69
2.7 Development of Whatit is not completely independent syntactically) and it is either surrounded or
followed by pauses. Because it must join with preceding material to form a
prosodic phrase, it cannot occur in sentence- initial position. They do not refer
to what in this function as a “pragmatic marker,” but I would classify it as such
(cf. Brinton 2006 : 76 on a similar use of say ).
2.7 The Development of What
Given its multifunctional nature, what has a complex history and the fol-
lowing account must be taken as, in part, speculative. In broad terms, what
seems to follow a course of development similar to that proposed by Traugott
( 1982 : 254– 255) for why and where , from interrogative in direct questions to
complementizer in indirect questions to pragmatic marker (see Chapter 1 ).
The interrogative function of hwæt in independent questions – i.e., the ref-
erential function – exists in the earliest Old English.^36 Hwæt serves as an inter-
rogative pronoun (the neuter singular nominative/ accusative of hwa ) meaning
‘what’ (OED: s.v. what , defs. A 1– 4) and as an interrogative adverb meaning
‘why, how, how much’ (OED: s.v. what , defs. A 19– 21; BT: s.v. hwæt ; Mitchell
1985 , I: 486). The adjectival use develops somewhat later from the partitive
genitive construction (OED: s.v. what , def. A 13a). The interrogative function
in dependent questions also exists in the earliest Old English (BT: s.v. hw ā ,
def. II), with the pronominal use (e.g., I asked what he wanted ) again predat-
ing the adjectival use (e.g., I asked what time it was ).^37 While in Old English,
at least, this form is not generally treated as a full- fl edged “conjunction,” it
clearly has conjunctive force in subordinating the dependent question to the
matrix clause. The complementizer function can be seen as belonging to the
textual component.
It has also been suggested that hwæt functions as a generalized “par-
ticle of interrogation” preceding sentences with declarative word order
(in Old English) and later interrogative word order (OED: s.v. what , def.
A 21; BTS: s.v. hwæt , def. I 2; Schmidt 1874– 75: s.v. what , def. e; Visser
36 In an aside, Bammesberger ( 2006 : 5) observes that exclamatory hwæt may not derive from the
interrogative at all, but may simply be homophonous with it as a result of phonological merger.
While Walkden ( 2013 : 476) admits that accidental homophony “cannot be ruled out,” and that
origins in OE hwæt ‘quick, active, vigorous’ or hwettan ‘to whet’ are possible, the use of inter-
rogatives in an exclamatory sense cross- linguistically supports an interrogative origin.
37 It is often diffi cult to distinguish a dependent question (e.g., I asked what he wanted ) from a
free relative clause (e.g., I knew what he wanted ), where what means ‘that which’ (see note
in OED under what entry; also Mitchell 1985 , II: 65– 68), and from a dependent exclamation
(e.g., We were shocked at what crimes he had committed ) (see Mitchell 1985 , II: 64– 71; Quirk
et al. 1985 : 1055– 1056), but the distinction is unimportant for our purposes. It is generally
agreed that what could not serve as a defi nite relative pronoun in Old English (Mitchell 1985 ,
I: 142, 207; 1985, II: 66– 68, 223).