GV&SP Discussion: Miscellaneous issues

[Sir C. M. Siervicül]

My quoted comments and King John's answers are from the Word docs posted in the main GV&SP thread. References to page and paragraph numbers in my quoted comments refer to the original GV&SP doc, also linked from the main GV&SP thread.

---

REGARDING THE PRONUNCIATION OF FINAL E AND I

My original comments:

I think the change proposed by specifying that final unstressed e and i are "silent (or very nearly so)" is unnecessary. This complicates Talossan pronunciation rules (by introducing an extra rule for the pronunciation of each letter that didn't exist before) rather than simplifying them, and doesn't seem necessary to accomplish the goals of regularising stress and accents. Recognising that these letters are unstressed is good enough for now. This document doesn't need to get into the phonetic detail of analysing precisely how much more unstressed e and i become than a, o, or u.

King John's answers:

Actually, this was an attempt to simplify (or at least regularise) the rules. We already have various endings in -i (like -etzi and -schti) where the i is silent; I think these aren't isolated cases, but examples of the general rule in Talossan.

The –etzi ending no longer exists. I don't think we should generalize too much from the silent i in the –schti ending, which may be due to the unvoiced sibilant followed by an unvoiced stop. In other words, after all that unvoiced energy, I wouldn't be surprised if a vowel besides i were very nearly silent in an unstressed final position – such as the â in feschtâ or revischtâ. And the i being silent in -schti doesn't necessarily translate into unusually reduced voicing for the final i in martini. So I believe there is a simpler explanation for that fact that SG declares the i in –schti to be silent than assuming that the i in –ici is also silent ("or very nearly so"), contrary to the actual phonetic transcription in SG.

---

REGARDING THE DEFAULT STRESS RULE

My original comments:

I see a minor problem with the default stress rule (which, put most simply, is "stress the vocalic before the last consonant in the word"). In the past, the bare-bones version of the default stress rule was "stress the penultimate syllable." The new version recognises that many Talossan words have dropped their final vowels, making the stressed syllable final rather than penultimate. However, I don't think this rule adequately accounts for words with multiple syllables after the final consonant in the word. For example, oroiâ (wetland) – the default rule tells us that the initial o should be stressed, which doesn't feel right to me. Theoretically, we could have polysyllabic words with no vocalic before the last consonant, like zoaeae (a crustacean larval stage). That word would have to be marked for stress under the rule as written, otherwise there'd be no way to pronounce it.

King John's answers:

I don't see any reasonable way to handle this (very small) group of words by rule. Why not just always mark their stress? That's not much of a burden.

That's not necessarily a problem, but it would technically require marking stress which ought to be fairly obvious on a few very common words, like ziuâ (day), douâ (two), and nouâ (nine). I don't care which course we take, but there you go.

My original comments:

Two suffixes that are in the list of unstressed suffixes in the current rules but not in the list in the GV&SP basic stress rule are -üm and -üs. It appears to me that the great majority of these words should be stressed somewhere other than the final syllable. Should these suffixes be added to the "always unstressed unless stressmarked" list? Look, this is a suggestion from me that results in a reduction of diacritics.

King John's answers:

I disagree. Ben (or someone) wrote, in an attempt at codifying stress:

"If the word contains only one accented vowel then it is stressed, excepting the endings -â, -âs, -escù, -ìci, -mînt, -päts, -phäts and the prefixes în-, ûn-.

"If a word contains two or more accented vowels it is usually the last one which is stressed, excepting the endings -â, -âs, -escù, -ìci, -mînt, -päts, -phäts, -üm, -üs."

I conclude that Ben did not consider words in -üm and -üs generally to have last-syllable stress.

I think you are correct here (I originally missed the significance of "two or more accented vowels" in the rule you quote), but I assume there is a typo in your final sentence. Words in –üs and –üm do generally have last-syllable stress, or else I was right that those suffixes should be added to the "always unstressed" list. So words like Bac'hüs, censüs, californiüm, and corigendüm, having only one accented vowel, are stressed on the final syllable under the current rules, and under the GV&SP rules. If we want any of them to be stressed on the penultimate syllable, we'll have to mark stress ( Bác'hüs, cénsüs, califórniüm, corigéndüm).

---

REGARDING NUMBERS

King John's answers:

Seifet (seven) is pronounced as if it were sáifet, too; maybe it should get a stressmark at least. I really don't know how 50 is supposed to be pronounced, either.

Maybe, since qator relaced ceatháir for 4, we should replace cearéinçe for 40. Qatreinçe anyone?

