Transcript She
GRS LX 700
Language Acquisition
and
Linguistic Theory
Week 2. The emergence of syntax
Syntax 101
Initially, children start off producing basically oneword utterances.
Though not impossible (comprehension), it is difficult to
conclude much about syntactic knowledge at this stage.
Somewhere around one and a half years, kids will
start putting words together: Syntax… of a sort.
Papa have it (Eve 1;6)
Marie go. (Sarah 2;3)
Eve gone (Eve 1;6)
Eve cracking nut. (Eve 1;7)
Kitty hiding (2;10)
Fraser not see him (Eve 2;0)
Eve talk funny
This is recognizably related to English,
and even comprehensible, but it’s not the
way adults talk.
3sg -s often missing.
Past tense -ed often missing.
Auxiliaries have, do, and be often missing.
In general, it seems like the grammatical
(functional) bits that are missing. Actually,
it’s kind of specific type of functional bit.
My need tea
The things that seem to be missing are
actually things that all were considered
part of “INFL” (a.k.a. “I”, a.k.a. “T”). Tense
and subject agreement.
Even if syntactic theory has gone on to the
view that there are multiple functional
heads there (AgrSP, TP, AgrOP), it’s still
the functional part of the tree (vs. lexical).
Ah, yes. Short human, short
tree.
A very natural suggestion to make about
kids’ syntax at this stage is that it lacks the
functional layers of structure.
The sentences are “small clauses”—just
the VP, and the NP.
Various people have run with this idea. For
example, Radford, Vainikka.
“Structure building” approach to
acquisition of syntax.
Small Clause Hypothesis
Radford (1990, 1995), Early Child English
Kids’ syntax differs from adults’ syntax:
kids use only lexical (not functional) elements
structural sisters in kids’ trees always have a qrelation between them.
VP
NP q
V’
man
V
q
chase
“Small Clause
Hypothesis”
NP
car
Small Clause Hypothesis
Adults:
Kids:
Absence of evidence for IP:
CP—IP—VP
VP
adult syntax ≠ child syntax
No modals (repeating, kids drop them)
No auxiliaries (Mommy doing dinner)
No productive use of tense & agreement (Baby ride truck,
Mommy go, Daddy sleep)
Absence of evidence for CP:
no complementizers (that, for, if)
no preposed auxiliary (car go?)
no wh-movement (imitating where does it go? yields go?;
spontaneous: mouse doing?)
kids bad at comprehending wh-object questions (out of
canonical order). (—What are you doing? —No.)
Small Clause Hypothesis
Adults:
CP—IP—VP
Kids:
VP
adult syntax ≠ child syntax
Absence of evidence for DP:
no non-q elements
no expletives (raining, outside cold)
no of before noun complements of nouns (cup tea)
Few determiners (Hayley draw boat, want duck, reading
book)
No possessive ’s, which may be a D.
No pronouns, which are probably D.
See also Vainikka (1993/4) for a similar proposal.
To sleep little baby
Turns out kids talk funny around this time in lots of
languages. A particularly popular funny way to talk is
to use infinitives.
Danish:
køre bil
German:
Thorstn das habn
French:
Dormir petit bébé
Dutch:
Earst kleine boekje lezen
‘drive[inf] car’
‘T that haveinf’
‘sleepinf little baby’
‘first little book readinf’
Further evidence for missing functional projections?
Sleeps baby
Well, but maybe not. At the very same time as
they’re using these superfluously infinitive
verbs, they are also using finite verbs.
Well, yeah, sure, but they hear finite verbs. But they
don’t have the clause-structural support for it yet (so
they don’t know the verbs are finite or not—that’s
information one gets from INFL). It’s just that you can
pronounce ‘sleep’ either as dort (sleeps) or as dormir
(sleep). Right? Yes?
Well, it’s easy to check. See if they can tell the
difference. See if they make errors—finite verbs
come in various kinds, do they use 1st person
agreement when they should have used 3rd?
Do kids get I/T?
Radford points out that the overt realization of I
(T) is often missing (morphology, modals,
auxiliaries).
But is it random? Are kids just arbitrarily using
tense morphology when they do?
When tense is there, does it act like tense would
for an adult?
Do kids differentiate between tensed and infinitive
verbs, or are these just memorized Vs at this point?
If kids differentiate between tensed and infinitive
verbs, there must be some grammatical
representation of tense.
Adult German
Poeppel & Wexler (1993). Data: Andreas (2;1, from
CHILDES).
Adult German is SOV, V2
The finite verb (or auxiliary or modal) is the
second constituent in main clauses, following
some constituent (subject, object, or
adverbial).
In embedded clauses, the finite verb is final.
V2 comes about by moving the finite verb to
(head-initial) C.
German clause structure
CP
C
DP
C+I
Hans kaufte
This “second position” is
generally thought to be C,
where something else
(like the subject, or any
other XP) needs to
appear in SpecCP.
This only happens with
finite verbs. Nonfinite
verbs remain at the end
of the sentence (after the
object).
IP
I
—
—
VP
V
DP
den Ball
—
German clause structure
CP
DP
den
Ball
C
IP
C+I
hat
DP
I
Hans VP
V
—
V
gekaufte
Things other than
subjects can appear in
“first position”.
