More MNCL5 Problems

I discovered a couple more problems. Hopefully, the two main issues are of
general interest.

11. Embedded WH-Questions and Relative Clauses
An embedded WH-question (and the normal variety of non-embedded WH-
questions, for that matter) begins with an interrogative pronoun; this is either
koi or a word form beginning with the k- morpheme. A relative clause begins
with a relative pronoun, which begins with the y- morpheme. Often, this isn't a
problem, due to the fairly free order of phrases and phrase components and
because required possessors/objects precede their heads. Here are some
examples (using non-embedded questions for simplicity).

Koi hote tafo java? -- "How hot is your coffee?"
Kanok ta vid'he? -- "What (thing) have you seen?"
Kamak dok zan disek? -- "Whose mother told them that?"
Kilko hauso cabi korek zo hundak? -- "What kind of house did the dog run to?"

The problem occurs when the interrogative or relative pronoun would be
contained within a subordinate clause. English pulls the pronoun (or the phrase
containing the pronoun) out of the subordinate clause, as in "What do you
want us to give you?" where "what" is really an object of "give", not "want"
(Note: the English infinitive construction corresponds to an MNCL5 complement
clause). I don't see this working for MNCL5. Possibly some kind of "such that"
construction could be used, with perhaps a 3rd person pronoun remaining in
place; I haven't figured out the other details yet. I wonder what other
strategies might be used?

12. Order of Medials
I've tried to arrange things so that all inflectional medials follow all derivational
morphemes. However, I've come across a possible exception. It seems to make
more sense for -apt-, which means "likely to" and is clearly derivational (it
changes the argument structure), to be applied to a stems ending in -es-,
which patterns with the aspect medials, as in:

Marxok disesapte. -- "Marsha is talkative." (literally, "likely to continue talking")

Another possible exception is mentioned below.

Regarding the earlier problems, I've made a little progress.

3. Trivalent Imperatives
5. Conatives
I haven't made a decision on these yet, but I think they're related: both
involve making a distinction between two arguments. So, either the -m- or
the -s- (grammatical voice medials) could be used with the conative marker,
whichever is used with the imperative. This brings up another issue, however:
must the conative medial precede the grammatical voice medial or can it
follow?