Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
A History of the English Language (Hogg).pdf
Скачиваний:
307
Добавлен:
01.03.2016
Размер:
9.85 Mб
Скачать

158 O L G A F I S C H E R A N D W I M VA N D E R W U R F F

concerning the general rise of periphrastic constructions), so it cannot come as a surprise that this order became the preferred one in negative sentences and thus, as it were, forced do into the negative clause pattern.

Empirical evidence for the idea that the awkward place of adverbial not played a role can be found in Elleg˚ard (1953: 195), who notes that the use of do in negative sentences is consistently higher with transitive verbs than with intransitive verbs, thus stressing the importance of the need to keep main verb and object together. Secondly, there are a number of verbs that resist pre-placement of not and take longer to accept the do-periphrasis. This group consists of verbs such as say, think, hope, know, doubt, trow, woot, fear, etc. It is precisely this group of verbs with which negatives typically can have two different scopes: the scope can be over the matrix verb or just its complement. In most instances in which these verbs are used, however, the scope of the negative will be the complement rather than the verb itself, which means that the position of not after the verb is more appropriate for these verbs. Note that in tag expressions like I hope not, I think not, where the negative clearly concerns the content of the hoping and thinking, and not the absence of the mental action itself, postverbal placement of not is still the rule.

3.3.7

Internal structure of the Aux phrase

 

We have already seen that a large number of new periphrastic constructions developed in the ME period and beyond, to express tense, mood and aspect distinctions that had formerly been part of the inflectional form of the verb. Two points need to be discussed in connection with these new auxiliaries, namely the diachronic order in which they develop and occur combined with each other, and the linear sequencing of these auxiliaries at each synchronic stage. The more grammaticalised these auxiliaries became, the more strict we would expect their ordering principles to be (grammatical items are usually strictly bound in the syntax of the clause, which is much less true for fully referential lexical items). This expectation is fully borne out in the case of the grammatical markers in the verbal group; in fact, the developments here ultimately resulted in one of the most orderly and systematic areas of English syntax. Denison (2000a) provides a very useful overview of the earliest occurrence of various combinations of auxiliaries (see Table 3.3). Table 3.3 makes clear that the modals were the first to find a firm position in the auxiliary system; already in OE they freely combine with the passive auxiliary (It can be sung, not shown in Denison’s table), with perfect have

(He must have sung it) and with progressive be (He will be singing it), even though the latter two were only just developing. This is presumably because from the very beginning the modals were followed by infinitives, and it was only a small step from a full verb infinitive to an auxiliary-like infinitive. The combinations of other auxiliaries was in each case later because they first had to develop firm auxiliary status before they could be combined with verbs that they had never before been combined with. For instance, have was in OE only found with a past participle of a full, transitive verb. Only when have had lost its independent

Syntax 159

Table 3.3 Combinations of auxiliaries in the verbal group (adapted from Denison, 2000a: 139)

 

 

first

second

three

Pattern

Date:

pair

pair

auxiliaries

 

 

 

 

 

Modal + Perfect + Progr. + V (he will have been

 

OE

a.1325

?a.1425

singing it)

 

 

 

 

Modal + Perfect +Passive be + V (it will have

 

OE

c.1180

c.1300

been sung)

 

 

 

 

Perfect + Progr. + Passive be + V (it has been

 

a.1325

1772

1886/1929

being sung)

 

 

 

 

Modal + Progr. + Passive be + V (it will be

 

OE

1772

1915

being sung)

 

 

 

 

Modal + Perfect + Passive GET + V (it will have

 

OE

1832

1950–

gotten sung)

 

 

 

 

Modal + Progr. + Passive GET + V (it will be

 

OE

1819

PDE

getting sung)

 

 

 

 

Perfect + Progr. + Passive GET + V (it has been

 

a.1325

1819

PDE

getting sung)

 

 

 

 

 

 

 

 

 

status did it come to occur in combinations with passive be (It has been sung) and progressive be (He has been singing it) (all in ME). The combinations of passive and progressive be are all late (rows 3 and 4 in Table 3.3), much later than even the four-verb combinations that did not result in double be (the first two in Table 3.3). It is very likely that the awkwardness of double be played a role in this. There is a natural tendency in language to avoid repetition of grammatical forms in immediate succession. Presumably the two be’s first had to grammaticalise into separate lexical items (see also below) before they could co-occur.

Warner (1993) emphasises that the combinatory possibilities of the auxiliaries is very much constrained by the category and subcategorisation properties that each auxiliary has. For instance, modals must be followed by an infinitival form. Because the modals do not possess an infinitival form, they cannot occur consecutively to another modal. This then forbids combinations of modals syntactically in standard English even though semantically they make perfect sense (as the occurrence of ‘double modals’ in American and Scottish English shows; see Section 3.3.3 above). Likewise, because modals do not have participial forms, they cannot occur after perfect have and progressive be, which are subcategorised for a past and present participle, respectively. In other words, in English the modal must occupy the first slot in the VP. The case with perfect have is slightly different. Unlike the modals, it possesses an infinitival form, but like the modals it lacks a participle (note that the present participle of perfect have may occur in non-finite clauses). Consequently, perfect have can occur after a modal, but it cannot occur after passive or progressive be because both of these are subcategorised for a participle. Warner thus shows that the ordering constraints are formal rather than semantic, as follows:

Соседние файлы в предмете [НЕСОРТИРОВАННОЕ]