When the tense appears
on an auxiliary, the verb
stays in place.
What to look for
in Child German
Poeppel & Wexler found that Andreas will
sometimes use a finite verb, sometimes a
nonfinite verb.
In adult German: finite verbs move to 2nd position,
nonfinite verbs are clause-final.
Does this also happen in kid German?
Look for a correlation between finiteness and verb
position:
ich mach das nich
I do that not
du das haben
you that have
Results
There is a strong contingency.
Conclude: the finiteness distinction is made
correctly (at the earliest observable stage).
Conclude: children do represent tense.
Andreas: 33 finite, 37 nonfinite verbs. 8 in both: finite, V2;
nonfinite final. Remaining verbs show no clear semantic core that
one might attribute the distribution to.
+finite
-finite
V2, not final
197
6
V final, not V2
11
37
Verb positioning =
functional categories
In adult German, V2 comes
from V I C.
If we can see non-subjects
to the left of finite verbs, we
FP
know we have at least one
functional projection (above Object
F
the subject, in whose Spec
F+V
VP
the first position nonSubject V
subject goes).
—
—
Is it really V2 (not SVO)?
V2 (German) is different from SVO in that the
preverbal constituent need not be the subject.
Is Andreas really using adult-like V2 (not SVO)?
Look at what’s preverbal:
Usually subject, not a big surprise.
But 19 objects before finite V2
(of 197 cases, 180 with overt subjects)
And 31 adverbs before finite V2
Conclude: Kids basically seem to be acting like
adults; their V2 is the same V2 that adults use.
Full Competence Hypothesis
(Poeppel & Wexler 1993)
The morphosyntactic properties associated with
finiteness and attributable to the availability of
functional categories (notably head movement)
are in place.
The best model of the child data is the standard
analysis of adult German (functional projections
and all). The one exception:
Grammatical Infinitive Hypothesis:
Matrix sentences with (clause-final) infinitives are a
legitimate structure in child German grammar.
CP
The Full Competence Hypothesis says not only
that functional categories exist, but that the child
has access to the same functional categories
that the adult does.
In particular, CP should be there too.
Predicts what we’ve seen:
finite verbs are in second position only
(modulo topic drop leaving them in first position)
nonfinite verbs are in final position only
subjects, objects, adverbs may all precede a finite
verb in second position.
Comparing FCH to SCH
SCH (Radford, et al.) pointed to lack of
morphological evidence for CP.
But P&W showed syntactic evidence for a
functional category (V2 with finite verbs) to
which V moves. Adults use CP for this.
But they also tend not to use embedded clauses.
Which causes which?
finite verb is second
non-subjects can be first
“Absence of evidence ≠ evidence of absence.”
Andreas uses agreement correctly when he
uses it—adults use IP for that.
Is it really CP and IP?
Or just FP?
Can we get away with only one functional
category?
The word order seems to be generable this way
so long as F is to the left of VP.
subject can stay in SpecVP
V moves to F
non-subject could move to SpecFP.
…though people tend to believe that IP in
German is head-final (that is, German is headfinal except for CP). How do kids learn to put I
on the right once they develop CP?
Is it really CP and IP?
Empirical argument for CP & IP:
negation and adverbs mark the left edge of VP.
A subject in SpecVP (i.e. when a non-subject is
topicalized) should occur to the right of such elements
(if there’s just an FP).
So, look for non-subject-initial sentences with
negations or an(other) adverb.
There were 8 that matched the criteria.
All eight have the subject to the left of the
adverb/negation:
[CP
Object C+I+V [IP Subject [VP neg/adv tSubj tV] tI ]]
Kid structures
Hypothesis: Kids have full knowledge of the
principles and processes and constraints of
grammar. Their representations can be basically
adult-like.
But kids seem to optionally allow infinitives as
matrix verbs (which they grow out of).
(And when they use an infinitive, it acts like an infinitive.)
What’s happening when kids use an infinitive?
Harris & Wexler (1996)
Child English bare stems as “OIs”?
In the present, only morphology is 3sg -s.
Bare stem isn’t unambiguously an infinitive form.
No word order correlate to finiteness.
OIs are clearer in better inflected languages.
Does English do this too? Or is it different?
Hypotheses:
Kids don’t “get” inflection yet; go and goes are
basically homonyms.
These are OIs, the -s is correlated with something
systematic about the child syntax.
Harris & Wexler (1996)
Hypothesis: RIs occur when T is missing from
the structure (the rest being intact).
Experiment: Explore something that should be a
consequence of having T in the structure: do
support.
Rationale:
Main verbs do not move in English.
Without a modal or auxiliary, T is stranded:
The verb -ed not move.
Do is inserted to save T.
Predicts: No T, no do insertion.
Harris & Wexler (1996)
Empirically, we expect:
She go
She goes
She not go (no T, no do)
She doesn’t go (adult, T and do)
but never
She not goes (evidence of T, yet no do).
Note: All should be options if kids don’t “get” inflection.
Harris & Wexler (1996)
Looked at 10 kids from 1;6 to 4;1
Adam, Eve, Sara (Brown), Nina (Suppes), Abe
(Kuczaj), Naomi (Sachs), Shem (Clark), April
(Higginson), Nathaniel (Snow).
