Does imperative 'do' (in 'do'-support) share the same form with infinitive 'do'?

Question

Without do-support, all imperative verbs are in the same form as infinitive verbs.

(1) Shut up. [imperative]
(2) I want you to shut up. [infinitive]

I can't think of any exceptions.

But, with do-support, it's not clear whether the same is true.

(1') Do shut up. [imperative]
(2') *I want you to do shut up. [infinitive]

Infinitives can't take do-support, so (2') doesn't work. The only way you can have infinitive do is by having it as a lexical verb, as in (4):

(3) Do the work. [imperative]
(4) I want you to do the work. [infinitive]

Does this mean that the 'do' in imperative 'do'-support doesn't share the same form with infinitive 'do'?

Answer 1

I think that you're getting at the rule that the auxiliary verb "do" must be in the primary position of any verb catena, i.e., it is not the dependent of any other verb.¹ (Some people would say that it can only function as an "operator".²) Thus, "do shut up" is fine but "I want you to do shut up" (which places "to do" subordinate to "want") isn't.

---

The only way you can have infinitive do is by having it as a lexical verb

This isn't entirely correct, because a primary auxiliary verb (including "do") can sometimes take an infinitive form. Your example demonstrates that:

Do shut up.

---

Does this mean that the 'do' in imperative 'do'-support doesn't share the same form with infinitive 'do'?

I believe that most people would agree that in such sentences "do" is in the infinitive form.

---

¹ I am speaking for AmE. According to comments (see below), this rule doesn't necessarily hold for other dialects.

² According to one grammar book, "Auxiliary do is the dummy operator: in the absence of any other auxiliary, it functions as the operator to form (for example) interrogative and negative sentences." (Sidney Greenbaum, The Oxford English Grammar, 1996, section 4.29)

Answer 2

I can’t speak authoritatively on CGEL, but since I have access to this doorstop of a tome, I can quote a few places that might answer your question, according to CGEL’s grammar framework:

The imperative, subjunctive, and infinitival all have the same form, and CGEL calls that form plain:

. . . the plain form, is used in the following three constructions:
[imperative] [subjunctive] [infinitival]
We pointed out in §1.2 that there is never any morphological difference between the form a verb has in the imperative construction and the form it has in the subjunctive construction, and we can now add that the form concerned is also identical with that used in the infinitival construction.
_{p.83 / Chapter 3 The Verb — 1.5 / The plain form}

Emphatic imperatives use do-support:

Supportive do is also used in imperatives to emphasise the positive polarity and again it occurs unconditionally, not just in the absence of an auxiliary, as in non-imperatives:
[16] a. Do hurry up. b. Do be careful.
. . .
_{p.929 / Chapter 10 Clause type and illocutionary force / 9.2.3 Imperatives with auxiliary do / Emphatic imperatives}

In emphatic imperatives, the plain do is used:

The auxiliary is used only in the four do-support constructions; it has only tensed forms, except that plain do/don’t occur in emphatic or negative imperatives.
_{p.1535 / Chapter 17 Deixis and anaphora / 7.6 Do it, do that, etc. / Summary of different uses of do}

So, yes, the plain form of do is what appears in emphatic do constructions.

_{Source: The Cambridge Grammar of the English Language}

Answer 3

This is part of an answer (so not aiming at the bounty). There is a historic 'periphrastic' usage of 'do' as auxiliary.

[https://www.jstor.org/stable/43343095]

You could say "I do go shopping every Tuesday". It is (was) very simple. Unlike the copula, 'do', of course, does not inflect at all: just the subject followed by base form 'do' followed by the base verb, as above.

This is grammatically unfamiliar, of course and so non-standard. But it is also grammatically reasonable. The surprise, in a way, is that this simpler form did not become standard.

This non-inflection applied also do the copula itself. I can remember sentences like "I be worried about the cabbages". I be, you be, s/he be, we be ... loses nothing that the crazy linguistic cocktail *I am, you are s/he is supplies.

But the cocktail or rather its users prevailed. So, of course did the users of was/were/was/were/were/were over the lower class users of just 'was'.

Of course the rural users of the uninflected 'do' and 'be' in this way was looked down on as a mark of uneducated rusticity, while the urban usage was looked down on a working class.

Now (I admit this is speculative), if you think of 'do' in this way (which has some history) it is not much of a stretch to see 'do' plus 'go' as a natural imperative. 'Do you go?' also fits the pattern. In fact, this usage of 'do', seen even in this dim light is not obviously 'periphrastic' - in its origin, at least.

This is not an answer, but I hope it is a thought.

Answer 4

47 days?
What happens then?

Do-support is a syntactic rule. As such, it has conditions and constraints for its application. Do-support is required only where another rule (like subject-auxiliary inversion, or negative adverb) requires an auxiliary verb, but there is no auxiliary verb to hand.

In that case, English produces a meaningless do that behaves like an auxiliary verb (in fact, it behaves like a modal, since it has to be followed by an infinitive verb form, and it has to be the first auxiliary, (because there weren't any other auxiliaries, duh), like modals do. But unlike modals, it has no meaning at all and merely facilitates the inversion or contraction or whatever.

Thus,

• Bill eats asparagus ~ Bill doesn't eat asparagus. ~ Does Bill eat asparagus?

But these are not imperative constructions. There is no requirement for an auxiliary verb in an imperative sentence. Consequently there is no requirement for Do-Support to apply, and no evidence of its having applied. Granted, do can occur in some varieties of imperatives. But that's not Do-Support. There are other varieties ('readings' or if you prefer, 'flavors') of do, rather like deontic and epistemic readings of modals, beside Do-Support do, which the simple verb has mutated into.

One of them is the so-called "emphatic do", as in DO be quiet now!, or Do have a little talk with him, won't you?. Another is the pro-verb do, which can stand for just about any verb (but not many predicate adjectives or nouns)

• What he said to do was to stand there/sing the blues/be back in 1 hour/*be tired

Then there's Ross's ACT do, which is restricted to syntactically active verbs and often takes noun phrase objects describing the activity

• I don't know how he finds the time to do it/do carpentry/do the work.

Do, like let, and sense verbs, and aspectual verbs like try and manage, are "small verbs" that don't follow the rules for big verbs with meanings; these guys sneak down the aisles and get into all kinds of trouble. Like kids