Notes 235
(^11) According to Brazil (1997) // the QUEEN of hearts // indicates that hearts is
projected as recoverable from the previous context. Here the lexical element
queen realizes an independent selection.
(^12) The process by which, in the history of a language, a unit with lexical meaning
changes into one with grammatical meaning (Matthews 1997: 151).
(^13) Caution is needed when considering the Spoonerisms (32) and (33). Potter
(1980: 30) conjectures that Spooner’s individual style of speech may have been
due to a cerebral dysfunction. He further speculates that Spooner’s condition
may not in fact have been unique. It was simply Dr Spooner’s exposed academic
position which highlighted his condition. Anderson (1990: 337), on the other
hand, suggests that Spooner’s style of speech may have been due to deliberate
attempts at humour by Dr Spooner himself.
(^14) The intended words are given in italics immediately after the word containing
the slip of the tongue.
(^15) Table 4.1 is an adaptation of Table 8–1 in Carroll (1994: 192) which included
all eight types of speech errors The remaining four classes Addition, Deletion,
Substitution, and Blends occur internally within orthographic words, and so cannot
shed any light on the issue of the extent of single lexical elements, and so are
excluded here.
(^16) Or perhaps the whole phrase getting your nose remodelled is treated as a single
meaningful unit.
(^17) This claim is neutral as to whether the chunk is stored at a single address in the
mental lexicon or assembled by speakers into a meaningful chunk prior to its
articulation as a single meaningful chunk.
(^18) To ensure methodogical consistency such instances will only be recorded if they
are so notated in Cobuild which has been chosen as the arbitrator of whether or
not word-like elements coalesce into larger elements because it is based upon
extensive corpus research.
(^19) The elements contained within the angled brackets were ellipted.
(^20) Brazil (1995: xvi) states that the diacritic & is used to code and and so. However,
on page (216) he somewhat oddly codes but with &. Section 4.5 p. 111 discusses
how to code linking elements such as but.
(^21) Brazil (1995: 216) coded just happens as a V element in (37) though a more
accurate coding would appear to be to code just as a suspensive adverbial
element. He codes and so as a single element in (38).
(^22) An alternative and intuitively satisfying defi nition of ellipsis is that it is the covert
realization of a word or words (Hudson 2006: 178) which entails that all ellipted
elements are ordinary words. This raises the possibility that a more delicate gram-
mar coding could easily devise a coding where the part of speech of the ellipted
element is included in the description.
(^23) The situational ellipsis of the lexical element I appears to mandate the ellipsis of
have as the utterance have got a cold appears unlikely possibly because it seems to
carry unwarranted and inappropriate interrogative implications.
(^24) In the interests of simplicity and because of their irrelevance to the present
discussion; key and termination selections have not been transcribed. The gram-
mar coding was not present in Brazil (1997).
(^25) This example presupposes that the pause is not a deliberate strategy aimed at
manipulating the hearers’ expectations.