Counted sentences…
with no or not before the verb
without a modal auxiliary
with unambiguous 3sg subjects
with either -s or -ed as inflected.
Harris & Wexler (1996)
Affirmative:
43% inflected
782
47
< 10% inflected
-inflec
neg
Negative:
aff
It not works Mom
no N. has a microphone
no goes in there
but the horse not stand ups
no goes here!
+inflec 594
5
Harris & Wexler (1996)
Small numbers, but in the right direction.
Generalization: Considering cases with no
auxiliary, kids inflect about half the time normally,
but almost never (up to performance errors)
inflect in the negative.
If presence vs. absence of T is basically
independent of whether the sentence is
negative, we expect to find do in negatives about
as often as we see inflection in affirmatives.
Also, basically true: 37% vs. 34% in the pre-2;6
group, 73% vs. 61% in the post-2;6 group.
Harris & Wexler (1996)
When kids inflect for tense, do they inflect for the
tense they mean?
(Note: a nontrivial margin of error…)
Inflected verbs overwhelmingly in the right
context.
present
bare stem 771
-s
418
-ed
10
past
128
14
168
future
39
5
0
Harris & Wexler (1996)
Last, an elicitation experiment contrasting affirmative,
never (no T dependence for adults), and not.
Does the cow always go in the barn, or does she never go?
Does the cow go in the barn or does she not go in the barn?
Do you think he always goes or do you think he never goes?
Do you think that he goes, or don’t you think that he goes?
Processing load? Extra load of not alleviated by
leaving off the -s? If that’s the case, we’d expect never
and not to behave the same way—in fact, never might
be harder, just because it’s longer (and trigger more -s
drops).
Harris & Wexler (1996)
Affirmatives inflected often, not inflected rarely,
never sort of inbetween.
Looking at the results in terms of whether the
question was inflected:
Kids overall tended to use inflection when there
was inflection in the question.
When the stimulus contained an -s:
affirmative: 15 vs. 7 (68% had an -s)
never: 14 vs. 16 (48%)
not: 4 vs. 12 (25%) —quite a bit lower.
An alternative to missing T
Much of what we’ve seen so far could also be
explained if kids sometimes use a null modal
element:
Idea:
RI?
I want to eat pizza. I will eat pizza.
I want to eat pizza. I will eat pizza.
First question: why modals?
Second, they don’t (always) seem to mean what
they should if there is a null modal. 20/37 seem
to be clearly non-modal (according to P&W93).
Thorsten Ball haben (T already has the ball)
Modal drop
Can we test this another way? What are the
properties of adult modals?
Adult modals are in position 2, regardless of what is in
position 1. If kids are dropping modals, we should expect
a certain proportion of the dropped modals to appear
with a non-subject in position 1.
But none occur—nonfinite verbs also seem to come with
initial subjects.
Why? Well, if V2 is a) movement of V to T to C, and b)
“topicalization” of something to SpecCP; and, if this is
triggered by V reaching C: There’s no need to move
anything to SpecCP if V remains unmoved. The subject
remains first.
Modal drop
Just to be sure (since the numbers are small),
P&W check to make sure they would have
expected non-subjects in position 1 with
nonfinite verbs if the modal drop hypothesis
were true.
17% of the verbs are infinitives
20% of the (finite) time we had non-subject
topicalization
So 3% of the time (20% of 17%) we would expect
non-subject topicalization in nonfinite contexts.
Of 251 sentences, we would have expected 8.
We saw none.
Subject case errors
Various people have observed that kids
learning English sometimes will use
accusative subjects.
It turns out that there’s a kind of a
correlation with the finiteness of the verb
as well. Finite verbs go with nominative
case, while nonfinite verbs seem to go
with either nominative or accusative case.
Finiteness vs. case errors
subject
Schütze & Wexler
(1996)
Nina
1;11-2;6
Finite
Nonfinite
Loeb & Leonard
(1991)
7 representative kids
2;11-3;4
Finite
Nonfinite
he+she
255
139
436
75
him+her
14
120
4
28
% non-Nom
5%
46%
0.9%
27%
EPP and missing INFL
If there were just an IP, responsible for both
NOM and tense, then they should go together
(cf. “IP grammar” vs. “VP grammar”)
Yet, there are many cases of root infinitives with
NOM subjects.
And, even ACC subjects seem to raise out of the
VP over negation (me not go).
We can understand this once we consider IP to
be split into TP and AgrP; tense and case are
separated, but even one will still pull the subject
up out of VP.
Syntax
Recall the basic structure
of adult sentences.
IP (a.k.a. TP, INFLP, …)
is the position of modals
and auxiliaries, also
assumed to be home of
tense and agreement.
CP is where wh-words
move and where I moves
in subject-aux-inversion
Splitting the INFL
Syntax since 1986
has been more or
less driven by the
principle “every
separable functional
element belongs in its
own phrase.”
Various syntactic tests
support these moves
as well (cf. CAS LX
523).
Splitting the INFL
Distinct syntactic
functions assigned to
distinct functional heads.
