A
transformation used in English grammar to produce complicated arrangements of
auxiliary verbs and endings, as in
I would have been eating, from a straightforward underlying structure.
For example, the perfect tense I have spoken is derived from an underlying I have+EN speak, representing a verb have which causes the following verb to take an EN ending. Before being translated into speech, the affix hops one position to the right, giving speak-EN, and then a mental dictionary translates this into the correct phonetic form spoken.
Now for more detail.
Excluding the irregular 'to be', verbs in English have five possible forms, for example see, sees, seeing, saw, seen. The -s and -ing forms are always formed regularly by adding those suffixes to the plain form. That leaves three principal parts, see, saw, seen. The second part is simple past tense, and the third part is used in two compounds, the perfect tense I have seen and the passive voice I am seen.
See the nodes English verbs, English verb tenses, and irregular verbs of English for details of the forms and irregularities.
But regardless of the complications of the actual form, in structural terms every verb has a plain present form, a simple past form, a present participle, and a past participle; and we may regard these as the verb stem plus an abstract grammatical element, which may be symbolized PRES and PAST and ING and EN respectively.
The use of -s is automatic in the present tense when the subject is third person singular, but structurally it is the same as the zero ending for other persons: I eat, you eat, she eat-s, they eat are all grammatically parallel. So the abstract ending PRES has two realization rules: in I eat-PRES the PRES comes out phonetically as zero but in she eat-PRES it comes out as -s.
All statements contain a tense, either present or past, and it's always marked on the first verb if there's a series of them:
- I smoke vs I smoked
- Mary eats vs Mary ate
- you don't agree vs you didn't agree
- I can come vs I could come
- I will have been waiting vs I would have been waiting
Additional verbs that go in front of the main verb are called either auxiliary verbs or auxiliaries or modal verbs (the distinction is not important for this discussion). Many of them are followed by the first part or plain form of the verb, sometimes with to between them.
- I see
- I can see
- I will see
- I want to see
- I used to see
The meanings are varied: we form the future tense, and various aspects that may be called habitual, volitive, and so on. Historically, would is the past tense of will and might is the past tense of may, but these days they have distinct meanings; and in fact used to is a past tense that no longer has a corresponding present *use to. However, for all the variety of meanings, they are structurally similar.
Other auxiliaries be and have govern (are followed by) participles, which are the plain form plus either ING or EN.
- I am eat-ING - continuous
- I have eat-EN - perfect
- I am eat-EN - passive
These may be combined with other modals:
- I will be eat-ING
- I could have eat-EN
- I might be eat-EN
- and even:
I might have be-EN eat-ING
The tense always goes on the first verb
I am, I was, I will, I would: these may be analysed as
I be-PRES, I be-PAST etc.
The three participial compounds are analysed as underlying units: continuous be+ING and perfect have+EN and passive be+EN. Affix hopping will move the affix onto the end of the next verb.
The same process deals with tense. The tense marker comes on the end of the first verb. Instead of analysing the sentence as
Subject Verb1 Tense (Verb2 ...) Predicate
it is clearer if all the verbs are treated alike, as
Subject Tense Verb1 (Verb2 ...) Predicate
Tense becomes either PRES or PAST. At least one verb must be present. It can be a main verb like speak or an auxiliary like be+ING. So we may have:
Mary PAST be+ING speak
This is the underlying structure (the deep structure) with the choice of words filled in. Now come a number of first syntactic then phonetic transformations to create the final form. Affix hopping moves both the tense marker and the participle marker onto the next verb:
Mary be-PAST speak-ING
The surface structure of be-PAST for the singular Mary is was, and the abstract speak-ING becomes the actual form speaking. The final transformed sentence is Mary was speaking.