I would love to learn some of this language once you have some interlinearized examples! tylakèhlpë'fö, Amanda On Wed, Oct 21, 2015 at 11:57:59PM -0600, Logan Kearsley wrote: > I went looking through old emails for a phonology sketch I did a while > ago (in 2012, in fact) in response to a Conlangery episode, and ended > up also finding that I did another sorta-Salish-inspired sketch back > in February of this year. It has a lot of similarities with this one, > so I'm stealing from it! We'll just go ahead and call this a > continuation of the same project, I think. > > I also went looking for more information of possessives in Salish > languages, and found these inspirational papers: > http://lingserver.arts.ubc.ca/linguistics/sites/default/files/1996_Burton_Davis-H.pdf > http://semanticsarchive.net/Archive/Tg0MzZhY/Underspecified%20Tense.pdf > > The second one there doesn't actually have anything to do with > possession, so I'm not sure why it came up in my search, but it's got > some cool ideas nonetheless. After reading that tense paper, I'm > thinking of having a basic underspecified non-future tense, where > different verbs are more likely to have a default interpretation as > present or past based on lexically-specified aktionsart. Manipulating > aspect would force perfective verbs to be interpreted as past, but > then there'll be an explicit past-tense clitic (like in St'at'imcets) > to force past interpretation in other cases. Meanwhile, I'm thinking > paraphrastic futures with modal distinctions, consciously stolen from > English. > > The possession paper gave me a lot to think about, and I think I'm > much closer to a decision on how to handle possessives, but not quite > there yet.... > > In the February stuff, I had the idea of a class of oblique pronouns- > essentially, intransitive determiners with the built-in preposition / > case that can be used in place of a relativized clause for an explicit > oblique argument. Those are useful because core argument positions are > all be marked on the verb, regardless of whether there is an explicit > DP/NP for them, but obliques can't be. > > The pronoun set I came up with then had a distinction between 3rd and > 4th person, where 3rds are always definite and 4ths are obviative > and/or indefinite. Before I rediscovered this, I had been thinking of > giving the language both 3rd and 4th person agreement affixes as well, > but with a slightly different distinction- third persons to be used > for proximate items (things that would be described as "here") and > people who are part of the discourse but not being directly addressed, > and 4th person to be used for things that are out-of-sight or people > who are not party to the conversation, with either one potentially > being capable of definite or indefinite marking. Keeping both systems > feels like a delicious bit of naturalistic complexity, and it even has > some internal logic to it- if an explicit DP is not provided for a 3rd > or 4th person argument, then 3rd person agreement would typically > indicate something definite, while 4th person would typically indicate > something indefinite. Thus, the pronouns can be seen as matching up > with the agreement affixes, and just making explicit what is implicit > in the agreement system. > > My February sketch also had an animacy hierarchy, so I'm gonna settle > on that as well- we get more use out of the inverse voice marker that > way. > If animacy rather than syntactic ordering determines core argument > roles, however, then we run into the problem of how to distinguish two > 3/4p arguments in the same clause. In my February sketch, I supposed > that a 3+4 combination would be OK, but a 3+3 or 4+4 wouldn't, but > that would require having different determiners for different persons, > and I don't really want to do that. We could still make an animacy > distinction for definite vs. indefinite arguments, but it's a weird > special case which I'd rather not deal with. So, there are two > options: either those kinds of clauses just are ambiguous, or they are > ungrammatical, and you can actually only ever have a maximum of one > explicit core argument to any verb. Making it ungrammatical is the > more typical Salishan solution, and results in more interesting > discourse structure constraints, so I'm going to go with that. This > reduces the possible extent of center-embedding, and also gives more > usefulness to the subject-marking clitics- if there's an explicit > argument to a transitive verb, and there's a subject clitic, then that > argument must be the object; and if there's no clitic, then that > argument must be the subject! > On a related note, reciprocals will also have to expressed with > multiple clauses. Reflexives, on the other hand, can be handled by > adding one more special detransitivizing affix. > > The hierarchy I have in mind is a very simple 2p > 1p > 3/4p. Further > gradations of animacy in non-discourse-participants should be > unnecessary, because again the subject marking clitic tells you > whether an explicit 3/4p argument is the subject or not. > > In my February musings, I noted that "With two [3/4p arguments], the > situation is ambiguous, and two conjoined clauses are required to > explicitly resolve the ambiguity. Typically, however, in running > discourse, one referent will have a higher expected animacy