What I have done:
- turned all < y > into IPA [j]
- turned all < ' > into IPA [?] glottal stop
- included missing data (I don't why my spreadsheet had so missing cells!)
- deleted languages with very few tokens
- eliminated all parentheses, I am using a hyphen "-" as morpheme break (very few polymorphemic forms)
- nasals: some vowels have ~ and the data also has < m > and < n >. The problem might be with words that are "all nasal", which I represent phonemically by placing ~ before the whole word. You will probably need to make tell the computer to make the following changes
"make all vowels following ~ to be nasalized" "make all voiced sounds to become nasal sounds b > m, d > n, g > ŋ , r > r̃ > , j > ɲ >, w > w̃,
-
aspiration: if an immediately follows or precedes a consonant, with now intervening vowel, it should be interpreted as part of the consonant, e.g [hp] or [ph]
-
palatalization: if a < j > immediately follows o a consonant, with now intervening vowel, it should be interpreted as part of the consonant, e.g [tj]