Episode 7b. Subjects, agreement, and case - PowerPoint PPT Presentation

1 / 36
About This Presentation
Title:

Episode 7b. Subjects, agreement, and case

Description:

CAS LX 522 Syntax I Episode 7b. Subjects, agreement, and case 6.1-6.3 Historical interlude Back in the old days, people hypothesized that Pat will eat lunch had a ... – PowerPoint PPT presentation

Number of Views:68
Avg rating:3.0/5.0
Slides: 37
Provided by: PaulHa53
Learn more at: https://www.bu.edu
Category:

less

Transcript and Presenter's Notes

Title: Episode 7b. Subjects, agreement, and case


1
CAS LX 522Syntax I
  • Episode 7b. Subjects, agreement,and case
  • 6.1-6.3

2
Historical interlude
  • Back in the old days, people hypothesized that
    Pat will eat lunch had a structure like this.
  • The subject NP Pat was in the specifier of IP
    (what we call TP), and the VP contained only
    the verb eat and the object NP lunch.
  • Pat got an Agent q-role by being in SpecIP, even
    though the fact that there is an Agent q-role to
    be had is determined by the verb down in the VP.

IP
NP Pat
I?
I will
VP
V eat
NP lunch
3
Historical interlude
  • Nevertheless, this predicts the normal word order
    pretty well, and so it was hypothesized that the
    verb simply assigned one of its q-roles directly
    to SpecIP.
  • No big deal, syntax works in strange and
    mysterious ways.
  • At a certain point, someone started thinking
    about sentences like these
  • All the students will take the exam.
  • The students will all take the exam.
  • Its fairly clear here that all the students is
    an NP, that it forms a coherent unit, a coherent
    concept. All really belongs with the students.

IP
NP Pat
I?
I will
VP
V eat
NP lunch
4
Historical interlude
  • All the students will take the exam.
  • The students will all take the exam.
  • Back in the even older days, the hypothesis was
    that there was a special rule that turned the
    first sentence into the second.
  • The Quantifier Float rule would move all over to
    the right, next to the VP.
  • all NP VP ? NP all VP

IP
NP Pat
I?
I will
VP
V eat
NP lunch
5
Historical interlude
  • Not all quantifiers are subject to Quantifier
    Float
  • Quantifiers every, some, all, most, several,
    many, both, four,
  • Every student will take the exam.Student will
    every take the exam.
  • Several students will take the exam.Students
    will several take the exam.
  • It works for both and allThe students will both
    take the exam.The students will all take the
    exam.
  • Whats a difference between every, some, several,
    many and both, all?

IP
NP Pat
I?
I will
VP
V eat
NP lunch
6
Historical interlude
  • Upon further reflection, some enterprising
    syntacticians hit upon the idea that rather than
    floating all to its position next to VP, all
    might instead have been left behind by a
    subject that had moved.
  • will all the students take the exam.
  • all the studentsi will ti take the exam.
  • the studentsi will all ti take the exam.
  • And why would all the students have been down
    there? Well, that would simplify assignment of
    q-roles.

IP
NP Pat
I?
I will
VP
V eat
NP lunch
7
The VP-Internal Subject Hypothesis
  • The verb (head of VP) can assign q-roles to other
    things within the VP, which is a natural
    explanation for how the choice of verb controls
    whether an Agent q-role is assigned or not.
  • This idea became known as the VP-Internal Subject
    Hypothesis.

IP
NP Pat
I?
I
VP
will
V?
ti
V eat
NP lunch
8
The VP-Internal Subject Hypothesis
  • For us, weve supposed from the beginning that
    assignment of q-roles is necessarily local. This
    may not seem like a very surprising hypothesis.
  • But it was at the time a rather unintuitive idea,
    and so various people set out to see if some of
    the predictions this makes are borne out in the
    grammatical data.

IP
NP Pat
I?
I
VP
will
V?
ti
V eat
NP lunch
9
The VP-Internal Subject Hypothesis
  • It turns out that as people looked, there were
    reasons to believe this.
  • The new analysis of Quantifier Float no longer
    relies on an idiosyncratic rule of English, but
    more general principles.
  • The assignment of q-roles can now be more
    directly related to the properties of the verb.
  • And we can make sense of there constructions in a
    more straightforward way.

IP
NP Pat
I?
I
VP
will
V?
ti
V eat
NP lunch
10
Back to the present
Quantifier stranding is still often referred to
as quantifier float to this day, even though
the name no longer reflects the analysis.
  • The basic components of the quantifier
    stranding phenomenon are
  • All the students is a constituent. The students
    is an NP inside all the students.
  • all NP the students
  • Either all the students or just the students can
    move to SpecTP, to satisfy the uN feature of
    T.
  • So all the students and the students are both
    NPs.
  • NP all NP the students
  • So all is essentially a noun, but one that takes
    an NP complement (all N, uN, ).

