First things first, I've got to develop a phonology, which following the historical approach, requires an ancestral phonology from which to develop it. However, I'm going to commit a cardinal sin of conlanging and design the modern phonology first, at least in broad strokes, so I know exactly what I'm going to be developing. I personally feel that conlangers over emphasise the necessity of going from "proto-language" to a modern descendant or multiple descendant languages. That does not mean I won't be developing the ancestor language, but rather, I will design it to suit the modern language, rather than the other way around.
Without further ado, here is, in broad strokes, the modern phonology I have in mind:
|
|
Bracketed sounds are those I am undecided upon, and will choose whether to include them depending on the sound changes I employ to develop this language. Recalling what I said in my previous update, I intend to have one vowel alongside the typical five vowel system. This vowel's quality is undecided, and that is why there are multiple vowels in brackets, which suggest the various qualities I could choose for this vowel.
Vowel length is contrastive in the language, but I have not decided which vowels will have contrastive length. In terms of diphthongs, there will almost certainly be the front/back pair of /aj aw/, of which the first segment might have varied production, potentially as a merger of all rising diphthongs. Other diphthongs may be added, depending on the sound changes I decide upon.
I decided that I wanted a syllable structure that is constrained; what I meant by this is a language that has a maximal syllable structure of (C)(j,w)V(C), which doesn't even consider what constraints exist for any given consonant initially, in clusters, between vowels, or finally. The constraints I am going for are as follows:
I will work out these constraints in more detail later.
Now comes the challenging part, figuring out what the ancestral phonology will be. Recalling my guidelines, I require the following changes to occur:
So, from these guidelines, we can construct the ancestral phonology. Obviously, I have a lot of creative liberty, but I have some ideas for the ancestor language. Most conlangers call their ancestral languages proto-languages, but these are not reconstructed languages, but rather an original form that they develop into descendant languages.
This is my first attempt at the ancestral language's phonology. In world, this language is meant to exist around three thousand years before the modern language, so the number of sound changes will be quite large, though I expect them to be concentrated around time periods where language shifts occurred.
| labial | dental | alveolar | palatal | velar | uvular | |
|---|---|---|---|---|---|---|
| nasals | /m/ | /n/ | /ɲ/ | /ŋ/ | ||
| stops | /pʰ/ | /tʰ/ | /cʰ/ | /kʰ/ | /qʰ/ | |
| /p/ | /t/ | /c/ | /k/ | /q/ | ||
| /b/ | /d/ | /ɟ/ | /g/ | |||
| fricatives | /T/ | /s/ | /ç/ | /x/ | /χ/ | |
| approximants | /l/ | /j/ | /w/ | /ʁ/ |
|
|
/T/ is the only undecided segment in the inventory so far. This is some kind of dental fricative, though the exact production is unknown. I want this sound to merge with different consonants depending on the context, so the most likely production will be [θ], [ð] or [ɬ]. The variable production of the short close/mid vowels is a design choice to allow some phonetic variation, as otherwise, there would be no short close vowels. The close production is found in open syllables, while the mid production is found in closed syllables.
In terms of phonotactics, for my own ease of generating vocabulary and handling sound changes, I will say that this ancestral language has a very simple phonotactic structure of permitting any consonant in the onset, and only sonorants (except glides) are permitted in the coda. Visualised, this is:
| Onset | Nucleus | Coda |
|---|---|---|
| Optional consonant | Vowel | Optional sonorant, glide, vowel lengthening, or gemination |
Long vowels and diphthongs cannot co-occur with codas. However, I will say that geminates are permitted in the ancestral language with all non-approximant consonants, however, in this analysis /ʁ/ is a fricative, as it can be geminated. Synchronically, /l j w/ become /dː ɟː gː/ when preceded by a "geminating syllable".
Now, even though I call this the 'ancestral' language, I only say that in the sense that it is the ancestor of all living languages belonging to the family that the modern language belongs to. I will now describe my rationale for the phonological inventory I have chosen, and the developments that led to the inventory I have here.
Most clearly, there is an asymmetry between short and long vowels. This can be simply explained by the shift/merger of short vowels, where earlier /æ ɑ/ merged as /a/, while /e o/ merged as /ə/, or into /i u/ in word-final position. Before diphthongs this was even more extreme, as earlier /i u/ merged with their mid vowel counterparts, and backness was neutralised entirely for the first segment of each diphthong.
The palatal/uvular/velar distinction arose from these vowel mergers, as they would have originally just been the allophonic variations of velar consonants. /g/ became /ʁ/, as [ɢ] is unstable and rare crosslinguistically. Palatalisation occurred before ancestral /i(ː) e(ː)/, uvularisation occurred before ancestral /ɑ(ː) o(ː)/, while ancestral /æ(ː) u(ː)/ were neutral. Note that these features of [+palatal] and [+uvular] spread to adjacent velar consonants. However, some consonants resisted/blocked these changes: /w/ blocks both palatalisation and uvularisation, and /j/ causes palatalisation of neighbouring velar segments.
The gap of *ɴ is typologically common, and my way of analysing this sound's absence is to say that /ŋ/ when uvularised simply merged back into its velar counterpart, unless it was directly adjacent to another uvular consonant. All of these instances of [ɴ] can be analysed as /ŋ/.
Well, I'd say this enough work for today. I'll be back some time next week to share some more work. Next time we'll be considering diachronic sound changes between this ancestor language and the modern language.