1.4 Pathways for Development 21semantically bleached and grammaticalized, serving a discourse (mitigating,
politeness ) function; they are prosodically integrated and lack stress. Kaltenböck
( 2011 ) shows that I think that and I think Ø show similar pro sodic patterns in
his corpus, with the occurrence of that in fact correlating with reduced prosodic
prominence and secondary function. He argues that that in these cases is not
functioning as a marker of subordination but rather as a fi ller.
The matrix clause hypothesis is explored in case studies in Chapters 5 (epi-
stemic parentheticals), 6 ( I admit ), and 7 ( I’m just saying ).
1.4.1.3 From Clausal Construction to Pragmatic Marker: Adverbial
Comment Clauses. Pragmatic markers that are “adverbial clause/ clausal
adjunct (fi nite)” parentheticals, or type (ii) comment clauses resembling fi nite
adverbial clauses ( Quirk et al. 1985 : 1115– 1117), e.g., as you know , as you
say , as I can see , cannot be accounted for by the matrix clause hypothesis.^23
The syntactic derivation of adverbial comment clauses is presumably simpler
than the one proposed for main- clause- like comment clauses, as it involves no
reversal of syntactic hierarchy. Adverbial clauses begin life as adjunct adverbi-
als with narrow syntactic scope within the predicate, and much like the process
of scope and semantic/ pragmatic expansion described by Traugott ( 1995a ; see
above), come to have wider scope as a disjunct adverbials (often in sentence-
initial or fi nal position). In this position they may undergo semantic change,
acquire pragmatic functions, and have scope over larger segments of discourse;
i.e., they become pragmatic markers. In some cases, the adverbial connectives
are lost.
An early study looking at an adverbial comment clause is Allen ’s argument
that the politeness marker please originates in the adverbial clause if you please
( 1995; see also Chen 1998 : 25– 27). Traditionally, please has been assumed to
originate in the impersonal construction please it you ‘may it please you’ >
please you > please (OED: s.v. please , adv. and int.; cf. please , v., def. 3b).
Such a structure would suggest the existence of a subordinate that- clause (sub-
ject) and argue for a version of the matrix clause hypothesis. However, the
OED allows that please may also be seen as a shortened form of if you please
(s.v. please , adv. and int.; cf. please , v., def. 6c). Allen notes that the personal
construction if/ when you please does not arise through reanalysis of the imper-
sonal if it please you ; rather, the two constructions exist independently and are
already clearly differentiated in Shakespeare’s time. The impersonal becomes
recessive and is lost, while the personal form is routinized as a polite formula
and ultimately (at the beginning of the twentieth century) shortened to please.
23 Quirk et al. ( 1985 ) consider only as , but not the other adverbial connectives, such as for , so ,
and especially if , which may introduce such comment clauses.