"
Suffix "-tura" is from suffix "-tor" (?) but it doesn't matches
perfectly since the feminine form for "-tor" is "toare" cf.
sunator/sunatoare, tsipator/tsipatoare. The suffix "tura" is to find in
words like muratura,acritura, etc.
I ask myself what we can get here. I cannot see now the construction
since /e/>/ie/ and if /g/+/i/ became a /gi/ like in /joke/, then the
words where the remaining /e/ is still present should be words with the
group /g^e/ or, for counter examples, the group /g^e/. If I am not wrong
with this assertion , I try to find some ( ^= the "i" which became part
of affricated /g/)
to moan = geme < gemere
knee = genunchi < geniculus
frost = ger < gelu
twin = geamãn < geminus
moan = geamãt < gemitus
eyelash = geanã < *genna(= gena=check, face)
- On Sat, 1 Feb 2003 10:22:01 +0100, "alex_lycos" <altamix@...>
>Excuse me Miguel, where should the "e" after palatal not be pronounced?In (standard) Romanian. Bourciez $460.3: "Derrière une palatale l'e
>In PRB or where?
ne se prononce pas dans roum. ceapã = ce:pa, sãgeatã = sagitta."
>Actually the sound is an "e" in the diphthong "ea".In general yes, but not after the palatal sounds (/s^/, /z^/, /c^/,
/dz^/ and /y/ [/iea/ > /ia/]).
>And in which dialect should be "sãdzatã" for "sãgeatã"? I am not awareYes: /dZ/, actually. I sometimes slip back into ASCII-IPA /S/, /Z/
>of it but you got me curious. Or maybe /dz/ here is just the way to write
>for affricated /g/ like in "joke"?
for what I prefer to write as /s^/, /z^/. And if /$/ is used for the
first, I'll also go along.
>Normally the /e/ how you said in the previous rules, diphtongued toPay attention. The chain of transformations was /se/ > /sE/ > /sie/ >
>"ie". The "i" from /ie/ became a part of the consonant before it , so
>/t/+/i/=/ts/, /s/+/i/= S and there should have remained the /e/ in the
>word , it was not elided or "mute".
>Now it seems that this "e" which remained "free" from /ie/ diphtongued
>once again to "ea" and monophtongued once again to "a".
>So the chain transformations for vocalism of Latin "septem" are
>/e/>/ie/>/e/>/ea/>/a/ for giving Rom. "Sapte".
/siea/ > /s^ea/ > /s^a/
>Interesting should be in the actual language that from an /a/ you canThe breaking of (i)e to (i)ea occurs only when the vowel is stressed
>get an /e/ when deriving: Sapte but înSepti= to make it seven times
(so not in the infinitive -í, -íre), and when the word ends in -e or
-a (and if it ends in -e, -ea- eventually goes back to -e- in
Daco-Romanian: Macedo-Romanian leadze, seate > Daco-Romanian lege,
>But how you showed these rules should work and there should beBefore doing so, you should have checked for the exact conditions.
>more examples for /e/ which became an /a/ after s,d,t,k,g having now the
>form /Sa/, /Za/, /Tsa/, /tSa/,/Ga/. We can try to find some with this
>form at the begin of the word.
Only stressed /e/ (ie. /E/ from Latin short /e/) before final -e and
-a are covered by the rule. All of your so-called counterexamples
have either unstressed /e/ or stressed /e/ before other vowels, and
can be thrown out right away. I won't even discuss them.
Miguel Carrasquer Vidal