T: tense/modality
AgrO: object agreement,
accusative case
AgrS: subject agreement,
nominative case
Neg: negation
Origins: Pollock (1989)
(split INFL into Agr and T),
Chomsky (1993) (split INFL
into AgrS, T, AgrO).
What to make of the case
errors?
Case is assumed to be
the jurisdiction of AgrSP
and AgrOP.
So, nominative case can
serve as an unambiguous
signal that there is an
AgrSP.
Accusative case,
conversely, may signal a
missing AgrSP.
Why are non-AgrSP
subjects accusatives?
Probably a default case in
English:
Who’s driving? Me. Me too.
It’s me.
Other languages seem
not to show this
“accusative subject” error
but also seem to have a
nominative default
(making an error
undetectable).
“ATOM”
Schütze & Wexler
propose a model of
this in which the
case errors are a
result of being able
to either omit AgrSP
or Tense.
For a subject to be
in nominative case,
AgrSP must be
there (TP’s
presence is
irrelevant).
For a finite verb, both TP and
AgrSP must be there. English
inflection (3sg present –s) relies
on both. If one or the other is
missing, we’ll see an infinitive
(i.e. bare stem).
Thus, predicted: finite
(AgrSP+TP) verbs show Nom
(AgrSP), but only half of the
nonfinite verbs (not both AgrSP
and TP) show Nom (AgrSP).
We should not see finite+Acc.
Agr/T Omission Model
(ATOM)
Adult clause structure:
AgrP
NOMi Agr
Agr
TP
T
ti
T
VP
ATOM
Kiddie clause, missing TP (—TNS):
AgrP
NOMi
Agr
Agr
VP
ATOM
Kiddie clause, missing AgrP (—AGR):
TP
ACC
defaulti
T
T
VP
Pronunciation of English
T+AgrS(+V) is
pronounced like:
/s/ if we have features
[3, sg, present]
/ed/ if we have the
feature [past]
Ø otherwise
Layers of “default”, most
specific first, followed by
next most specific
(“Distributed
Morphology”, Halle &
Marantz 1993).
Notice: 3sg present –s
requires both TP and
AgrSP, but past –ed
requires only TP (AgrSP
might be missing, so we
might expect some
accusative subjects of
past tense verbs).
One prediction of ATOM
+AGR+TNS: NOM with inflected verb (-s)
+AGR–TNS: NOM with bare verb
–AGR+TNS: default (ACC) with bare verb
–AGR–TNS: GEN with bare verb
(the GEN case was not discussed by Wexler
1998, but see Schütze & Wexler 1996)
Nothing predicts Acc with inflected verb.
Finite goes with nominative
subjects.
subject
Schütze & Wexler
(1996)
Nina
1;11-2;6
Finite
Nonfinite
Loeb & Leonard
(1991)
7 representative kids
2;11-3;4
Finite
Nonfinite
he+she
255
139
436
75
him+her
14
20
4
28
% non-Nom
5%
46%
0.9%
27%
ATOM and morphology
[+3sg +pres] = -s
[+past] = -ed
—=Ø
[+masc +3sg +nom]
play+[3sg+pres]
[+2sg +nom]
play+[2sg +past]
he plays.
you play.
But is this knowledge
built-in? Hint: no.
[+masc, +3sg, +nom] =
he
[+masc, +3sg, +gen] = his
[+masc, +3sg] = him
[+fem, +3sg, +nom] = she
[+fem, +3sg] = her
[+1sg, +nom] = I
[+1sg, +gen] = my
[+1sg] = me
[+2, +gen] = your
[+2] = you
ATOM and morphology
What if the child
produces a lot of
utterances like
ATOM predicts that
agreement and
nominative case
should correlate.
So does this child’s
use of her goes to
school mean ATOM is
wrong?
her sleeping
her play
and even
her sleeps
her goes to school
but never uses the
word she?
Schütze (2001, inter alia)
No.
Her goes to school is not
necessarily a
counterexample to ATOM
(although it is a
candidate).
Morphology must be
learned and is
crosslinguistically
variable.
She is known to emerge
rather late compared to
other pronouns.
If the kid thinks her is the
nominative feminine 3sg
pronoun, her goes to
school is perfectly
consistent with ATOM.
Hence, we should really
only count her+agr
correlations from kids
who have demonstrated
that they know she.
ATOM and morphology
Morphology (under “Distributed
Morphology”) is a system of
defaults.
The most specified form
possible is used.
Adult English specifies her as
a feminine 3sg pronoun, and
she as a nominative feminine
3sg pronoun.
If the kid doesn’t know she, the
result will be that all feminine
3sg pronouns will come out as
her. That’s just how you
pronounce nominative 3sg
feminine, if you’re the kid.
Just like adult you.
[+masc, +3sg, +nom] =
he
[+masc, +3sg, +gen] = his
[+masc, +3sg] = him
[+fem, +3sg, +nom] = she
[+fem, +3sg] = her
[+1sg, +nom] = I
[+1sg, +gen] = my
[+1sg] = me
[+2, +gen] = your
[+2] = you
Rispoli (2002, inter alia)
Rispoli has a proposal
about her-errors.
Pronoun morphology is
organized into “tables”
(paradigms) basically,
where each form has a
certain weight.
