Lecture Four: X’-theory

Ian Roberts 

04/21/23 1Roberts, Structure, Mich 2010, Lecture 4

The story so far:

 Syntactic theory provides structural descriptions of

sentences (labelled bracketings, tree diagrams) in terms of categories and constituents

PS-rules specify all and only the grammatical structures (well-formed structural descriptions) of a language

PS-rules can be recursive, and thereby specify an infinite set of well-formed structural descriptions

Constituency tests can be used to show which structural descriptions are correct

 Today we will make a significant refinement to the

system of PS-rules.  

04/21/23 2Roberts, Structure, Mich 2010, Lecture 4

Some PS-rules of English:a. NP (D) N (PP)

[NP [D the ] [N picture ] [PP [P of ] [NP [N John ]]]]

b. AP (Mod) A (PP) [AP [Mod very ] [A angry ] [PP [P with ] [NP [N John ]]]]

  c. PP (Mod) P NP [PP [Mod just ] [P beyond ] [NP [D the ] [N frontier ]]]

  d. VP (AdvP) V NP [VP [AdvP really] [V enjoy ] [NP [D the ] [N movie ]]]

04/21/23 3Roberts, Structure, Mich 2010, Lecture 4

Some impossible PS-rules:

a. NP V AP b. PP N V c. AP V NP

04/21/23 4Roberts, Structure, Mich 2010, Lecture 4

A generalisation over attested PS-rules

XP (highly optional stuff) X (fairly optional stuff)

In fact, the stuff following X depends on the lexical entry of X.

04/21/23 5Roberts, Structure, Mich 2010, Lecture 4

Some sample lexical entriesbuy(V) ___ NP So: VP V NPrely(V) ___ PP So: VP V PPput(V) ___ NP PP So: VP V NP PPsay(V) ___ CP So: VP V CP The material following X is known as the

complement of X; that preceding it as the Specifier of X (the Determiner “specifies” something about the Noun; the modifier “specifies” something about the A/P, etc.).

 04/21/23 6Roberts, Structure, Mich 2010, Lecture 4

Collapsing the PS-rules into the X’ schemaa. XP Spec X’b. X’ X complement 

-- here X is a categorial variable whose possible values are one of N, V, A, P.

04/21/23 7Roberts, Structure, Mich 2010, Lecture 4

In tree format XPruYP X’

ruX ZP

 structurally, therefore, a complement = the sister

of a head, while a specifier = the sister of the head + complement structure (X’) → different levels of syntactic “closeness” are structurally represented

04/21/23 8Roberts, Structure, Mich 2010, Lecture 4

Evidence for X’ from coordination a. the [ pictures of John ] and [ books

about Mary ]b. very [ angry with John ] and [ worried

about Mary ]c. just [ beyond the frontier ] and [ down

the road ]d. really [ enjoy the movie ] and [ dislike

the book ]

04/21/23 9Roberts, Structure, Mich 2010, Lecture 4

3-level structure:head (X) head (X) + its complement = X’[X’ head (X) + its complement] + specifier =

XP 

04/21/23 10Roberts, Structure, Mich 2010, Lecture 4

Some terminology:

X’ is an intermediate projection between the head (X) and its maximal projection (XP)

depending on the information in the lexical entry, the complement may, like the Specifier, not be generated: smile(V) ___ (so V’ V)

X-Bar Theory is nevertheless meant to operate in a “blind” fashion as far as its hierarchical structure is concerned: although complements and specifiers are generated depending on lexical considerations, it is always essential that X be dominated by X’ and that X’ be dominated by XP – cf. the X-Bar representation of Liz (strictly speaking; textbooks generally don’t bother to fill in all the redundant structure, but this is theoretically important)

04/21/23 11Roberts, Structure, Mich 2010, Lecture 4

NP|N’

  |N

Liz 

04/21/23 12Roberts, Structure, Mich 2010, Lecture 4

Extension to functional categoriesa. John

will/can/should/would/won’t/doesn’t talk to Mary.

b. John has/hasn’t left.c. John is/isn’t leaving.-- evidence for an Aux position in English

sentences, which is distinct from VP. 

