292 Concluding Remarks: Pathways of Change
the correct way in which to view the rise of all incomplete sentences (cf. also
Trousdale 2012b), and there may exist a gradient from those which clearly
arise as subordinate clauses to those for which it is not possible to reconstruct
a subordinate- clause source.
10.3.3 Relative Clauses
Nominal relative clauses such as what’s more surprising/ serious/ strange/ sig-
nifi cant of all/ what annoys me constitute Quirk et al.’s ( 1985 : 1117) type (iii)
comment clause (see Section 1.4.1.4 ). Quirk et al. suggest that a sentence
with such a comment clause, e.g., what’s more surprising, he didn’t inform
his parents , “corresponds” to a cleft sentence ( It’s more surprising (that) he
didn’t inform his parents ) where “the relationship of subordination between
the two clauses is reversed.” It also corresponds, they point out, to a fi nal
sentential relative ( He didn’t inform his parents, which is more surprising )
and to a pseudo- cleft sentence ( What’s more surprising is (that) he didn’t
inform his parents ). Quirk et al. do not claim that these correspondences
represent any sort of synchronic or diachronic derivation. Sections 9.3 – 9.4
examined the history of the bare what’s more pragmatic parenthetical. As
with the matrix clause hypothesis , it is intuitively appealing to think that this
parenthetical may have developed from a main clause– subordinate clause
structure, such as a cleft or pseudo- cleft sentence, as hinted at by Quirk et
al. Also, as with the elliptical if- clauses discussed in the previous section, it
seems plausible that what more must have developed from a non- elliptical
structure. However, the historical data presented in Chapter 9 do not con-
fi rm either of these hypotheses. Both the full- cleft and pseudo- cleft sentences
with what’s more (Adj) are extremely rare and thus cannot serve as a source
for the comment clause. A comment clause which is more alternates with the
what’s more comment clause but dies out in the eighteenth century; it likely
has a different origin than the what’s more construction. The sentential rela-
tive (as in We serve food, yes, but also hope, which is more crucial, really.
[2013 Essence (COCA)]) appears as early as the seventeenth century, but
its form, function, and sentential position are all quite different from those
of the what’s more construction: it retains its adjectival complement; it is
almost always clause fi nal rather than clause initial; and it has a commen-
tary but not a connective function. The second hypothesis – that what’s more
derives from a non- elliptical structure with an explicit adjective complement
(e.g., what’s more important/ strange/ interesting ) – also fi nds little historical
support. Although the constructions appear simultaneously, what’s more Adj
has different morphosyntactic behavior than bare what’s more. In addition,
the OED’s defi nition of more as a pronoun in the sense of ‘something of
greater importance or signifi cance’ (often used predicatively) suggests that