[YG Conlang Archives] > [Latejami group] > messages [Date Index] [Thread Index] >
The Shadow wrote: > Wow, great minds think alike! I was also doing verbs as CVCVC, with > modifiers as CVCCV with various endings. (Won't you have to choose > the part-of-speech consonants carefully, though, so that they don't > get mistaken for parts of the root?) I've been contemplating doing > nouns as VCCVC, though I haven't gotten far with it. Not really: if the last consonant belong to the root then the penultimate syllable will usually be open, and if the last consonant doesn't belong to the root the penultimate syllable will usually be closed. There are homophones, though. I haven't decided exactly what to do, if I should accept them or not at all. In either case, I accept them to a certain degree (lower than in natural languages): they have to be few, and I have to think they will be disambiguated in context, and they can't occur in the core of the vocabulary. I haven't accepted any homophones yet, though. My solution is to manually look for alternative interpretations of all words I create and see if there is some ambiguity. If there is, I remove it. I think I won't accept them. But I do accept that the listener has to know the language, though, to be able to parse it. That is, a word of the form CVCCum might be either a verb or a noun, but only one of them will exist. > voice structure in the middle of the word as possible. (What vowel > are you representing with 'y'?) /@/. I have the phonology of Lirakdom here: http://veoler.googlepages.com/phonology.html > Have you thought of using some biliterals for really common words? > They can even use the exact same vowel matrix, just shifted over: > CVCV for verbs, say, and VCCV for others. No. I used to play around with both CC and CCC sometimes in earlier sketches, but not for Lirakdom. Since any consonant could occupy any position in roots, I will get CCh, CC', CCw, CCj and similar, and introducing biliterals would only increase the ambiguity of my morpho-phonology without giving me much in return. > I'd love to see what else you've developed, if you don't mind. (I > don't really have enough concretely decided to share as yet, alas.) I haven't got much. Right now I'm working on expanding the lexicon, since that is what I usually do last, and the grammar is more or less finish, but I want some clay in my hands to work with. A thing which I'm proud of is my transfix stacking system I invented 2007, which I haven't seen in any other language, nat or con. It allows you to have an agglutinating language, where you can have infinitely many affixes in any order, and at the same time have a completely nonconcatenative morphology. It also solves a main problem with taxonomic languages. If you have a root/stem such as 'kam', then add the transfix 'b-e' you get 'bakem', then if you add 'l-f-i' you get 'lafbekim'... All affixes contain a consonant part (the onset and optional coda) and a vowel part. The consonants are prefixed to the stem and the vowels suffixed, and the whole word metatheses. http://veoler.googlepages.com/transfix.html > I'd even offer to collaborate, except that we are bound to have > different visions of what the language is for. Feel free to email me. Yes, proper collaboration is not a good idea. > P.S. What exactly is a 'transfix'? It is a discontinuous affix, excluding circumfixes I presume. A prefix adds its material in the front of the stem, a suffix adds it in the end, an infix adds it inside the root, a circumfix adds it both in the front and in the end. And a superfix adds it by superimposing, e.g. with vowel mutations or tone. If you have CaCaC to be a verb, and CeCiC to be a noun, and a-a is a single morpheme and e-i is a single morpheme, then they are transfixes. If the first "a" and "e" indicates part of speech and the second "a" and "i" aspect and number, they are infixes instead. The stackable affixes mentioned above is also an example of transfixes. But the vowel stem matrix in Lirakdom isn't - they are two infixes and one superfix. -- Veoler