Large Inventories

[thethief3]

I don't often see conlangs with large consonant phoneme inventories say 60+ phonemes. The one with the most i ever made was a collaborative project proto southern with a grand total of 37 phonemes. Personally i trend to more minimalist languages but i want to see some maximalist ones.

[Knox Adjacent]

Ya know, I had a Project Clicklang designed for maximum fun for a logography. It jived well with my fondness for minimal pairs and figuring out rebuses. Total of 102 consonants with last revision. The size is due to a voicing distinction overlaid with a tenuis/aspirated/glottalized distinction (yes, it involved voicing contours to allow the secondary phonations) and click-nonclick contours.

Vowels were also wacky with an incredible system of tones, but alas unasked for.

ɟ͡cˀo˧ǀ̬͡ɢo˦ "thunder" seems representative

[DesEsseintes]

My conlang TLFKAT had several iterations where the consonant segments numbered nearly a hundred. I believe this version contained 101 consonants?

I should go back to working on it.

[Omzinesý]

I find it difficult to control such a system. My languages with large inventories are few and last short.
Mhilva https://conlang.fandom.com/wiki/Mhilva is maybe the best attempt.

[WeepingElf]

I find it difficult to control such a system. My languages with large inventories are few and last short.

Yes, large inventories are difficult to handle. Many years ago, there was a guy on the CONLANG mailing list who worked on a language named Tech. One of the reasons why he made little progress was that he was using gargantuan phoneme inventories - one version had 327(!) consonant phonemes. (Another problem was that he used Nostratic as a starting point.) I prefer phoneme inventories of moderate size. Huge inventories don't necessarily lead to interesting phonologies.

[Nortaneous]

I have one conlang that could be analyzed as having 60 consonants - my usual analysis has 30, but labialization being preserved through Cə~CV- reduplication means doubling that could be defensible.

kø̀ɶ̯tsʌʔ > kœˈkø̀ɶ̯tsʌʔ /kʷə-kʷea̯ɦtsʌʔ/
iχsxœtɬuɔ > ɪʔiχsxœtɬuɔ /ɨ-ʔiχsxʷœtɬʷɯʌ/
tsʼwàpœn > tsœˈtsʼwàpœn /tsʷə-tsʷʼaɦpʷən/

There are problems with this analysis, though - certain rounded vowels don't have unrounded equivalents and vice versa. So the usual analysis has 31 vowels and 30 consonants:

/p pʼ b t tʼ d ts tsʼ dz tɬ tɬʼ dɮ k kʼ g qʼ ʔ/
/ɸ β s z ɬ l x ɣ χ ɦ/
/m n ŋ/

/i y u/
/ɪɛ ʏœ ɯʌ uɔ/
/ɛɪ œʏ ɔu/
/e ø o/
/eæ øɶ wʌ ʌ ɔ/
/æ wæ a wa/
/ɪu ɛu oɪ/
/ɛ̆ œ̆ ə̆ ŏ ɪ̆ ʏ̆/

Diachronically, front rounded vowels and wV diphthongs are from CwV sequences, which helps the analysis, but you'd probably have to say that there are vowels before which only rounded or only unrounded consonants can occur.