04.01.09
(Available on-line – please print out if you can!) 1. Review, with PRINCIPLES
2. Affix Hopping: how participles get their suffixes in English 3. T-C movement: Questions
4. T-C movement: German statements
1. Review, with P
RINCIPLES(1) Hans has eaten these apples.
(2) The D-structure: the tree that is built by TREE GENERATOR using the LEXICON and that conforms to THE PROJECTION PRINCIPLE, THETA CRITERION, WORD-TO-TREE RULE. (i.e. most of our course up to the midterm.)
TP
DP T’
Hans T VP [past] V VP
have
V’
V DP eaten
these apples
But the D-structure violates the WORD RULE
WORD RULE (PART 1) If a head X0
is occupied by a morpheme that cannot be pronounced as a separate word all by itself, do something to make sure it can be pronounced.1
Solution #1: V-T MOVEMENT: move the contents of V to T Solution #2: AFFIX LOWERING: move the affix to the verb Solution #3: DO-INSERTION: if all else fails, inset do into T.
1What kinds of morphemes can’t be pronounced as separate words is determined on a case-by-case basis. We will
VERB MOVEMENT PARAMETER: Determines what classes of verbs can move. Set by each language.
French: any verb (also Italian, German, Irish …) English: just auxiliary verbs (also Haitan Creole, ??, ...)
Solution #1: V-T Head Movement
(3) The S-Structure for English showing V-T movement
TP
DP T’
Hans T VP Have+[past]
V VP
V’
V DP eaten
these apples
(4) (Some) Rules of (English) morphology: “↣” = pronounce as
Have+[past] ↣ /hæd/ “had”
Have+[pres] ↣ /hæs/ “has”
Is+[past, 3rd person, sg] ↣ /wʌz/ “was” Is+[past, 3rd
person, plural] ↣ /wɛr/ “were” (we’ll just write tense info on T;
Is+[pres, 3rd
person, sing] ↣ /ɪz/ “is” subject determines person/number)
Is+[pres, 3rd
person, pl] ↣ /ær/ “are”
Is+[pres, 1st
person, sg] ↣ /æm/ “am”
⋮
Regular Verbs:
walk+[pres, 3rd
person, sg] ↣ /wɑks/ “walks”
walk+[pres, all other] ↣ /wɑk/ “walk”
walk+[past] ↣ /wɑkt/ “walked”
Past Irregular Verbs:
eat+[past] ↣ /ejt/ “ate”
swim+[past] ↣ /swæm/ “swam”
read+[past] ↣ /rɛd/ “read”
(5) Don’t skip heads: *Hans was have eating these apples.
TP
DP T’
Hans T VP = ungrammatical movement be+[past]
V VP have
be VP
V’
V DP eating
these apples
HEAD MOVEMENT CONSTRAINT (aka don’t skip heads!)
Short version: neither head movement (V-T, T-C) nor affix lowering can “skip” heads.
Long version: A head Z can move to a head X if there is no head Y such that X c-commands Y and Y c-c-commands Z.
Why can’t V stop off at V auxiliary?
WORD RULE (PART 2)
If two morphemes occupy one head position (via head movement or affix lowering), they must form a word in the language.
Solution #2: Affix Lowering
(6) The S-Structure for English after Affix Lowering
TP
DP T’
Hans T VP N.B. Adv does not “get in the way”
according to HEAD MOVEMENT CONSTRAINT V’
AdvP V’
Adv’ V DP eat+[past]
Adv these apples
today
(7) Why not Affix Lowering to put tense on Auxiliaries?
TP
DP T’
Hans T VP
V VP Have+[past]
V’
V DP eaten
these apples
(8) *I not have eaten these apples
Solution #3: DO-INSERTION
Adding Negation: and how English and French (and English Auxiliaries) respond differently
(10) Hans ne-mange pas des pommes. (Tense V precedes Negation) Hans ne-eat not of.the apples
‘Hans didn’t eat the apples’
(11) The D-Structure
TP
DP T’
Hans T NegP [pres]
pas Neg’
Neg VP ne-
V’
V DP mange
des pommes
V to T movement technically has to be V to Neg to T movement to satisfy what?______
(12) Step One:
TP
DP T’
Hans T NegP [pres]
pas Neg’
Neg VP
ne-mange
V’
V DP
(13) Step two: the S-structure (to satisfy the WORD RULE)
TP
DP T’
Hans T NegP ne+mange+[pres]
pas Neg’
Neg VP
V’
V DP
des pommes
English:
(14) *Hans not eats the apples.
Affix lowering from T to V violates what? _________________.
TP
DP T’
Hans T NegP
not Neg’
Neg VP
V’
V DP eat+[pres]
(15) But why not two affix lowerings? T down to Neg, then down to V?
