Rob wrote:

"What's a "reasonable" amount of use for a pitch-accent? Or, when does the
use of pitch-accent start to feel too unnatural or too bothersome?

"Vilani apparently has six 'tones', which I take to mean six pitch-accent
tonal patterns (or uses, even); perhaps the pitch-accents are tonemes.  For
example, nouns could be case-marked using pitch, as in

"Shar'ik.      <= agent
"Shar.ik.      <= patient
"Shar.ik'      <= dative/benefactive obj

"And verbs could be aspect-marked using pitch:

"' Shar.ik.   "Sharik is bothersome."
" Shar.ik.   "Sharik is becoming bothersome."
"Na'su. Shar.ik.   "Sharik is becoming less bothersome."

"Na.da'su. Shar'ik.   "Sharik bothers Eneri."
" Shar'ik.   "Sharik is beginning to bother Eneri."
"Na' Shar'ik.   "Sharik no longer bothers Eneri."

"My question is: how much is too much?  How much farther can this be
stretched out?  Or is it too stretched already?

"I guess my root question is: given that this is intended to be a human
language, what is a reasonable limit of the amount of semantic information
that tonemes ought to carry?"

Interesting question. I guess you could survey tonal languages and find an
average of how many tones each has and draw conclusions about them.

I don't think one single toneme should carry a lot of meaning, because the
meaning will become too ambiguous--it's as if you were to use a suffix /-an/
for, say, accusative case, augmentatives, negation, and plural. Especially
if this were in a language like English, where there are words like 'walk',
and you don't know its part of speech right away, it would be very difficult
to understand the language.

My engelang Kel uses three tones, rising, mid, and falling, as well as
length, nasalization, and rhotacity. Each vowel type is used for a different
grammatical function (and affixes are also used), for example valency is
marked with tone, opposite is marked by lengthening the vowel,
nominalization is marked by nasalizing the vowel, agentive nouns are marked
with rhotacity. The basic monophthongial system is /i, e, a, o, u/, so there
isn't any ambiguity; these vocalic "inflections" are only grammatical. I
chose three tones because I had also decided to use other vowel types in the
grammar; if I had chosen any more, I would have had too many to work with.