554 JAMES K. WATTERS

same level; i.e., the operator tense, though a clausal operator, is not neces­

sarily tied to the temporal modifier of the clause. The tense operator and

the temporal adverbial in such constructions do not appear to be structur­

ally linked. I'll return to this matter in the conclusion. The second problem

is more methodological rather than posing a clear counter-example to

F&VV's claims. In instances of core juncture such as that above, we expect

the nuclear operators, such as aspect, to be potentially independently

specified for each verb. The fact that the progressive aspect cannot occur on

a -mEk infinitive is apparently a detail of Turkish morphology. But as men­

tioned above, the morphological position of the aspectual suffixes present a

counterexample to another claim of F&VV's — that the layering of mor­

phemes will reflect the layered structure of the clause. The result is that to

independently specify aspect within the dependent clause, an additional

juncture is required:

{

-vyor ol-mak ist-iyor-um

next month work < become-iNF want-PROG-lsG

(^ -AOR

"Next month I want to be working."

Thus, the infinitival form is the verb olmak "to become" which itself is

linked to the preceding verb (involving one of the Turkish linkage types not

explored in this paper). This latter verb can be independently marked for

aspect.

7. Core subordination

Instances of istiyor which, unlike that above, involve noncoreferential sub­

jects do manifest subordinate linkage: the dependent clause fully manifests

the characteristics of a clausal argument.

(56) 1ιş---ι ist-iyor.

WOrk-VN-lSG-ACC Want-PROG

"S/he wants me to work."

(57) Kitap oku-ma-n-ι ist-iyor.

book read-vN-2sG-Acc want-PROG

"S/he wants you to read a book."