When a kid is trying to
pronounce a pronoun,
s/he attempts to find the
entry in the table and
pronounce it.
The kid’s success in
finding the form is
affected by “gravity”.
“Heavier” forms are more
likely to be picked when
accessing the table, even
if it’s not quite the right
form. If it’s close and it’s
heavy, it’ll win out a lot of
the time.
Her by virtue of being
both acc and gen is extraheavy, and pulls the kid in
fairly often.
Her plays
ATOM and Rispoli make
different predictions with
respect to her plays.
ATOM says it should
never happen (up to
simple performance error)
Rispoli says case errors
are independent of
agreement, her plays is
perfectly possible, even
expected.
Rispoli’s critique of
Schütze’s studies:
Excluding kids who
happen not to produce
she in the transcript
under evaluation is not
good enough. The
assumption is that this
learning is monotonic, so
if the kid ever used she
(productively) in the past,
the her errors should not
be excluded.
Monotonicity
Schütze assumes that
use of she is a matter of
knowledge of she. Once
the kid knows it, and
given that the adult
version of the kid will
know it, it’s there, for
good.
Rispoli claims that the
“weight” of she can
fluctuate, so that it could
be “known” but misretrieved later if her
becomes too heavy.
Rispoli (2002) set out
to show that there is a
certain amount of “yoyo’ing” in the
production of she.
We’ll focus on Nina,
for whom we can get
the data.
Nina she vs. her
Rispoli’s counts show
Nina using she from
basically the outset of
her use of pronouns,
and also shows a
decrease of use of
she at 2;5.
2;2
1315
2;3
1619
2;4
2023
2;5
she
her
2
4%
43
96%
1
8%
12
92%
1
14%
6
86%
7
73
Nina’s early shes
2;2
*CHI: she have hug a lady .
*CHI: she have jamas@f on .
These are the
times when Nina
used she (twice at
2;2, once at 2;3,
once at 2;4).
Rispoli found 7 at
2;5, we’ll deal with
them later.
2;3
*MOT: does she like it ?
*CHI: she drink apple juice .
*CHI: her like apple juice .
2;4
*MOT: he's up there ?
*CHI: no # she's not up there .
*CHI: he's up there .
What if we include early files
in evaluating ATOM?
2;2
*CHI: helping her have a
yellow blanket .
*MOT: she has a yellow
blanket ?
*CHI: yeah [= yes] .
*CHI: her's ok .
*CHI: her ok .
*MOT: she's ok ?
*CHI: ok .
*CHI: her's ok .
*CHI: her ok .
*CHI: her's ok .
*MOT: she's ok .
These three and one other
time Nina said her’s ok are the
only candidate
counterexamples at 2;2.
At 2;2, 45 her+bare verb.
At 2;3, no candidate
counterexamples, 14 her+bare
verbs.
(R got 43, possibly including
her’s ok)
(R got 12)
At 2;4 none, 7 her+bare.
(R got 6)
Including early files…?
*MOT: what happened when I
shampooed Miriam yesterday ?
*CHI: her was cried .
2;5:
*MOT: oh # there's the dolly's bottle .
*CHI: her's not going to drink it .
*MOT: I'll start washing it .
*MOT: see how clean it comes ?
*MOT: you want to use the pot ?
*CHI: a little bit .
*CHI: her don't .
*CHI: her's not dirty .
*CHI: not dirty .
I found about 76
her+bare/past
verbs.
I found 3 potential
counterexamples.
Bottom line?
It doesn’t seem like
anything was
particularly affected,
even if Nina’s early
files were fully
included.
The number of
possible
counterexamples
seems well within the
“performance error”
range.
The point about variation in
usage of she is valid, worth
being aware of the
assumptions and being sure
we’re testing the right things.
Rispoli made the point that if
we’d accidentally missed a
she in the early files, we
might have excluded
counterexamples there. Yet,
even including everything, the
asymmetry is strong.
Two hypotheses about
learning (Wexler 1998)
VEPS (very early parameter setting)
Basic parameters are set correctly at the
earliest observable stages, that is, at least
from the time that the child enters the twoword stage around 18 months of age.
VEKI (very early knowledge of inflection)
At the earliest observable stage (two-word
stage), the child knows the grammatical and
phonological properties of many important
inflectional elements of their language.
Very Early Parameter Setting
As soon as you can see it, kids have:
VO vs. OV order set (Swedish vs. German)
VI [yes/no] (French vs. English)
V2 [yes/no] (German vs. French/English)
Null subject [yes/no] (Italian vs. Fr./E.)
So, at least by the 2-word stage, they have
the parameters set (maybe earlier)
VEKI?
Generally, when kids use inflection, they
use it correctly. Mismatches are
vanishingly rare.
English (Harris & Wexler 1995)
German (Poeppel & Wexler 1993)
Again, this is kind of contrary to what the
field had been assuming (which was: kids
are slow at, bad at, learning inflection).
Ok, but…
So: Kids have the full functional
structure available to them, and they set
the parameters right away and know the
inflection.
What then do we make of the fact that
kids make non-adult utterances in the
face of evidence that they aren’t
learning the parameters?
KW: Certain (very specific, it turns out)
properties of the grammar mature.