TP
DP T’
Hans T NegP
not Neg’
Neg VP
V’
V DP eat+[pres]
des pommes
TENSE MORPHOLOGY RULE: (other languages may have different morphological rules)
Tense morphology ([PRES],[PAST]) can only form a word with (i.e. affix to) verb roots.
i.e. Neg+[pres] is not a word in English.
WORD RULE (PART 2) …from above…
If two morphemes occupy one head position (via head movement or affix lowering), at any point, they must form a word in the language.
Sucks for you English! What are you going to do?
(16) DO-INSERTION: if all else fails, insert do into T.
(17) The do in DO-INSERTION is not an auxiliary that starts in a V. It appears in T only when needed.
*John should not do eat apples.
(19) The S-Structure after English DO-INSERTION
TP
DP T’
Hans T NegP do+[pres]
not Neg’
Neg VP
V’
V DP eat
des pommes
Important: DO-INSERTION do is not in D-structure; it appears in S-structure as a “last resort.”
2. Affix Hopping
You (will) have noticed that the form of verbs that follow auxiliaries is determined by the auxiliary:
Perfect have: following verb appears as a past participle have __+en/ed e.g I have eaten/read/traversed/
- past participle is either –ed or –en (and irregular ones)
Progressive be: following verb appears as a progressive participle be __+ing e.g John is eating/traversing/reading
- the suffix –ing is added to the main vebr form the progressive participle.
Passive (be): following verb appears as a passive participle be ___+en/ed e.g The bread was eaten/traversed
Chomsky’s (1957!) solution: the affixes are “born” on the auxiliary verbs. The affixes then lower onto the verb below them.
(20) V V V
We are going to let some “word trees” like these appear in sentence trees.
D-structure for John has eaten.
(21) TP
DP T’
Hans T VP [pres] V VP
have [perf] V’
V DP eat
these apples
have and [perf] part company. Have raises (it has to by RAISE!) and [perf] lowers.
(22) TP
DP T’
Hans T VP have+[pres]
V VP
V’
eat+[perf] ↣ /itәn/ “eaten”
V DP eat+[perf]
these apples
It is a mystery why have and [perf] have to part company (why can’t [perf] stay on have?). We are going to have to say that have can’t have a perfect suffix, progressive be can’t have a progressive suffix, and passive be can’t have a passive suffix. (This might be like our TENSE MORPHOLOGY RULE, something about English morphology, and WORD RULE (PART 2) kicks in)
And here’s where things get fun … Combinations!
D-Structure for (b)
(24) TP
DP T’
Hans T VP should
V VP
have [perf] V VP
be [prog] V’
V DP eat
these apples
(25) S-Structure
TP
DP T’
Hans T VP should
V VP
have V VP
be+[perf] V’
V DP eat+[prog]
these apples
[Perf] is usually [-en] (or [-ed]) depending on the verb stem it attaches to. [Prog] is always [-ing].
be+[perf] ↣ be+[en] ↣ /bin/ “been”
3. T-C Movement: English Questions
We will treat this movement like other instances of head movement – there is some “silent” morpheme that needs help in pronunciation.
(26) Should Hans have seen this film?
The D-structure for a question is:
(27) CP
C
C TP [Q]
DP T’
Hans T VP should V VP
have
V’
V DP seen
these films
[Q] violates our WORD RULE
Putting something in C would satisfy [Q] and the WORD RULE. Movement appears to be
T-to-C movement relocates the contents of T and puts it in C, making a new word [should+Q].
(28) CP
C
C TP
should+[Q]
DP T’
Hans T VP
V VP have
V’
V DP seen
these films
It is the element in T that must satisfy [Q] and the WORD RULE. Moving have won’t work:
(29) *Have Hans should seen these films.
Why is this?
Interaction between V-T and T-C
(30) *Eat you the apples?
(31) Mangez-vous des pommes? Eat-you of.the apples? ‘Do you eat apples?’
Why is English different from French?
4.
T-C: German
In German, the word order differs in main and embedded clauses.
Embedded Clauses
(32) Jan glaubt… Jan believed…
a. … [dass Hans das buch kauft] that Hans the book buys ‘that Hans buys the book’
b. … [dass Hans das Buch gekauft hat] that Hans the book bought has ‘that Hans has bought the book’
c…. [dass Hans das Buch gekauft haben muss.] that Hans the book bought have must. ‘that Hans must have bought the book
And the one from the assignment (page 281 ex.d)
(33) a. Er sagt [dass die Kinder diesen Film gesehen haben] He said that the children this film seen have ‘He said that the children saw this film’
Basic Structure of an Embedded Clause in German.
