Unfortunately, I don’t own any Latin or Greek dictionaries or etymological texts, but I tried to research this topic on the internet. Here is what I found:
Perseus: words ending in “iasis” in L&S mostly have short -ăsis
- ĕlĕphantĭăsis, mydrĭăsis, phthīrĭăsis, hypŏcŏrĭăsis, sătyrĭăsis, sīrĭăsis, plătycŏrĭăsis, stĕnŏcŏrĭăsis, trĭchĭăsis
Oddly, all of them are transcribed with a short “a” in Latin except for “anathymiasis”.
“Short a” is also implied by the antepenultimate stress in the usual pronunciations of English words with this ending (e.g. English “satyriasis” = /sætəˈraɪəsɪs/), so this doesn't seem to be just a Lewis and Short thing.
OED: but they're from Greek words with -ᾱσις?
The Oxford English Dictionary (OED) has a entry that, if I’m reading it correctly, says that the vowel was long in Greek:
Etymology: < Latin -ăsis, Greek -ᾱσις.
Forming names of diseases, really nouns of state or process from verbs in -άειν (contracted -ᾶν); as from ϕθείρ louse, ϕθειρᾶν to be lousy, ϕθειρίᾱσις phthiˈriăsis; so elephantiasis, psoriasis, and many modern words, more or less analogical, as allantiasis, arseniasis, etc.
There is also the following note in the OED about a historical pronunciation of “myiasis” (not in Lewis and Short) with penult stress:
N.E.D. (1908) gives the pronunciation as (məiˌiˌēi·sis) /maɪɪˈeɪsɪs/.
The length of the alpha in some of the above words in Greek seems to be supported by some Ionic spellings with η listed in LSJ on Perseus: “μυ^δ-ίησις” (for mydriasis) and “σα^τυ^ρί-ησις” (for satyriasis). (Wikipedia tells me that “Proto-Greek ā > Ionic ē; in Doric, Aeolic, ā remains; in Attic, ā after e, i, r, but ē elsewhere.”)
What is the Greek etymology of this apparent suffix -ιασις? I think that I recognize the last part as the nominalizing suffix -σις from PIE *-tis. But I'm confused by the -ια-. The OED says verbs were a step in the derivation of these words, and gives as examples psoriasis (ψωρίασις) and phthiriasis (ϕθειρίασις). But the corresponding verbs are apparently ψωριᾶν and ϕθειρᾶν: one has an iota before the ᾶν and the other doesn't. Why does ϕθειρᾶν turn into "ϕθειριασις" instead of "ϕθειρασις" (if the OED's explanation is right)?
Was the alpha long in Greek? I found a few pages through Google Books that seem to say that in general, Greek nouns ending in ασις could have short /a/ or long /aː/, with a rule for choosing between them based on the conjugation of the verb from which they are derived (long when derived from verbs in -αω; short elsewise) (Introduction to Greek Prosody, Peter Wilson; Practical rules for Greek accents & quantity, by Philipp Buttmann). Unfortunately, the citation forms used by the OED for ϕθειρᾶν and ψωριᾶν don't end in -αω, and I don't know how to conjugate verbs in Greek so I can't tell if they have this ending in other forms.
Why are most of these words pronounced with "ă" in Latin, except for "ănăthȳmĭāsis", according to Lewis and Short?
I'm really stumped. I did come up with one tentative hypothesis, which is that these words were late borrowings that were stressed on the antepenult after the Greek (like certain nouns in -ia were stressed on the penult after Greek in Romance languages; e.g. Spanish filosofía < Gr. φιλοσοφία). If that's the case, maybe it would be better for L&S to say that they belong to a period of Latin after vowel length began to be lost, rather than transcribing the "a" with a breve? But I'm not confident about this. I'd really appreciate an explanation if anyone knows of one.