NP
all
NP
thestudents
11
All the students will take the exam
  • We start by building our vP.
  • Merge the NP the exam and the V take (checks
    uN on V)
  • Merge v and VP (HoP)
  • Move V to v (checks uV on v)
  • Merge the N all and the NP the students (checks
    uN on all)

vP
v?
NP
VP
v
all
NP
ltVgt
vagentuInfl,
Vtake
NP
thestudents
the exam
12
All the students will take the exam
  • We Merge the T will with vP (HoP)
  • This values uInfl on v.
  • Will is really a modal (present tense version of
    would).
  • Policy Modals value uInfl features as
    uInflnone(none meaning uninflected, but
    still checked.)

T?
vP
T willuN,
v?
NP
VP
v
all
NP
ltVgt
vagentuInflnone,
Vtake
NP
thestudents
the exam
13
All the students will take the exam
Is all the students closer to T than the students
is? Not if we define closer as we did, in terms
of c-command. Where X c-commands Y and Z, Y is
closer to X than Z is if Y c-commands Z.
  • Now, there are two possibilities
  • Move the NP all the students.
  • Move the NP the students.

T?
vP
T willuN,
v?
NP
VP
v
all
NP
ltVgt
vagentuInflnone,
Vtake
NP
thestudents
the exam
14
All the students will take the exam
  • Now, there are two possibilities
  • Move the NP all the students.
  • Move the NP the students.

TP
T?
NP
vP
T willuN,
all
NP
v?
ltNPgt
thestudents
VP
v
ltVgt
vagentuInflnone,
Vtake
NP
the exam
15
The students will all take the exam
  • Now, there are two possibilities
  • Move the NP all the students.
  • Move the NP the students.

TP
T?
NP
vP
T willuN,
thestudents
v?
NP
VP
v
all
ltNPgt
ltVgt
vagentuInflnone,
Vtake
NP
the exam
16
Expletive constructions
  • An expletive is an element that can be in subject
    position without having received a q-role from
    anywhere.
  • It had been raining.
  • There are fans rioting on Comm Ave.
  • Weve seen it before. But there is also there,
    which well concentrate on now. Neither means
    anything, neither gets a q-role, both appear to
    satisfy the uN feature of T.
  • Both can be used in other ways I saw it over
    there.

17
Expletive constructions
  • There are fans rioting on Comm Ave.
  • Fans are rioting on Comm Ave.

IP
TP
TP
NP there
I?
fans
there
T?
T?
I are
VP rioting
T
T
ProgP
ProgP
T
be
T
be
ltbegt
ltbegt
vP
vP
ltfansgt
fans
v?
v?
ltVgt
vVriot
ltVgt
vVriot
18
Case
  • Recall that pronouns in English show distinctions
    in case
  • Subject pronouns are in nominative case
  • Object pronouns are in accusative case
  • I saw her. She saw me. They saw him.
  • How can we ensure that this happens?

19
Nom case
  • Nominative subjects generally appear in the
    specifier of a finite T.
  • Finite T is pretty much any kind of T except the
    infinitive to marker.
  • We can treat case like we treated tense
    inflection
  • Suppose T also has a ucasenom feature.
  • Suppose nominative NPs have a ucase feature.
  • Suppose the ucasenom on T can value ucase
    on the NP, checking both.
  • So T needs a nom NP, and a nom NP needs T.

20
Acc case
  • Subjects check nominative case with T. Objects
    have accusative case, which we can treat in the
    same kind of way.
  • Suppose v has ucaseacc.
  • Suppose accusative NPs have ucase
  • Suppose the ucaseacc on v can value the
    ucase feature on the NP, checking both.
  • Nominative case is a relation between (finite) T
    and an NP, accusative case is a relation between
    v and an NP.

21
Notes on case
  • Nominative case is associated with finite T.
  • She will eat lunch.
  • I want her to eat lunch.
  • I expect her to eat lunch.
  • Non-finite T is not associated with nominative
    case. Its not actually associated with
    accusative case either, but well come back to
    that later.
  • Because NPs have an unvalued ucase feature, we
    can suppose that pronouns always enter the
    numeration the same way, and are valued based on
    where they are Merged.
  • pronoun N, ucase,

22
Notes on case
  • Although in English we only see the morphological
    effect of case on pronouns, we assume that all
    NPs have an unvalued ucase feature.
  • Plenty of languages other than English show case
    on all NPs, not just on pronouns. Case is
    something that goes with being an NP. Its just
    something you often dont hear in English.
  • Notational shortcuts
  • nom is used for ucasenom (on T, or NP when
    checked)
  • acc is used for ucaseacc (on v, or NP when
    checked)
  • case is used for ucase (on an NP)

23
Subject-verb agreement
  • Recall that in English, the f- features of the
    subject have an effect on the morphology of the
    verb
  • Fans are rioting on Comm Ave.
  • A fan is rioting on Comm Ave.
  • While were here, we might as well account for
    this as well. It is also an agreement relation,
    between the subject and, eventually, the verb.