Root infinitives vs. time
The timing on root
infinitives is pretty
robust, ending around
3 years old.
NS/OI
But some languages appear not to
undergo the “optional infinitive” stage. How
can this be consistent with a maturational
view?
OI languages: Germanic languages studied
to date (Danish, Dutch, English, Faroese,
Icelandic, Norwegian, Swedish), Irish,
Russian, Brazilian Portuguese, Czech
Non-OI languages: Italian, Spanish, Catalan,
Tamil, Polish
NS/OI
What differentiates the OI and non-OI
languages?
Agreement? Italian (non-OI) has rich agreement,
but so does Icelandic (OI).
Null subjects!
Null Subject/OI Generalization:
Children in a language go through an OI stage iff
the language is not an INFL-licensed null subject
language.
NS/OI and Hebrew
(Rhee & Wexler 1995)
Hebrew is a NS language but only in 1st
and 2nd person, non-present tense.
Everywhere else (3rd past, future, present)
subjects are obligatory.
Hebrew-learning 2-year-olds showed
optional infinitives except in 1/2-past, and
allowed null subjects elsewhere, with
infinitives.
NS/OI and Hebrew
(Rhee & Wexler 1995)
% of RIs
kids up to 1;11
null subjects
overt subjects
1/2 past/fut (NS)
0 (of 21)
0 (of 6)
else (non-NS)
32% (36/112)
0 (of 28)
all OI kids
1/2 past/fut (NS)
else (non-NS)
null subjects
0.6% (1/171)
25% (85/337)
overt subjects
1.4% (1/72)
0.6% (3/530)
Where we are
There is evidence that children around the age
of 2:
Allow nonfinite verbs in main clauses—in non-NS
languages.
Differentiate between the syntax of finite and nonfinite
verb forms.
Show evidence from word order of functional
structure above the VP.
Subject case and do-support suggest that TP/AgrP
can be missing from the child representations.
Proposals concerning
TP/AgrP
Wexler (1998): Unique Checking Constraint.
Rizzi (1993/4): Truncation.
For adults, subjects need to check a feature both on T
and on Agr. For kids, only one is possible (so either T
or Agr must be left out, or UCC must be violated).
Predicts: NS/OI. Other reflexes, perhaps object shift
in Korean.
Adults know that CP=root. Kids don’t, so they will
sometimes stop early.
Predicts: If TP is missing, so is CP.
Legendre et al. (2000): Like UCC, within OT.
Implementing the UCC
The basic idea: In adult clauses, the
subject needs to move both to SpecTP
and (then) to SpecAgrP.
This needs to happen because T “needs”
something in its specifier (≈EPP) and so
does Agr.
The subject DP can “solve the problem”
for both T and for Agr—for an adult.
Implementing ATOM
Implementation: For adults:
T needs a D feature.
Agr needs a D feature.
The subject, happily, has a D feature.
The subject moves to SpecTP, takes care of
T’s need for a D feature (the subject “checks”
the D feature on T). The T feature loses its
need for a D feature, but the subject still has
its D feature (the subject is still a DP).
The subject moves on, to take care of Agr.
Implementing ATOM
Implementation: For kids:
Everything is the same except that the subject
can only solve one problem before quitting. It
“loses” its D feature after helping out either T
or Agr.
Kids are constrained by the Unique Checking
Constraint that says subjects (or their D
features) can only “check” another feature
once.
So the kids are in a bind.
Implementing ATOM
Kids in a pickle: The only options open to the
kids are:
Leave out TP (keep AgrP, the subject can solve Agr’s
problem alone). Result: nonfinite verb, nom case.
Leave out AgrP (keep TP, the subject can solve T’s
problem alone). Result: nonfinite verb, default case.
Violate the UCC (let the subject do both things
anyway). Result: finite verb, nom case.
No matter which way you slice it, the kids have
to do something “wrong”. At that point, they
choose randomly (but cf. Legendre et al.)
Minimalist terminology
Features come in two relevant kinds:
interpretable and uninterpretable.
Either kind of feature can be involved in a
“checking”—only interpretable features survive.
The game is to have no uninterpretable features
left at the end.
“T needs a D” means “T has an uninterpretable
[D] feature” and the subject (with its normally
interpretable [D] feature) comes along and the
two features “check”, the interpretable one
survives. UCC=D uninterpretable on subjects?
Are kids really UGconstrained?
So, aren’t TP and AgrSP required by UG?
Doesn’t this mean kids don’t have UGcompliant trees?
Actually, perhaps no. UG requires that all
features be checked, but it isn’t clear that
there is a UG principle that requires a TP
and an AgrP in every clause.
Are kids really UGconstrained?
Perhaps what requires TP and AgrP are
principles of (pragmatic) interpretation…
You need TP so that your sentence is
“anchored” in the discourse.
You need AgrSP … why? Well, perhaps something
parallel…? Wexler doesn’t really say…
Regardless, kids can check all the
uninterpretable features even without TP or
AgrSP; hence, they can still be considered to be
UG-constrained.
NS/OI via UCC
An old idea about NS languages is that they
arise in languages where Infl is “rich” enough
to identify the subject.