than the > other, and the usual direct/inverse marking is used to manage your > *expectations* of relative argument animacy/agency." I still think > that's a cool idea, so perhaps I shall save it for a dialectal > variant, or re-use it for a different project. But definitely > ungrammatical here. > > An overview of the phonology (slightly tweaked from what I had in > February, and definitely *not* objectively Salish-like at all, so be > warned) and orthography are as follows: > > Stops come in a voiced and unvoiced series: the usual p/t/k, b/d/g > distinctions, and then a voiced glottal stop and an unvoiced > pharyngeal, which I'm tentatively transcribing as <g'> and <k'>. I > briefly considered <gq> and <kq>, but I figured that would be > confusing in light of the *clicks*, of which there 2 (and sort of 4): > <tq> for a dental/alveolar click, and <lq> for a lateral. > The sort-of-4 comes from the fact that all stops and clicks can be > prenasalized, indicated by an n-digraph, for a total of 16 stop/click > phonemes. Voiced prenasalized stops, however, are only realized as > stops pre-vocalically; in other positions, they become homorganic > nasals. There are no phonemic simple nasals, so this creates no risk > of homophony. > > There is only a single series of fricatives: > > <l> - /K/ (lateral fricative) > <x> - /S/ > <s> - /s/ > <h> - /T/ > <f> - /f/ > > The fricatives and clicks default to unvoiced, but gain voicing > between two voiced sounds (i.e., intervocalically, or between a vowel > and a voiced stop), in a sort of Old-Englishy way. Yeah, /T/ is > over-represented in conlangs that voicing rule is reminiscent of > English/Welsh, but I am fond of /T/ and I think the rest of the > phonology makes this quite distinct enough that these concessions can > be forgiven. > Now, that large fricative inventory seems a little front-heavy, so to > balance things out and make 'em a little more interesting, /k/ and /g/ > will become fricatives in intervocalic positions, thus providing some > back-of-the-mouth fricative phones (even though they're not > distinctive phonemes), one phonemic voicing distinction between > fricative phones, and one unvoiced intervocalic phone. And it's not > even too difficult to come up with a few different plausible > historical explanations for such a scenario. > > There are four basic vowels: > a - /a/ > e - /E/ or /i/ (this may be free variation, or there may be some > consonant-conditioning involved here that I haven't figured out yet) > o - /A/ > u - /M/ > > Vowels come in two series, like Russian, but with rounding rather than > palatalization, for a total of 8. The rounded vowels are written with > a w-digraph, and induce rounding-assimilation on preceding consonants > and preceding hiatus vowels. This can induce mutation of > previously-final vowels during affixation. > > Rounded vowels can occur in hiatus with each other, but unrounded > vowels can only occur individually. I'm imagining that this is due to > a historical process where originally there were no vowels in contact, > but the loss of an original /w/ consonant that produced the rounding > features on neighboring vowels resulted in some instances of > rounded-vowel hiatus. Identical vowels in hiatus are half-long, and > might undergo some subtle quality shifts. > > Syllable onsets must consist of one consonant or a cluster of one > fricative and one non-nasal plosive, unless the vowel is rounded, in > which case null onsets are allowed (again, due to the loss of an > original */w-/, which is still in the romanization as a digraph > character for rounding). Internal codas may be empty or consist of any > single non-click consonant. Clicks can only occur word-initially or > intervocalically. Word-final codas may also include stop+fricative and > fricative+stop clusters, as a result of the historical loss of some > final vowels. I expect this to result in some interesting infixing or > metathesis or something when adding consonant-initial suffixes. The > simple solutions would, of course, be the insertion of epenthetic > vowels (irregular ones that reflect the historical form, of course!) > or loss of one of the consonants, but Infixing & Metathesis Are Cool > as well. > > In the event that the coda of one syllable and the onset of another > syllable are brought together such that a nasalized stop is placed > after a fricative, metathesis occurs to put the nasalized stop first. > If a click ends up after any other consonant (due to affixing), it > assimilates to a corresponding stop: tq -> t/d, lq -> k/g, ntq -> > nt/nd, nlq -> nk/ng. > If a nasalized stop ends up following a normal stop, the initial stop > changes into a prenasalized stop instead; thus, k+nt -> nkt. > > Note that a word ending in a voiced prenasalized stop will have its > pronunciation change depending on whether or not the next word starts > with a vowel or a consonant- it may be realized as a labialized voiced > prenasalized stop, or as a non-labialized voiced nasal. > > With the phonology and romanization basically worked out, I can start > making up words! My initial vocab is likely to come largely from > borrowings out of examples in papers I'm referencing on Salish, but > one big advantage of a very-not-Salish phonology is that, by the time > the borrowings are suitably adapted, you probably won't be able to > tell. > > -l.