Lao Syllables outside the Gedney Box

[Richard W]

I've been trying to make sense of the system of Lao tones. Ignoring the 7-tone systems (e.g. Xieng Khuang, Roi Et), there seems to be a 5-way contrast on live syllables not starting with a mid consonant, e.g.:

High consonant, no tone mark: [R]khaa 'leg'
Low consonant, no tone mark: [H]khaa 'stuck'
High consonant, mai ek: [M]khaa 'galangal'
Low consonant, mai ek: [M]khaa 'value' (same tone as above - but not in 7-tone Lao)
High consonant, mai tho: [FL]khaa 'kill; slave' (low falling tone)
Low consonant, mai tho: [FH]khaa 'trade'.

So far this is straight forward, though there is a lot of variation in the actual tones.

Now Enfield's grammar says that for live syllables starting with a mid consonant, there are only three possible tones:

Live consonant, no tone mark [L]kaa 'crow' (sometimes described as low rising - Enfield actually says that [L] = [R] in Vientiane speech)
Live consonant, mai ek [M]kaa 'black perch'
Live consonant, mai tho [FH]kaa 'bold, dare'

The grammar also says that the tones with dead syllables are also restricted according to vowel length and whether the initial consonant is mid:

High consonant, short vowel: [H]khat 'contrary'
Mid consonant, short vowel: [H]kat 'to bite'
Low consonant, short vowel: [M]khat 'to select'

High consonant, long vowel: [FL]khaat 'torn'
Mid consonant, long vowel: [FL]kaat 'cabbage'
Low consonant, long vowel: [FH]khaat 'hope'

Now all these combinations can be fitted into the 'Gedney box', a table for neatly summarising the regular tone correspondences between Tai dialects.

Is Enfield correct when he says that for the ordinary Lao of Vientiane, the combinations of initials, vowel lengths and tones have not expanded beyond what one sees in regularly developed inherited words? This would be quite different from Thai, where there is no constraint on the tone of live syllables and the three tones of dead syllables can all occur with long or short vowels and mid or non-mid consonants.

Now, Lao has tone marks mai ti and mai chatawa corresponding to Thai mai tri and mai chattawa, but there usage seems quite unsystematic and accounts of their meaning differ! It does seem that they are generally restricted to mid consonants, as in Thai. The SEAlang Lao dictionary incorporates Kerr's dictionary, which records tones, but the usage of mai ti and mai chatawa seems unsystematic. Kerr's dictionary distinguishes the tones I have marked as [L] and [R]. In so far as I can see any pattern, it is that:

Dead, mai ti seems to be the [M] tone.
Live, mai chatawa is equivalent to no tone mark (tone [L]) - which makes it completely redundant!
Dead, mai chatawa, long seems to represent the [H] tone.

However, it seems quite chaotic.

There is a second dictionary within the SEAlang dictionary, Patterson and Severino, but its tones have been derived automatically from the spelling.

I had been about to deduce that Lao in fact had a 5-way phonemic tone contrast, with the [L] and [FL] tones being variants of the same tone, realised as [L] after mid consonants in live syllables, but that doesn't seem to stack up.

What is going on here?