(34) CP
C TP
dass
DP T’
Hans VP T
muss VP V
haben
V’
DP V
gekauft
das buch
When there is no modal word in T, the verb moves to that position (by the way, we don’t have incontrovertible evidence yet for this…but hold on!)
So the D-structure tree for embedded clauses with auxiliaries looks like this in German:
(35) CP
C TP
dass
DP T’
Hans VP T [pres] VP V
haben
V’
DP V
gekauft
And it is transformed into the following S-structure tree via V-T head movement:
(36) CP
C TP
dass
DP T’
Hans VP T
hat
VP V V’
DP V
gekauft
das buch
Where German differs from English (beside that TPs and VPs are right headed) is that main verbs also move to T. (Of course, in these sentences you can’t “hear” the difference, but go with me on this one; we’ll get more evidence later!)
When there is no Auxiliary verb, then raise the finite V to T.
(37) CP
C TP
dass
DP T’
Hans VP T kauft+[past] DP V
das buch
So German has V to T for auxiliaries and main verbs!
German is then just French with T and V on the right instead of left.
Matrix Clauses
(38) a. Hans kauft das Buch Hans buys the book.
b. Hans hat das Buch gekauft. Hans has the book bought. ‘Hans has bought the book’
c. Hans muss das Buch gekauft haben. Hans must the book bought have.
‘Hans must have bought the book’.
The tensed verb (a tensed main verb, a tensed aux, or a modal) here appears to the left of the object.
First try: “Maybe the tensed verb has moved to T in main clauses, but not in embedded clauses?”
No: because we see that T, when filled by a modal, is on the right in embedded clauses. So T must be on the right of VP.
Second try: “ok maybe T is in different places in main and embedded clauses?”
No: We would have to have a different headedness parameter for TP depending on whether it is in a main clause or embedded clause. This would not be a good way of describing the difference between main an embedded clauses because we have no way of telling our grammar to put things on different sides depending on other material in the sentence!
Third try: “maybe the verb has moved around nearer the front of the sentence in the matrix clause but not in the embedded cluase?”
Yes. Your assignment showed you that in main clauses, the tensed verb doesn’t always just follow the subject, but follows any one phrase, called ‘topic’.
(39) Gestern sahen die Kinder den film. Yesterday saw the children the film. ‘Yesterday the children saw the film.’
(40) Gestern haben die Kinder den film gesehen. Yesterday have the children the film seen. ‘Yesterday the children have seen the film.’
But in Embedded Clauses: if there is a complementizer dass, the finite verb stays at the end of the sentence, meaning it hasn’t moved. So the following is ungrammatical:
(41) a.*Er sagt [dass die Kinder haben diesen Film gesehen] He said that the children have this film seen ‘He said that the children have seen this film’
b. *Er sagt [dass haben die Kinder diesen Film gesehen] He said that have the children this film seen ‘He said that the children have seen this film’
c. Er sagt [dass die Kinder diesen Film gesehen haben] He said that the children this film seen have ‘He said that the children saw this film’
Generalizations:
• The tensed verb in German appears at the front of its cluase after one constituent – which is the topic, and we put it in the specifier of CP. This phenomenon is called the “V2”. • If there is a complementizer dass this doesn’t happen. When we find such complementary
distribution, we assume two things are in the same position.
Conclusion: In German finite/tense verbs move to C. (but not to form a question)
(42) CP
XP C’
gestern C TP
DP T’
Hans VP T
[present] VP V
hat
V’
DP V
gesehen
(43) CP
XP C’
gestern C TP
DP T’
Hans VP T
hat+[present] VP V
V’
DP V
gesehen
Diesen film
(44) CP
XP C’
gestern C TP hat+[pres]
DP T’
Hans VP T VP V
V’
DP V
gesehen
(45) CP
C TP
hat+[pres]+dass
DP T’
Hans VP T VP V
V’
DP V
gesehen
Diesen film
(46) hat+[pres]+dass is not a word in German!
Same for English when we try to move an auxiliary to T that is filled by a stand-alone word, like a modal:
(47) TP
DP T’
Hans T VP
Have+should V VP
V’
V DP seen
these films
Have+should is not a word in English.
WORD RULE (PART 2)
If two morphemes occupy one head position (via head movement or affix lowering), they must form a word in the language. (e.g. V+[pres] is a word. But Neg+[pres] is not, at least in English and other familiar languages. Nor is
Some mysteries – things we won’t explain – about German V2:
In matrix clauses, there must always be a phrase in Specifier of CP, called the topic.
