One of the questions I face as a teacher of English grammar is how
accurate should a description of English be.  Jeff provides us with a
standard description of the English auxiliary system from an early
generative perspective.  A version of which I use in my own teaching
English grammar to pre-service teachers.

> T + (Modal) + (Have + EN) + (BE + ING) + MV

And, I agree

> It is a more direct route to the understanding of the English
> verb than memorization of a whole list of tenses, modes, and voices that
> even we who teach grammar have a hard time remembering.

As he showed with his examples, it is very useful to teach what past
perfect progressive actually means.

Note 4 is not nearly the problem he thinks it is.

> Note #4:  We are dealing here only with simple sentences that occur in the
> forms of the unelaborated basic sentence patterns. Used in other sentences,
> slight differences occur in the formula which we will not go into at this
> time. Just as an example, the auxiliary verb "do" might be present.

Do occurs when there is nothing else in AUX and T is unable to be
attached to the main verb.  We can all recognize structures when that
happens and when that does not happen, can't we?

In noting that NOT is located after T, we have an insight into the
system which explains why children and second language learners of
English go through what is referred to in the literature as a NO VP
stage.  They need to figure out that NOT goes after tense in English
This property of verbal negation is very unusual in the world's
languages.

However, there are aspects of the description above which present
problems.

> Note #3:  The above formula is for finite clauses. For nonfinite VPs, omit
> T.

This is not accurate and shows what is problematic with the
description.

In (1), to "describe English accurately" is nonfinite.  "Describe" has
no tense.

        1) We want to describe English accurately.

Likewise, in (2), "be able to" is nonfinite.

        2) We want to be able to describe English accurately.

The description above predicts that (3) should be perfectly grammatical.

        3) *We want to can to describe English accurately.

The explanation I have for this (it is not an original insight) is that
in modern English all of the modals have inherent tense.  As a
consequence, they can never occur in a nonfinite position as "can" in
(3).  By the way, we now have an explanation for why no modal every
takes the agreement -s.  In English, you can never mark a verb twice for
tense.  German modals, by the way, don't have this property. Because
they have nonfinite form, they all can be marked for tense.

(An aside. The ungrammaticality of (3) can not be explained by saying it
makes no sense.  (2) is perfectly acceptable and "to be able to" and
"can" in this context mean essentially the same.  This same argument can
be made with "must" and "have to".  Such an interesting formal
constraint raises questions about theories of grammar which assert that
grammar rules must in some way be related to meaning.)

To return to Jeff's description, it is much more accurate and insightful
to say that ALL FINITE clauses in English have either T or a modal.

We can now dispense with all of this description of the modals.

> Modal.  We have quite a list. Memorize it. Can, could, will, would, shall,
> should, may, might, must. The modal, if it occurs, is the first word in the
> AUX. Therefore, if it occurs, it carries tense (past or not past). Can,
> will, shall, and may are Not Past. Could, would, should, and might are Past.
> Must is fuzzy in terms of tense and not a problem.

Historically, the past and non-past distinction in the list above is
correct.  If you know German, you can see recognize that past/non-past
distinction in German modals  In reported speech, there are some
remnants of this past/non-past distinction.

        4) Last week, Bush said, "I will not interfere with the execution.
        5) Last week Bush said that he would not interfere with the execution.

However, there is a serious complication if the past/non-past
distinction still exists.  If might, could, should, would are past, why
are they all possible in the following?

        6) The Cubs might/could/should/would win the pennant this year (for
"would" insert: only if God intervenes).

This is one of the ways modality is signaled in English, as Judy
Diamondstone has just told us.

I do not know how important it is for teachers of native speakers of
English to know all of this. I think it is very important in being able
to recognize passive constructions.  I also teach pre-service ESL
teachers.  Understanding the formal properties of English auxiliary
system for them is very important.

Bob Yates, Central Missouri State University