> >     First off, what's the opposite of tonogenesis?


> pitch accent is realised as a rise in pitch,
> but the rules to position this rise of pitch
> can be formulated in terms of accented syllable.
> Basically, instead of a rise in volume you have a rise in pitch..

Instead of "instead", I would say that
the stress and the rise in pitch go always together.

> Pitch accent can also have rules for propagation of the high tone
> (meaning that the high tone can be put on more than one syllable,
> unlike a stress in a stress accent system), but those are secondary.
> So basically, in a pitch accent system there are two pitches:
> high and low, and in one word only one syllable
> (or mora, or whatever the prosodic unit is in that case)
> receives a high pitch
> (just like in a stress system only one syllable receives the stress).
> The high pitch may also propagate after the stressed syllable
> (like in Japanese), even beyond the word limit itself,
> but the main phenomenon is the rise of pitch, not its lowering.

From another point of view, it is the lowering what matters,
because the rising takes no time.
We coud say that there are two *pitchemes:
_falling_ on moras with word-stress and
_flat_ on unstressed moras.
The falling *pitchem has three *alopitches:
_high to mid_ before the sentence-stress,
_high to low_ on the sentence stress,
_mid to low_ after the sentence stress.
The flat *pitchem has two *alopitches:
_low_ before the first stress,
_mid_ from the first stress to the sentence-stress,
_low_ again after the sentence-stress.


The default place of the sentence-stress is at the end:


(Actually, all moras are more or less falling,
but the more they are accented,
the more rapid their falling is. )

> On the other hand, in a tone system,
> each prosodic unit receives a distinct tone,
> a single word can receive more than one high pitch,

It can be compared with vowel length:
in some languages, only the stressed sylable can be long,
while in other ones, any sylable can be long.