When that happens, the tensed verb moves from T to C. Why T moves to C in German in this case we will leave as an open question. Also, you can’t have anything preceding dass in the specifier position of CP. Another mystery. Also, T-C here is not “driven” by the WORD RULE – i.e. we haven’t put a morpheme that can’t be pronounced all by itself in here – or maybe we could?
5. A Transformational Generative Grammar
Movement, what were once called “transformations”, allow one sentence to obey conflicting constraints. E.g. PROJECTION PRINCIPLE and the WORD RULE. We do this by giving a sentence more than one tree structure, and use movement a mechanism for relating these trees – or changing one into the other.
LEXICON
Tree Generator
D-structure S-Structure
PROJECTION PRINCIPLE WORD RULE
THETA CRITERION EPP
WORD-TO-TREE RULE CASE ASSIGNMENT
X-BAR X-BAR
Transformations (=movements)
Morphological rules (e.g. have+[past] ↣ had)
NOTE: This picture isn’t suppose to describe what goes on in your head when you produce a sentence. It is a model of knowledge – what you know when you know the syntax of a language. (It just happens to look “chronological”, but the arrows are just here to make it easier to think about.)
D-structure is buit by RANDOM TREE GENERATOR using the languageʼs LEXICAL ITEMS.
A. THE LEXICON
Every morpheme of the language is listed in the lexicon with its category,
subcategorization and (if a V,P,A,N…) its theta-grid. Morphological rules, like the ones we saw above, are put here too.
B. TREE GENERATOR
Lexical items can be put into phrase structures in any way you want. The labels of these phrases are random, the structures are random. The trees must be well formed: no
crossing branches; no multiple immediate dominance (two mothers), the tree must be “connected”, dominance is represented top-to-bottom.
D-structuremust conform to the following principles, however.
D-structure Principles
A. i. WORD-TO-TREE RULE
a. A word/morpheme’s category label is the node that’s put in the tree. It is an X0
, a “head”. The head carries all the information that the lexical entry does (we’ll write its pronunciation or spelling below the X0 often).
b. Every X0
must head a well-formed XP.
c. A well-formed XP conforms to the X-BAR SCHEMA.
ii. X-BAR SCHEMA iii. X-BAR PARAMETERS
XP → (ZP) X’ Specifier: first/last
X’ → QP X’ Adjunct: first/last
X’ → X0
(YP) (UP) Head: first/last
iv. α → α CONJ α (Conjunction rule; doesn’t conform to X-bar)
C. THE THETA CRITERION
a. For every θ-role there is one unique position to which that θ-role is assigned.
b. For every θ-position, there is one unique phrase with an appropriate semantic value
D. PROJECTION PRINCIPLE
i. If X assigns a theta-role to Y, X and Y are sisters.
ii. If anything is sister to a head X0, then X0 assigns a theta-role to it.
There are also some principles that D-structure cannot conform to. These are called S-structure principles.
S-structure Principles
A. WORD RULE (PART 1) If a head X0
is occupied by a morpheme that cannot be pronounced as a separate word all by itself, do something to make sure it can be pronounced.
B. EXTENDED PROJECTION PRINCIPLE (the “EP P”)
All clauses must have subjects. (i.e. the specifier of TP must be filled by DP or CP)
D-structure and S-structure Principles often conflict, making it impossible for one tree to satisfy both. Movement Rules solve this by “transforming” D-structures to S-structures. Or, less metaphorically, transformations are ways of relating (at least) two trees: one that satisfies D-structure Rules and another that satisfies S-structure rules. We will think of transformations as moving parts of a D-structure to form an S-structure, and we will usually draw one tree with arrows.
Movement Rules
V-T MOVEMENT: move the contents of V to T T-C MOVEMENT: move the content of T to C AFFIX LOWERING: move an affix down the tree.
Not just anything can move around anywhere. We have seen some constraints on Movement, and we will see more.
Constraints on Movement
A. VERB MOVEMENT PARAMETER: What can move:
B. HEAD MOVEMENT CONSTRAINT
Short version: neither head movement (V-T, T-C) nor affix lowering can “skip” heads.
Long version: A head Z can move to a head X if there is no head Y such that X c-commands Y and Y c-c-commands Z.
C. RAISE!: Move if you can.
Constraints on movement imposed by Morphology
A. WORD RULE (PART 2)
If two morphemes occupy one head position (via head movement or affix lowering), at any point, they must form a word in the language.
B. TENSE MORPHOLOGY RULE: (other languages may have different morphological rules) Tense morphology ([PRES],[PAST]) can only form a word with (i.e. affix to) verb roots.
In some cases, Movement is completely prevented by various things and it is impossible for the language to satisfy all the constraints. In this case, Last Resort rules kick in.
Last Resort Rules
A. DO-INSERTION: if all else fails, insert do into T