SenorPez wrote: ↑Thu Jan 24, 2019 3:58 pm
I'm trying to develop a phonotactic structure to build syllables that, while not English, aren't as strangely alien.
What you define as alien is a matter of taste. Nuxalk (a real-world language) has words like
clhp’xwlhtlhplhhskwts’. Looks alien? And yet it's a real word spoken by real people. Which is why I am asking for more specifics on what you're trying to achieve.
As an example, lättom (stadium, in Verdurian) is a great example of a word that, while not English, might still be a terrestrial language. I've had some luck (gasä for example), but as soon as I try to add a bit of complexity for syllables that aren't CV or VC, it ends up looking like I just punched my keyboard to create a word.
First off, there is nothing wrong with only having CV or VC syllables. If you want to create more complex syllables, it's a good idea to have fixed rules on what consonants can cluster together, and to have rules for resolving illicit clusters if they appear due to inflection or something.
Sonority hierarchy is a good start.
A random example of resolving illicit clusters. Let's say you have a prefix
k- meaning "inside", and a noun
pag meaning house. So *k-pag would mean "house interior". But you don't want two stops in the onset because they are of equal sonority. You could do one of the following (among other things):
- lenite the second stop to a fricative so it's more sonorous: *k-pag -> kfag
- insert an extra vowel to avoid the cluster: *k-pag -> kapag
- decide that you allow long (geminate) consonants, and the prefix assimilates to a following stop: *k-pag -> ppag
In real languages, there are a lot more rules like this. A lot of them arise due to sound changes over time, and can be hard to replicate without simulating at least a bit of a language's history. But this is a start.
When you say "unstable inventory," do you mean balance between voiced and unvoiced, or balance between articulations? I have no real problem moving the θ and ɖ back to t and d, for example. The main design goal (aside from putting the LCK to use) from a phonetic standpoint was to reduce the number of phones (as well as, for reasons peripheral to the conlang, having a consistent phone-to-symbology process).
What I mean that "gaps" in a phonemic inventory should make sense. You have plenty of alveolars. You also have both voiceless and voiced stops at other points of articulation. So why no alveolar stops? This creates a gap and a real language would gravitate towards filling it. That's what I mean with "unstable". /θ ɖ/ would be likely candidates for becoming /t d/ instead. Or maybe /s z/ would become /t d/ and then /ʃ ʒ/ would become /s z/ to fill that new gap.
Having some gaps is fine but they need to be justified. /t/ is one of the most common consonants. It is missing only in a few real languages, and these languages have very few consonants in general. So not having /t/ is very weird in itself, and completely implausible in such a large inventory as yours.
(Another thing is the sole retroflex /ɖ/. It looks like the odd one out here. A retroflex rhotic would be less jarring perhaps as a one-off. But it honestly looks like you added it for the sake of having something interesting without a real plan behind it. If you do want to keep retroflexes, I'd recommend adding at least /ʈ/, as well.)
