Tsakonian orthographic reform

If you're Hubert Pernot, the great hellenist whose grammar of Tsakonian is a Neogrammarian masterwork, Tsakonian gets written in some bizarre adaptation of a bizzare French dialectological alphabet (Gillieron-Rousselot, which seems to have evolved into the French Romance transcription.)

If you're Agathocles Haralambopoulos, former prof at Aristotle Uni who did his doctorate on Tsakonian phonology, you use the IPA, and call Nick's favourite phoneme /tsʰ/ instead of /tɕ/, thereby not making Nick your fan.

If you're any other Greek linguist, and particularly the native speaker Thanasis Costakis (who was Pernot's main informant), you use historical Greek orthography, with any diacritics your printer allows (as is normal Greek dialectological practice)—and different printers will allow different diacritics. So it'll be κ̔ τ̔ π̔, σ̌, ρ̌, λ̣ ν̣, λ̑ ν̑, and τσ̕ or τσ̑.

If you are Michael Deffner, enthusiast who stayed in Tsakonia for fifty years but did not become any better as a linguist as a result (Pernot has some gallically charming things to say about his phonology), you have even more diacritics than you actually need, because you haven't quite grasped that whole phoneme concept.

If you're a lay Tsakonian—as happened in the few times in the 19th and early 20th century when Tsakonians wrote Tsakonian—you just use the 24 letters of the Greek alphabet, and a *lot* of digraphs. The system is pretty much laid out in Wikipedia (citing me, and with my corrections). So σχ is /ʃ/; ρζ is /r̝/ or /ʒ/ (socially conditioned); πφ τθ κχ is /pʰ tʰ kʰ/; λλ and νν are /l/ and /n/ before front vowels, where you'd expect /ʎ ɲ/. (That one's subtle, and much less often complied with. In Tsakonian /ne/ goes to /ni/, and /ni/ goes to /ɲi/.) [c] is just κ, which occurs before front vowels as in standard Greek (but is in fact a reflex of palatalised /p/ or /t/: κίνω /cino/ < πίνω /pino/).

The reflex of palatalised /k/ is not [c] in Tsakonian, but what Haralambopoulos transcribed as /tsʰ/, and what I'd like to think is closer to /tɕ/. (Alas, even though Pernot actually did some pioneering instrumental phonetics, and Haralambopoulos did spectrograms too, it's hard to tell.) Lay practice was to render this as τζ. So κήπος [cipos] went to τζήπε [tɕipe].

The problem with τζ is, Modern Greek speakers will read it as /dz/, and there are already /dz/ in Tsakonian. (Admittedly they're overwhelmingly /ndz/, but a couple of Turkish loan words are all you need to throw that off.) I've long thought that, if Tsakonian keeps getting written in lay transcription, that τχ [tç] would be a less problematic rendering of /tɕ/. And by its resemblance to the digraph tx, it would finally establish a connection between Basque and Tsakonian! :-)

So τχήπε instead of τζήπε. There, I've said it, now I just wait for the Googlers to find it...