Good points. I agree that seifet should at least get a stressmark, which is how it used to be spelled until a previous Arestadâ. I agree also that keeping cearéinçe after adopting qator was an oversight that should be corrected. Let's look at the numbers "four" and "forty" in a number of Romance languages:

Catalan: quatre, quaranta
Corsican, Sicilian: quattru, quaranta
French: quatre, quarante
Italian: quattro, cuaranta
Portuguese: quatro, quarenta
Rhaeto-Romance: quatter/quater/cater, quaranta/curonta/curanta (dialectal variation)
Sardinian: cuatru/batro/bàtoro/quattru, caranta/barànta (dialectal variation)
Spanish: cuatro, cuarenta
Old Talossan: quattuor, quadraginta
1997 Talossan: ceatháir, cearéinçe
2003 Talossan: qator, cearéinçe

Excluding the anomalous Sardinian bilabials, the general pattern is the triliteral QTR for "four" (although the q has been changed to c in some modernised orthographies). One difference between "four" in Talossan and the other Romance languages is that in Talossan the r doesn't shift forward to be adjacent to the t like it did in the other Romance languages (r also holds its original position in Rhaeto-Romance). Then in "forty," the t changes to d in Old Talossan/Latin, and the combination dr simplifies to just r in all the daughter languages. The final t from OT quadraginta is retained in the other Romance languages, but mutates to ç (/s/) in Talossan. I would expect the Talossan word for "forty" to be something like *qaréinçe, unless the fact that the final r did not shift in qator suggests that the d from OT quadraginta should survive as something like *qadréinçe. But Rhaeto-Romance lost the d in "forty," too.

Regarding 50, I assume that Ben got [ü] from the Talossan Phonetic Alphabet and [y] from the International Phonetic Alphabet confused to arrive at [ÿ].

---

REGARDING THE Ä VOWEL

My original comments:

Number 2 says "Replace all instances of äi with ai." Why? Nothing in sections I through III would seem to require this change. If it's solely due to perceived difficulty of pronunciation (I personally don't see a problem with it, and neither would Aussies, Finns, or Norwegians), perhaps my proposal in the next paragraph would help here as well.

King John's answers:

Yes, I think äi is hard to pronounce, and it certainly doesn't strike me as very natural for a Romance language. Nuke it I say.

The combination äi is rare enough that I won't sweat the change too much, but I'll note that "hard to pronounce" in this case is completely subjective and based on speaking a Western variant of American English, because [æi] or [æI] is a standard diphthong in Australian English, Cockney English, and some Southern variants of American English, as well as Finnish and Norwegian. You may respond that none of those are Romance languages. Granted, but that's just a reflection of the fact that [æ] is uncommon in Romance languages. When [æ] does occur in Romance languages (as in Quebec French), though, it is found in environments where you, but not the Quebeckers, find it difficult to pronounce. Like before r (where you also suggest replacing ä, as my next paragraph discussed).

My original comments:

Number 3 calls for respelling words ending in –är or –ärC. I have a suggestion which would not require any respelling of these words. In most languages that have ä (German, Swedish, Estonian, Slovak), the pronunciation thereof varies between [ æ] and [ε]. The same goes for æ in Norwegian and Danish (both æ and ä developed from a ligature of a and e, and both have historically been used for the same purpose in Norwegian and Danish, although æ is standard now). This is understandable because phonetically, [æ] and [ε] both fall in between and [e]. When ä is difficult to pronounce as [æ], then, I think the most natural solution is to turn to [ε] as a regular allophone (or a permissible allophone). Given the existence of words like bär (berry), ärör (error), and Tärtuglhán (Tertullian), I really suspect that this is what Ben was actually doing without realising it, just like in German. I don't think Tartuglhán makes as much sense. I think this approach would also be more consistent with the actual history of Latin æ, which by Late Latin had devolved from [ai] to [ε] (although I realise that few instances of Talossan ä actually derive from Latin æ).

King John's answers:

So you're proposing a pronunciation rule that ä is pronounced like e before r? That strikes me as pretty reasonable, and I'd be willing to modify my proposal if people would rather go that way than respell those words.

I think that would be a workable solution. Personally, however, I don't think ä is difficult to pronounce before r. Like I mention above, that sound combination is found in Quebec French (one of the few Romance languages that has [æ]) , and it is a common context for [æ] in Norwegian.

[Sir C. M. Siervicül]

siervicul said:

That's not necessarily a problem, but it would technically require marking stress which ought to be fairly obvious on a few very common words, like ziuâ (day), douâ (two), and nouâ (nine). I don't care which course we take, but there you go.

Replying to my own message, I will point out here that His Majesty has since helped me clarify a misreading on my part of the proposed default stress rule. None of the words I cite would require a stressmark, because they will all be stressed by default on the first vocalic.

[King John]

1. OK, I can go along with Sir Cresti on this. Let's drop my proposed observation about final unstressed i and e, and list the silent i in -schti as an exception.

2. I think we're in agreement here?

3. So ... lets spell seifet with a stressmark — séifet —, and list it among the exceptionally-pronounced words. And I'm very happy with Sir Cresti's proposed qaréinçe for forty. (Although it won't need the stressmark.) Any objections?

4. All right, all right. Before r, ä is "often" pronounced e. (Like, by me, for instance.)

— John R

[Sir C. M. Siervicül]

john said:

2. I think we're in agreement here?

I think so.

3. So ... lets spell seifet with a stressmark — s&eacue;ifet —, and list it among the exceptionally-pronounced words. And I'm very happy with Sir Cresti's proposed qaréinçe for forty. (Although it won't need the stressmark.) Any objections?

Sounds good to me. And you're right, qareinçe won't need a stressmark. I was stuck thinking OS.