Maybe in NS languages, AgrS does not need
a D (it may in some sense be nouny enough
to say that it is, or already has, D).
If AgrS does not need a D, the subject is free
to check off T’s D-feature and be done.
Is there any way to see the
effects of UCC even in NS
languages?
Italian: Mary has laughed.
Suppose that auxiliaries (like have) also have
a D-feature to be checked as the subject (in
the adult language) passes through.
Not crazy: (All) the students (all) have (all) left.
UCC-constrained kids will have to drop
something (the auxiliary or T), even in Italian.
Lyons (1997) reports that a “substantial
proportion of auxiliaries are omitted in OI-age
Italian.”
Ok, maybe. Consistent, anyway.
One open question…
The UCC says you can only use a D-feature
on a DP to check against a functional
category once.
This explains why sometimes TP is omitted
(keeping AgrSP) and sometimes AgrSP is
omitted (keeping TP).
but if GEN infin. comes from omitting both TP
and AgrSP, what could ever cause that
(particularly given Minimize Violations)?
One adult result
If the UCC/ATOM approach is right, this is
one of the only places where we might
actually have evidence that both TP and
AgrSP (exist and) require the subject to
move through their specifiers. Generally,
it’s hard to tell in the adult syntax whether
it’s just one or the other.
Rizzi and truncated trees
Rizzi (1993/4): Kids lack the CP=root axiom.
The result (of not having CP=root) is that kids
are allowed to have truncated structures—trees
that look like adult trees with the tops chopped
off.
Importantly: The kids don’t just leave stuff out—
they just stop the tree “early.” So, if the kid
leaves out a functional projection, s/he leaves
out all higher XPs as well.
Truncation: < TP < CP
If kid selects anything lower than TP as the
root, the result is a root infinitive—which
can be as big as any kind of XP below TP
in the structure.
Note in particular, though, it can’t be a CP.
So: we expect that evidence of CP will
correlate with finite verbs.
Truncation: TP < AgrSP
Pierce (1989) looking at French observed
that there are almost no root infinitives
with subject clitics—this is predicted if
these clitics are instances of subject
agreement in AgrS; if there is no TP, there
can be no AgrSP.
Truncation: TP <> NegP?
There is some dispute in the syntax literature as to
whether the position of NegP (the projection
responsible for the negative morpheme) is higher or
lower than TP in the tree.
If NegP is higher than TP, we would expect not to find
negative root infinitives.
But we do find negative RIs—(Pierce 1989): in the acquisition of
French, negation follows finite verbs and precedes nonfinite verbs
(that is—French kids know the movement properties of finiteness,
and thus they have the concept of finiteness).
So, is TP higher than NegP?
Hard to say conclusively from the existing French data because there
are not many negative root infinitives—but further study could lead to
a theoretical result of this sort about the adult languages.
S O Vfin?
Usually (Poeppel & Wexler 1993) German
kids put finite verbs in second position, and
leave nonfinite verbs at the end.
Occasionally one finds a finite verb at the
end.
Rizzi suggests we could look at this as an
instance of a kid choosing AgrSP as root,
where CP is necessary to trigger V2.
*Truncation?: Where train go?
Truncation predicts: If TP is missing, then CP
should be missing.
But Bromberg & Wexler (1995) observe that bare
verbs do appear in wh-questions in child English. Whquestions implicate CP, bare verbs implicate
something missing (TP or AgrP). So, truncation can’t
be right.
Guasti notes that although the logic here works,
English is weird in this respect: pretty much all other
languages do accord with the prediction.
Theories of missing structure
No functional projections. (Radford) Kids
don’t have any functional projections (TP,
CP, and so forth). This comes later. No TP,
no tense distinction.
Structure building. (Vainikka, Guilfoyle &
Noonan) Kids start with no functional
projections and gradually increase their
functional structure.
Theories of missing structure
“ATOM” (Full competence). (Wexler, …) Kids
have access to all of the functional structure and
have a very specific problem with tense and
agreement that sometimes causes them to leave
one out.
Truncation. (Rizzi) Like structure building but
without the time course—kids have access to all
of the functional structure but they don’t realize
that sentences need to be CP’s, so they
sometimes stop early.
Legendre et al. (2000)
Wexler: During OI stage, kids sometimes omit T,
and sometimes omit Agr. Based on a choice of
which to violate, the requirement to have T, to have
Agr, to have only one.
(cf. “Kids in a pickle” slide)
Legendre et al.: Looking at development (of
French), it appears that the choice of what to omit
is systematic; we propose a system to account for
(predict) the proportion of the time kids omit T, Agr,
both, neither, in progressive stages of development.
Optimality Theory
Legendre et al. (2000) is set in the
Optimality Theory framework (often seen
in phonology, less often seen applied to
syntax).
“Grammar is a system of ranked and
violable constraints”
Optimality Theory
In our analysis, one constraint is Parse-T,
which says that tense must be realized in
a clause. A structure without tense (where
TP has been omitted, say) will violate this
constraint.
Another constraint is *F (“Don’t have a
functional category”). A structure with TP
will violate this constraint.
Optimality Theory
Parse-T and *F are in conflict—it is
impossible to satisfy both at the same
time.
