Linguistics 522


Lecture 8

V Movement

V Movement has been proposed in two forms. First General V Movement:

This takes any verb and moves it into Infl. The agreement feature present in Infl would then force the verb to be realized as the third person singular form (in this case annoys).

In this class we don't assume General V movement. We assume V movement for auxiliaries, which is to say, for have and be, the only auxiliaries which don't start out in Infl.

Note first of all that have and be cannot start in Infl in general, because they can occur when Infl is filled:

This is unlike other things that start out in Infl, which cannot occur when Infl is filled:

This being the case we may assume one of two things, either something like the Emonds analysis for have and be, which we all argued for so persuasively in the midterm:

    John [Infl may] [V'' [V' have [V'' [V'left ]]]]
Or the Aux specifier analysis [used in the textbook in examples like (68), p. 425]. For this example that looks like:
    John [Infl may] [VP [Aux have] [V' left ]]

Footnote for the Aux junkies. One advantage of the Emonds analysis is that it accounts naturally for examples like:

    John [Infl may] [V'' [V' have [V'' [V' been [V''leaving ]]]]]
Do you see the problem this poses for the Aux specifier analysis? The problem with double specifiers?

Be that as it may, we will accept either analysis. Pick your fav on assignments. In these notes I will use the Emonds analysis.

So on a V-movement analysis, have and be start out in V'' and get moved to Infl when it is empty. They then receive agreement features. For example:


Consider facts like these:

  1. John [Infl will] not play football.
  2. * John [Infl will] play not football.
  3. John [Infl does] not play football.
  4. John wants [Infl to] not play football.
On the basis of these facts we can tentatively conclude that not should follow whatever is in Infl. In particular, it can't follow any ordinary verb.

But note that not can follow either have or be:

  1. John has not played football.
  2. John is not playing footbnall.
  3. John has not been playing football.
We can account for these facts very simply if we continue to claim that not must immediately follow Infl, but that verb V-movement may move have and be in front of not. For example:
  • John [Infl e] not [V'' [V' have [V'' [V' been [V'' playing football ]]]]]

  • => [V-Movement]

  • John [Infl has] not [V'' [V' [V'' [V' been [V'' playing football. ]]]]]]


Consider facts like these, which we have seen before:

  1. John [Infl will] probably play football.
  2. * John [Infl will] play probably football.
These examples are here as a reminder that a sentence adverb like probably can follow whatever is in Infl. Being a sentence adverb it must be dominated by S. In particular, it can't occur between an ordinary verb and its complements, which would put it in V'.

But note that probably can follow either have or be:

  1. John has probably played football.
  2. John is probably playing footbnall.
  3. John has probably been playing football.
  4. * John has been probably playing football.
We can account for these facts very simply if we continue to claim that probably must immediately follow Infl, but that verb V-movement may move have and be in front of probably. For example:
  • John [Infl e] probably [V'' [V' have [V'' [V' been [V'' playing football. ]]]]]

  • => [V-Movement]

  • John [Infl has] probably [V'' [V' [V'' [V' been [V'' playing football. ]]]]]]


For a more elaborate argument, we next look at the phenomenon of Have contraction:

  1. I have played tennis all summer.
  2. I've played tennis all summer.
  3. You/they/we have played tennis all summer.
  4. You/they/we 've played tennis all summer.
  5. He/she has played tennis all summer.
  6. He/she 's played tennis all summer.
Singular and plural forms of have both contract after certain pronouns.

We will assume there two conditions:

  1. Contraction to pronominal subject (I, you, we, he, she,they, etc.).
  2. Not across gap.

We've illustrated the working of the first condition. AS to the second, consider what is called gapping:

  1. I could have been playing tennis and you could have been playing football.
  2. I could have been playing tennis and you Ø football.
  3. I could have played tennis and you Ø have played football.
  4. * I could have played tennis and you Ø 've played football.
The second condition accounts for the ungrammaticality of the last sentence. If we assume there is a gap (denoted Ø) marking the site of the omitted material in gapped sentences, then the second principle predicts that contraction isn't possible in such sentences.

Now consider our verb movement case:

  • John [Infl e] [V'' [V' have [V'' [V' been [V''leaving ]]]]]

  • => [Contraction blocked because of e]

  • => [V-Movement]

  • John [Infl has] [V'' [V' [V'' [V' been [V''leaving ]]]]]]

  • => [Contraction okay now]

  • John [Infl 's] [V'' [V' [V'' [V' been [V''leaving ]]]]]]

I Movement

This is a transformation that used to be called Subject Auxiliary Inversion:

The first question is: Why do this with a movement rule? Why not generate the auxiliaries in alternative ways? For example, let the base component (X-bar theory and the lexicon) generate the aux at the front of the sentence.

The second question is why generate the Aux in Comp position?

Why posit a movement transformation? One reason: There is always one finite auxiliary:

  • *Must John can go?
We explain this fact if inverted forms are created by movement, that is, if they leave behind an empty Infl that can't be refilled.

Another reaons is morphological dependency. The verb forms following auxiliaries are constrained:

  1. Modals:
    • John may go.
    • * John may going.
    • * John may gone.
  2. Have:
    • John has gone.
    • * John has going.
    • * John has go.
  3. Be:
    • John is sleeping.
    • * John is slept.
    • * John is sleep
Exactly the same dependencies hold for inverted forms
  1. Modals:
    • May John go?
    • * May John going?
    • * May John gone?
  2. Have:
    • Has John gone?
    • * Has John going?
    • * Has John go?
  3. Be:
    • Is John sleeping?
    • * Is John slept?
    • * Is John sleep?
Generalization: Get the same form you get after modal/have/be whether or not inversion has happened. This "coincidence" is explained if the dependencies of what verb form can follow an auxiliary are determined at D-structure, and I-movement (subject-auxiliary inversion) happens later.
Uniformity   At some level, direct and indirect questions look alike.
  • Can John go?
  • D-structure [Comp e ] John [Infl can ] go
  • Betty wonders whether John can go.
  • D-structure Bettey wonders [Comp whether] John [Infl can ] go

The difference between the two kinds of comp at D-structure is boils down to whether Comp is filled or not. We capture their underlying uniformity.

Why land
in Comp?
  The argument for landing in Comp position is first of all purely descriptive. Comp position is the right place, right at the front of the S-bar. But there's also a more compelling argument, namely that inversion and complementizers are in complementary distribution. You never get the two together, even when the opportunity arises.
  1. Comp filled: John wondered whether Mary would come.
  2. Inversion : John wondered would Mary come?
  3. Inversion and Comp filled : * John wondered whether would Mary come?
If inversion can only happen when comp is empty, because inversion means moving into an empty comp position, this is explained.