558 JAMES K. WATTERS
so should have scope over the entire sentence in instances of core juncture.
Yet, in section 6.3 we found instances of core juncture in which it had scope
over only one clause. A solution that may be preferable to the one men
tioned in the last section may be to consider the layering of the clause in
regard to operators to be of a different sort than the layering of the clause
in regard to constituents. As Van Valin has pointed out (personal com
munication) there are two pictures of the layers of the clause given in
F&VV: besides the primary one of concentric layers there is also another
one of linear chaining. The former consistently applies to the discussion of
operators while the latter seems implicit in some of the discussion regarding
constituents. The data surveyed here suggests that there is, in fact, such a
distinction. Consider the following picture of the clause:
Operators: (Illoc. Force, Evid., Tense, Status (Modality (Aspect)))
PERIPHERY CORE NUCLEUS
Constituents: [Temporals, Locatives, Instr., etc. ] [NP(NP)(NP)][Verb]
Figure 2
In such a model, sentences such as that in (67) are not so troublesome:
the temporal can still be considered a constituent of the periphery, in fact of
a periphery that is only linked to the dependent core. The scope of the
clausal operators, however, must be over both clauses, as it is an instance of
core juncture.
In conclusion, though many of the details remain to be worked out,
such a model appears to be more adequate, though weaker in some of its
claims than the model given by F&VV: the scope of operators is concentri
cally layered over the nucleus, core and periphery while the layers of the
clause involving the constituents are related as links on a chain.
Notes
* I'm especially grateful to Karl Zimmer both for suggestions and for catching some
embarassing mistakes; and to Robert Van Valin, Sabahat Tura, Eric Pederson