When constraints conflict, the choice
made (on a language-particular basis) of
which constraint is considered to be “more
important” (more highly ranked)
determines which constraint is satisfied
and which must be violated.
Optimality Theory
So if *F >> Parse-T, TP will be omitted.
and if Parse-T >> *F, TP will be included.
Optimality Theory
Grammar involves constraints on the
representations (e.g., SS, LF, PF, or
perhaps a combined representation).
The constraints exist in all languages.
Where languages differ is in how important
each constraint is with respect to each
other constraint.
Optimality Theory: big picture
Universal Grammar is the constraints
that languages must obey.
Languages differ only in how those
constraints are ranked relative to one
another. (So, “parameter” = “ranking”)
The kid’s job is to re-rank constraints
until they match the order which
generated the input that s/he hears.
Legendre et al. (2000)
Proposes a system to predict the
proportions of the time kids choose the
different options among:
Omit TP
Omit AgrSP
Omit both TP and AgrSP
Include both TP and AgrSP (violating UCC)
French v. English
English: T+Agr is pronounced like
/s/ if we have features [3, sg, present]
/ed/ if we have the feature [past]
/Ø/ otherwise
French: T+Agr is pronounced like:
danser
a dansé
je danse
j’ai dansé
NRF
(3sg) past
1sg (present)
1sg past
The idea
Kids are subject to conflicting constraints:
Parse-T
Parse-Agr
*F
*F2
Include a projection for tense
Include a project for agreement
Don’t complicate your tree with
functional projections
Don’t complicate your tree so
much as to have two functional
projections.
The idea
Sometimes Parse-T beats out *F, and then
there’s a TP. Or Parse-Agr beats out *F,
and then there’s an AgrP. Or both Parse-T
and Parse-Agr beat out *F2, and so there’s
both a TP and an AgrP.
But what does sometimes mean?
Floating constraints
The innovation in Legendre et al. (2000)
that gets us off the ground is the idea that
as kids re-rank constraints, the position of
the constraint in the hierarchy can get
somewhat fuzzy, such that two positions
can overlap.
*F
Parse-T
Floating constraints
*F
Parse-T
When the kid evaluates a form in the
constraint system, the position of ParseT is fixed somewhere in the range—and
winds up sometimes outranking, and
sometimes outranked by, *F.
Floating constraints
*F
Parse-T
(Under certain assumptions) this
predicts that we would see TP in the
structure 50% of the time, and see
structures without TP the other 50% of
the time.
French kid data
Looked at 3 French kids from CHILDES
Broke development into stages based on a
modified MLU-type measure based on how
long most of their utterances were (2 words,
more than 2 words) and how many of the
utterances contain verbs.
Looked at tense and agreement in each of
the three stages represented in the data.
French kid data
Kids start out using 3sg agreement and
present tense for practically everything
(correct or not).
We took this to be a “default”
(No agreement? Pronounce it as 3sg. No
tense? pronounce it as present. Neither?
Pronounce it as an infinitive.).
French kid data
This means if a kid uses 3sg or present
tense, we can’t tell if they are really using
3sg (they might be) or if they are not using
agreement at all and just pronouncing the
default.
So, we looked at non-present tense forms
and non-3sg forms only to avoid the
question of the defaults.
French kids data
We found that tense and agreement
develop differently—specifically, in the first
stage we looked at, kids were using tense
fine, but then in the next stage, they got
worse as the agreement improved.
Middle stage: looks like
competition between T
and Agr for a single node.
A detail about counting
We counted non-3sg and non-present verbs.
In order to see how close kids’ utterances were to adult’s
utterances, we need to know how often adults use non3sg and non-present, and then see how close the kids
are to matching that level.
So, adults use non-present tense around 31% of the
time—so when a kid uses 31% non-present tense, we
take that to be “100% success”
In the last stage we looked at, kids were basically right at
the “100% success” level for both tense and agreement.
Proportion of non-present
and non-3sg verbs
40%
35%
30%
25%
non-present
non-3sg
adult non-pres
adult non-3sg
20%
15%
10%
5%
0%
3b
4b
4c
Proportion of non-finite root
forms
35%
30%
25%
20%
NRFs
15%
10%
5%
0%
3b
4b
4c
A model to predict the
percentages
Stage 3b (first stage)
no agreement
about 1/3 NRFs, 2/3 tensed forms
*F2
ParseT
ParseA
*F
A model to predict the
percentages
Stage 4b (second stage)
non-3sg agreement and non-present tense
each about 15% (=about 40% agreeing,
50% tensed)
about 20% NRFs
*F2
ParseT
ParseA
*F
A model to predict the
percentages
Stage 4c (third stage)
everything appears to have tense and
agreement (adult-like levels)
*F2
ParseT
ParseA
*F
Predicted vs. observed—
tense
40%
35%
30%
25%
non-present
predicted non-pres
20%
15%
10%
5%
0%
3b
4b
4c
Predicted vs. observed—agr’t
40%
35%
30%
25%
non-3sg
predicted non-3sg
20%
15%
10%
5%
0%
3b
4b
4c
Predicted vs. observed—
NRFs
35%
30%
25%
20%
NRFs
predicted NRFs
15%
10%
5%
0%
3b
4b
4c