[YG Conlang Archives] > [ceqli group] > messages [Date Index] [Thread Index] >
Rex May - Baloo wrote: > > on 2/27/02 10:03 PM, Mike Wright at darwin@hidden.email wrote: > > > Regarding glides, I don't think they should coexist with certain types > > of vowel combinations. It's too difficult to distinguish the following > > pairs in connected speech: > > > > bua bwa > > By the rules as they stand, these two would be clearly different: > > BOO-ah and BWAH > > > bau baw > > These would be different, but not different enough to suit me > > BAH-oo and BAW. > > The first would tend to squish into the second, so I'd prohibit the first. > > > bai bay > > BAH-ee and BAY > > Prohibit the bai > > > bia bya > > > > BEE-ah and BYAH Note what Rob Speer wrote: ==================== > bua bwa > bau baw > bai bay > bia bya Is there anything to distinguish? To me it seems those are the same in each case, [...] ==================== Many languages have only one or the other. Speakers of those languages have a hard time making the distinction. Those that do not, like Japanese, tend to be ones that have a phonology in which the core of each syllable is a single vowel. On the other hand, you could define a Txeqli syllable as "one or more vowels delimited by consonants or weaks" (this is assuming the dropping of the [w] and [j] glides). Since words must start with consonants, you wouldn't have to worry about including word breaks in the definition, though you could. > > My tendency would be to simply drop the glides, as that would make the > > rule simpler. Then you could have <w> and <y> for other purposes. (I > > recall that we tried <w> as a vowel and didn't like the look of > > it--neither of us being Welsh, I suppose.) > > Hm. If by drop the glides you mean the sound as well as the letters, then > there will be no diphthongs. See above. You can define a string of vowels as a syllable, which is all you need. Syllable-hood is tied to the stress system, I believe, so the phonology needs to be defined as a package, not just as individual elements. > Rather than go that far, I'd be willing to > complicate the rules to say that ai, ei, oi, au, and eu automatically form > diphthongs and are therefore single syllables. And you could even have triphthongs, like <iei>, <iau>, or <uai>. In practice, I doubt you'll find yourself borrowing anything more complex than that. And, since you're borrowing your vocab from natural languages, there's little point in defining a phonology that can generate sounds that will never actually occur, is there? > If we do that, 'dia' can exist, pronounced 'DEE-ah', but 'dya' cannot, > pronounced 'DYAH.' Or, we can say that i and w before a vowel becomes a > glide, therefore eliminating DEE-ah, which I'd hate to do. > > > Also, do you expect nCnN to be true in practice? If not, why not > > narrow it down? Try listing all reasonable combinations, then look for > > a relatively compact set of rules. But I would prefer a more complex > > set of rules yeilding fewer bizarre possibilities such as /bdomen/ and > > /kfey/--not to mention the theoretical possibility of horrors such as > > /ksbqrmnlq/ or /bpbpbpqmqmqm/. > > That I will do next. I'll use Italian as the model. [...] I wonder how Italian does with forcing loan words into its own sound system. That could be a good question to address to sci.lang (but I don't post there anymore). If it accepts loans that violate the native phonology and/or syllable structure, you should probably ignore those, at least in the beginning. Italian does have a couple of potentially difficult sounds, like initial <sf>. Remember to look at the phonology, not the orthography, which has the usual historical baggage. :-) -- Mike Wright http://www.CoastalFog.net _______________________________________________________ "When they wired us humans up, they really should have labeled the wires--don't you think?" -- Ed