I saw this question in reddit, which I found interesting. I quote:
“Why is it that the antepenult could be accented when the penult is long and the ultimate is short but not the other way round?”
I don’t know the answer, but I’m sure some of you do. It is interesting because the two cases have the identical in the count of morae, right?
It goes back to the “rule” of contonation:
You can only have one mora after contonation (rising tone followed by return to standard tone on next syllable), so if the ultima was long, you would have two morae following contonation. Another way of putting it: when the antepenult is accented, contonation occurs in the antepenult and penult (tone rises on antepenult, returns to standard on penult, whether the penult is long or short);after this you can only have one mora, which by definition is a short syllable. Mastronarde has a nice tutorial on this:
http://atticgreek.org/accent/contonation1.html
And if you really want to dig into this:
https://www.yumpu.com/en/document/fullscreen/12201185/vox-graeca
Have a look at Chapter 6, Accent.
I think this is the real question: how can the length of the penult have no bearing? This is where the super stars of the board may have to step in, but I’ll venture the following: contonation involves intonation of syllables, whereas morae are a measurement of vowel length, which may or may not determine syllable length.