04/21/23 13Roberts, Structure, Mich 2010, Lecture 4

VP-fronting a. We expected John (not) to talk to Mary,

and [VP talk to Mary ], John will/can/should/would/won’t/doesn’t t.

b. We thought John leave, and [VP left ] he has t /[VP leaving ] he is t.

04/21/23 14Roberts, Structure, Mich 2010, Lecture 4

VP-proform (do so)Angela will [VP finish the race in under an

hour] and Michael will do so too.

04/21/23 15Roberts, Structure, Mich 2010, Lecture 4

ElisionRichard has [VP entered the competition] → I

know he has [ ]

04/21/23 16Roberts, Structure, Mich 2010, Lecture 4

S → NP (Aux) VP Is Aux really optional? What (grammatical)

information does it contain?

04/21/23 17Roberts, Structure, Mich 2010, Lecture 4

a. Michelle has finished her paper → I know she has [ ]

b. Richard has entered the competition → I know he has [ ]

04/21/23 18Roberts, Structure, Mich 2010, Lecture 4

Tag questionsAux and subject of the main clause,

confirming it’s true:

a. So her work is finished, is it?b. So he has entered the competition, has

he? 

04/21/23 19Roberts, Structure, Mich 2010, Lecture 4

DoIf no auxiliary is present in the main clause,

do appears:a. Michelle finished her paper. I know

she did [ ].b. Richard entered the competition. I

know he did [ ].c. So she finished her work, did she?d. So he entered the competition, did he?

04/21/23 20Roberts, Structure, Mich 2010, Lecture 4

 Aux always expresses at least tense and

agreement with the subjectThe appearance of do where the verb is

deleted implies that tense and agreement (i.e. Aux) must always be expressed. If there is no verb or auxiliary independently available, do shows up.

04/21/23 21Roberts, Structure, Mich 2010, Lecture 4

S NP Aux VPdoesn’t fit the X’-schema, but there is

evidence from coordination for an intermediate-level category:

John [ should talk to this man ] and [ will then act on the consequences ].

04/21/23 22Roberts, Structure, Mich 2010, Lecture 4

S NP ???? Aux VP-- What category is this?

04/21/23 23Roberts, Structure, Mich 2010, Lecture 4

AuxP AuxP

ru

NP Aux’ he ru Aux VP should ru V PP talk ru P NP to ru D N this man

04/21/23 24Roberts, Structure, Mich 2010, Lecture 4

What about when there is no auxiliary?There is still a tense. So we rechristen AuxP

as TP (better for other languages, as nearly all languages have tense but not all languages have auxiliaries).

04/21/23 25Roberts, Structure, Mich 2010, Lecture 4

TPTP

ru NP T’ he ru T VP should r u V PP talk r u P NP to ru D N this man 

04/21/23 26Roberts, Structure, Mich 2010, Lecture 4

What about CP?We can treat complementisers as heads, so

we have:  C’ C TP

04/21/23 27Roberts, Structure, Mich 2010, Lecture 4

CP(John said) CP

|C’

ruC TPthat ru

NP T’ he ru

T VPshould ru

V PPtalk ru P NP to Mary

04/21/23 28Roberts, Structure, Mich 2010, Lecture 4

“Sluicing” and the Specifier of CP

John knew he had to talk to one of the students, but he couldn’t remember

[CP which one ([TP he should talk to])

04/21/23 29Roberts, Structure, Mich 2010, Lecture 4

(He couldn’t remember) CPruNP C’

which oneru ru C TP

ruNP T’he ru

T VPshould ru V PP

talk ru P NP

to t

- note the wh-movement here

04/21/23 30Roberts, Structure, Mich 2010, Lecture 4