142 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

presumably because in hypothetical contexts the emphasis would be on the activity rather than on the resultative state, because that state is never in fact reached. This again reduces the ‘intransitive’ domain of be. Thus in (32a) have is used rather than be because the context is one of ‘irrealis’, in contrast to (32b), where we have a factual situation:

(32) a. He wende have cropen by his felawe John, | And by the millere in he creep

b.noot I for-why ne how/ That jalousie, allas . . . | Thus causeles is cropen

For similar reasons have is more frequent in iterative and durative contexts. Frequency is a crucial factor in grammaticalisation processes, so it is not surprising that slowly have gains on be.

At the same time, there are formal and functional factors that account for the decrease of be. The combination of be + past participle was functionally at a disadvantage because it was ambiguous: it was also used for the passive construction. Thus a phrase like she was returned could mean ‘she had returned’ as well as ‘she had been returned’. In addition, the common abbreviation of both has and is to ’s led to a further falling-together of the be and have verbs, under the more frequent perfect auxiliary have. The decline of be follows the typical S-curve: it is a slow but steady one in ME and early ModE. It picks up speed in Late ModE and comes to a climax in the nineteenth century, which witnesses a dramatic drop in the frequency of be.

Mood as a formal category is typically connected with the verb. Mood inflections on the verb are used to express the attitude of the speaker towards the factual content of a proposition. Thus in OE the indicative mood is commonly used in the context of an objective, factual report and is the rule in most constructions that do not involve grammatical dependence. The subjunctive is a subjective expression and is found especially in volitional, conjectural and hypothetical contexts, which are, as it were, one step removed from the situation as fact, from our purely socio-physical experience of the world. In OE the subjunctive occurs most frequently in dependent clauses, following a main clause which introduces modality lexically. In other words, the subjunctive is not (or no longer) a primary marker of modality: it has grammaticalised. Main clauses, where modality needed a stronger expression, already usually contained a modal verb in OE; a mere subjunctive inflection on the main verb did not suffice. In some cases the grammaticalisation of OE subjunctives has gone so far that the subjunctive form has become semantically meaningless. Thus when a subordinate clause depends on another subordinate clause, the subjunctive is the rule when the first clause has a subjunctive. We also often find an ‘empty’ subjunctive in indirect speech:

144 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

‘that it is no wonder that the holy king can heal sickness now that he lives in heaven’ (ÆLS (Oswald)272)

It is clear from these examples that mood can be expressed in many different ways. The marking may be primarily lexical, i.e. by means of full verbs, ‘hedges’ (e.g. phrases like I guess, I think), adverbs, idiomatic expressions, etc. One step down the ladder of grammaticalisation would be the use of modal auxiliaries, while the subjunctive is an even further grammaticalised form.

As far as the expression of mood in the history of English is concerned, we see two main shifts, both involving the modal verbs. First of all, there is the replacement of (subjunctive) inflections by periphrastic (modal) constructions. In grammaticalisation terms, this is the beginning of a new cycle, in that a bleached inflection becomes replaced by a new, more expressive modal verb. We see this already happening in subordinate clauses in later OE. Where, before, a subjunctive was sufficient to express mood (as shown in the (a) examples above), we more and more see the insertion of a modal verb in that position, often itself in the subjunctive. Compare the instances in (37) with a modal verb, to (34a) and (35a), which still have only a subjunctive inflection on the finite verb:

Syntax 145

‘it is necessary therefore that the monasteries will be moved from that place to

‘the Saviour said to them, do you believe me now that I can heal both of you?’ (Mt (WSCp) 9.28)

This development was possible because modals were already fully employed in main clauses. The development was also necessary. In OE the verb had different forms for the following moods: indicative, imperative and subjunctive (see Chapter 2). However, already in OE not all endings were distinctive: there was no distinction between indicative and subjunctive in the past tense of strong verbs in the second-person singular, nor in the past tense of weak verbs in the firstand third-person singular. Distinctiveness decreased further in ME, where all -e, -on, -en endings fell together under -e, thus obliterating many earlier differences between indicative and subjunctive forms. In PDE practically all these endings have disappeared: only the stem form is left and one single inflection to mark the third-person singular present indicative, i.e. -s. The imperative is now distinguished only by its front position and the fact that usually no subject is present; the form itself does not differ from the indicative form (or rather the stem) in PDE. The PDE subjunctive has been reduced to a few relics like were (If I were you) and the occasional use of the stem form in subordinate clauses such as It is necessary that every member inform himself of these rules, and in main clauses (in its optative function) in set expressions like Heaven forbid that . . .

The second important shift concerns the modal verb itself. When it slowly grammaticalised into an auxiliary through replacing the subjunctive, it also began to enter the epistemic or discourse domain, taking the place of earlier lexical markers of epistemic modality. Thus whereas in OE it was usual to employ adverbs and predicative phrases such as wenunga, eaþe, wen is þæt ‘(it is) possible, probable’, or verbal constructions such as me þynceþ ‘me seems’, to indicate epistemic modality, the modals begin to play a much larger role here in later English. In fact, this development has led to some modals becoming virtually restricted to the epistemic domain. Thus the form might, which used to have dynamic and deontic meanings (i.e. it expressed ability and – later – permission), is now (almost) exclusively used as an epistemic modal, so that He might come a bit later can only be understood as the description of a possibility, but not ability or permission; in its wake, an utterance like He may come a bit later, with the modal may, is now also gradually losing its permission sense, especially in American English.

Finally, one other development must be noted here, that of the so-called modal preterite. In OE it was usual to employ a past-tense subjunctive in hypothetical constructions, both in present and past-time contexts: