226 MARK HANSELL
There is no way other than a Complement of Result to unambiguously state
that an action has been carried out to an end result. That is, even a word
like shä (usually translated "kill") does not mean categorically that the
object of the verb has died, and the only way to ensure that interpretation
is to use the CR construction shä sí ("kill+die"). Shä should perhaps be
better translated as "to assault murderously", and by analogy to (47b) it can
be hypothesized that the logical structure of shä si is as follows:
(51) Wo shä sí tä.
I "assault" die him
"I kill him."
Logical Structure: [assault' (I, him)] CAUSE [BECOME dead'
(him)]
The existence of this abstract operator CAUSE might be doubted, as is any
abstract entity that never appears at the surface level, except for the fact
that traces of it DO appear at the surface level in the Potential Complement
[PC].
This paper has largely ignored the PC so far because most of the gross
structural properties of the CR are shared by the PC. The PC is in a sense
parasitic on the CR, in that the PC can be seen as a CR with bu or de
infixed between V\ and V 2. I will propose, however, that the bu or de of the
PC in fact does not have scope over either of the verbs in the CC, nor over
any other lexical material, but rather over the abstract operator CAUSE.
Consider the following examples:
(52) a. Wo kán döng le. Complement of Result
I read understand ASP
"I understood it (by way of reading)"
b. Wo kán de döng. Potential Complement
I read PART understand
"I can understand it (by reading.)"
Wö kán bu döng.
I read NEG understand
"I can't understand it (by reading.)"
Li and Thompson (1976) state that the NEG marker bu in a PC like (52c)
has scope over only the V 2 , literally that "I read and didn't understand".
This is a mistaken analysis, because while (52a) asserts that I actually read
the book (i.e. asserts that both "read" and "understand" have a positive