10.3 Clausal Sources of Pragmatic Markers 289the matrix clause hypothesis – is tenuous at best. Rather, Old English contains
fi rst- person know - verbs in adverbial clauses following þæs (þe) and swa ‘so,
as’ which may be glossed ‘so/ thus/ as I think’; these continue to be common in
Middle English. Therefore, I argue that the rise of parentheticals can be traced
back to the existence of these adverbial clauses with as or so. The derivation of
parentheticals is simpler than that postulated by the matrix clause hypothesis
since it does not involve a reversal of syntactic hierarchy (the adjunct adverbial
clause already has a degree of independence) but merely deletion of an adver-
bial connective (in those cases where the logical relation between main and
adverbial clause expressed by this connective is contextually inferable).
Less detailed examinations of the matrix clause hypothesis occur in Chapters
6 and 7. In Chapter 6 , the rise of pragmatic parenthetical I/ you (modal) admit
is shown to support the matrix clause hypothesis, with the matrix clause I/ you
(modal) admit (that) S antedating the parentheticals. The development of the
parentheticals represents a direct line from the procedural uses (performative
and non- performative) of admit. There is no evidence that adverbial structures
such as as I admit contributed to the rise of the parenthetical, but the more fre-
quent and earlier passive construction it be admitted construction may have
provided support for the rise of the parenthetical through merger of the two
constructions, though the syntactic mechanisms involved in this merger are
not clear. In Chapter 7 , a number of different sources for I am/ ’m just sayin(g)
are considered, including the full structure with a complement that - clause, an
adverbial clause with as , a conjoined structure ( I’m not Ving, I’m just say-
ing ), I’m saying without the adverbial, and I just/ only say. The historical data
suggest that the only plausible source for I’m just saying is the main- clause
structure with a clausal complement, thus conforming to the matrix clause
hypothesis. However, the evidence is rather weak, as I’m just saying that/ Ø S
is not particularly common before the period that sees the appearance of the
comment clause.
The studies here (as well as in Brinton 2008 ) thus point to the validity of the
matrix clause hypothesis in the case of some comment clauses but also suggest
that in the case of other comment clauses we must seek other origins.
10.3.2 Adverbial Clauses
As noted in Section 1.4.1.3 , an adverbial source for many pragmatic parentheticals
is immediately obvious because of the continued presence of the adverbial con-
nector: if you will , as it were , if you choose/ like/ prefer/ want/ wish , so to speak/ say ,
if you ask me , as you see. These adverbial clauses are often elliptical, as they may
be missing a necessary complement; for example, the verbs choose/ like/ prefer/
want/ wish are all standardly followed by an NP or clausal complement, which is
absent in the parenthetical. Furthermore, in the case of an if- clause, it is also missing