24
Subject-verb agreement
  • The verb gets its tense inflection specified by T
    when, e.g., the tensepres feature of T values
    the uInfl feature of v.
  • Since the subject already agrees with T (the
    nom feature of T checks the case feature of
    the subject), well incorporate subject agreement
    into this process.
  • Notice that we still want this agreement to be
    mediated by T (sometimes it values, e.g., Perf)
  • They have been reading novels.
  • She has been reading novels.

25
Subject-verb agreement
  • Suppose then that T has a uf feature as well.
  • The subject has (interpretable) f-features that
    value the uf feature of T.
  • Fans are rioting on Comm Ave.
  • T T, uN, uf, nom
  • fans N, fpl, case
  • So, once T is in the structure, c-commanding fans
    in SpecvP, we get
  • T T, uN, ufpl, nom
  • fans N, fpl, nom

26
Subject-verb agreement
  • Finally, we suppose that the (checked) ufpl
    feature of T, also values a uInfl feature on a
    lower v (or Perf, or Prog).
  • The rules of pronunciation will tell us that a v
    with the verb riot adjoined to it sounds like
  • riots if v has the feature uInflpres,sg
  • riot if v has the feature uInflpres,pl.
  • Notice that T values a uInfl feature all at
    once, with any relevant feature(s) it has (so,
    tense and f-features both).

27
She likes them
  • So, lets walk through it.
  • We start by merging like and the 3pl pronoun.

VP
NPpronounN, f3pl, case
VlikesV
28
She likes them
  • v v, uN, uInfl, uV, acc
  • We Merge v with VP (HoP).
  • The acc on v matches, values, and checks the
    case on the pronoun, checking itself as well.

v?
VP
v v, uN, uInfl, uV, acc
NPpronounN, f3pl, acc
VlikesV
29
She likes them
  • The V moves up to adjoin to v to check the uV
    feature of v.

v?
VP
v
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInfl, uV, acc
VlikesV
30
She likes them
  • The V moves up to adjoin to v to check the uV
    feature of v.
  • The 3sg feminine pronoun is Merged to check the
    uN feature of v.

vP
v?
NPpronounN, f3fsg, case
VP
v
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInfl, uV, acc
VlikesV
31
She likes them
  • The T is Merged with vP (HoP).
  • T has the features T, tensepres, uf, uN,
    nom.
  • The nom feature of T matches, values, and
    checks the case feature of the pronoun,
    checking itself in the process.

T?
TT, tensepres, uf, uN, nom
vP
v?
NPpronounN, f3fsg, nom
VP
v
VlikesV
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInfl, uV, acc
32
She likes them
  • The f3fsg feature of NP values and checks the
    uf feature of T.

T?
TT, tensepres, uf3fsg, uN, nom
vP
v?
NPpronounN, f3fsg, nom
VP
v
VlikesV
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInfl, uV, acc
33
She likes them
  • The uf3fsg and tensepres features of T
    value and check the uInfl feature of v.

T?
TT, tensepres, uf3fsg, uN, nom
vP
v?
NPpronounN, f3fsg, nom
VP
v
VlikesV
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInflpres3fsg, uV, acc
34
She likes them
  • Finally, the NP is moved up and Merged with T? in
    order to check the EPP feature (the uN
    feature) of T.

TP
NPpronounN, f3fsg, nom
T?
TT, tensepres, uf3fsg, uN, nom
vP
v?
ltNPgt
VP
v
VlikesV
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInflpres3fsg, uV, acc
35
She likes them
  • All uninterpretable features are checked, the
    pronunciation rules give us she likes them.

TP
NPpronounN, f3fsg, nom
T?
TT, tensepres, uf3fsg, uN, nom
vP
v?
ltNPgt
VP
v
VlikesV
NPpronounN, f3pl, acc
ltVgt
v v, uN, uInflpres3fsg, uV, acc
36
?
  • ? ?
  • ?
  • ? ?
  • ? ?
  • ?
  • ?
Write a Comment
User Comments (0)
About PowerShow.com