MODALS and SEMI-MODALS

The SEMI-MODALS—be going to, have to, have got to, be able to, be about to, get to and be bound to—could on syntactic grounds be regarded as main verbs taking a Modal (FOR) TO complement clause (with subject that is coreferential with the main clause subject and thus omitted), similar to attempt or fail. The main reason for our linking them with MODALS lies in the semantic parallels between will and be going to, must and have (got) to, can and be able to.

But we did demonstrate a recurrent semantic difference, in examples (1a/ b), (2a/b) and (3a/b) of MODALS tend to indicate prediction, ability, necessity, etc. under specific circumstances, while SEMI-MODALS are more likely to have an ‘unconditional’ sense and relate to the subject’s involvement in an activity. Compare (61a), with MODAL must, and (61b), with SEMI-MODAL have to:

(61a) You mustn’t mind what he says when he’s sleepy

(61b) You have to watch out for muggers after dark in the town centre

It is thus entirely appropriate for all SEMI-MODALS to include to, in terms of the meaning of the Modal (FOR) TO complement.

We illustrated the two uses of dare and of need. When used to refer to some general circumstance (almost always in questions and negatives) they may behave syntactically like MODALS, e.g.

(62) No one dare question my credentials

But when need and dare refer to a definite subject becoming involved in some activity, then they behave like a non-MODAL and take complementiser to (although this may sometimes be omitted), e.g.

(63) That man dares (to) question my credentials

Just three of the fifteen MODALS include to. This is semantically appropriate with be to, a verb which most often has a human subject who is scheduled to become involved in an activity, e.g. I am to call on the Vice-Chancellor tomorrow. However, it is hard to discern any semantic reason for the to of ought to and used to; we should perhaps search for a historical explanation in these instances.