theta roles and ergativity in a fairly isolating new language [longish]

First off, Hello, my name is Joe Mondello and I have lurked on this list for
years.  I am writing today because I have a dilemma.  In my current conlang,
I have been grappling with the question of intention.  I will use the word
súdo "find" as an example.  I wished to distinguish between finding through
the process of searching (i.e. subject as agent) and finding by chance
(subject as experiencer).  The structure I have come up with is essentially
Agent-Verb-Experiencer-Theme:

nu súdo-ka    pi    bráme.
1s find-Perf. Det horse.
I've found the horse (by searching, in the process of looking for it).

súdo-ka    nu (pi)    bráme
find-Perf  1s  (Det) horse
I've happened upon a(the) horse

The matter becomes fairly confused when dealing with subordinate clauses:

ki nu [(nu) súdo pi brame]
ki      nu súdo pi    brame
hope I    find   Det horse
I hope I find (by looking) the horse

ki      mo       nu-z      [nu súdo zúyo]
hope mother 1s-Gen [1s  find  work]
My mother hopes I find work


Thus far I think this is fairly reasonable.  The hairy part is that when the
subject of the subordinate clause is the experiencer, the subject and verb
of the main clause follw the verb of the subordinate clause (follwing
this?), e.g.:

súdo-ga  ki       mo slívil-o
find-Fut  hope 1s   money-PL
I hope I will find/happen upon a substantial sum of money.

Is there anything similar to this structure?  It comes from the
proto-languages treatment of the intransitive verb, which i think may be
considered ergative. the subject of the intransitive verb follows the verb
like the object of a transitive verb, but there is no overt marking, so we
have a situation as follows:

[protolanguage]
        no sutu bradmay
        I    seek horse
        I am looking for a/the horse

        ta   bradmay
        big horse
        the horse is big

from this structure, more complex ones came about:

        du-ko         no [ta bradmay]
        cause-Past 1s  [big horse]

        bradmay ta-go
        horse      big-be
        the horse that is big

   Which led to intransitive verbs becoming causatives and relative clauses
becoming adjectival verbs:

[daughter language]
        nu  ta-m               pi   brame-vo nu-z
        1s  enlarge-Prog Det horse-Pl  1s-Gen
        I'm fattening up my horses

        brame-vo nu-z      tayo
        horse-Pl   1s-Gen big
        My horses are large


This type of language is somewhat new to me from a design point of view, so
I look forward to any feedback.

Joe Mondello