Transcript Lecture: Psycholinguistics Professor Dr. Neal R. Norrick

Lecture: Psycholinguistics
Professor Dr. Neal R. Norrick
_____________________________________
Psycholinguistics
Universität des Saarlandes
Dept. 4.3: English Linguistics
SS 2009
Organizational matters
Klausuranmeldung:
 on HIS LSF POS
July 01-10, 2009
(https://www.lsf.uni-saarland.de/qisserver/rds?state=user&type=0)
According to the Studiensekretariat, sign-ups will
take place for all courses during that time!!!
5.3
Piaget
5.4 Innateness Debate
Chomsky (1986: 150) writes:
What we "know innately" are the principles of the
various subsystems [phonology, syntax, thematic
structure etc.] of So [the initial state of the child's mind]
and the manner of their interaction, and the
parameters associated with these principles. What
we learn are the values of the parameters and the
elements of the periphery (along with the lexicon to
which similar considerations apply).
That is:
We "know innately" as part of Universal Grammar (UG)
that sentences will have noun phrases and verb
phrases in some order, but we have to learn the order.
 Chomsky argues children must know innately what
they can not learn by observation.
Poverty of Stimulus Argument:
Some patterns in language are unlearnable from
positive evidence alone (due to the hierarchical
nature of languages)
You are happy. / Are you happy?
possible rules:
• the first auxiliary verb in the sentence moves
to the front
• the main auxiliary verb in the sentence moves
to the front
but compare:
The girl who is on the bus is happy.
*Is the girl who __ on the bus is happy?
Is the girl who is on the bus __ happy?
 Children don't see sentences like this enough to
decide which rule works but nobody ever chooses
the wrong rule
Grammaticality judgments:
Who do you think Mary knows?
Who do you think that Mary knows?
Who do you think knows Mary?
*Who do you think that knows Mary?
 Note translations!
We need input to learn the whole vocabulary of our
language, including the special syntactic properties of
the vocabulary we learn. We need input to set
parameters like word order, use of cases versus
prepositions etc. and we need input for the periphery,
i.e. all the structures and rules characteristic for the
particular language we hear.
But if input supplies all this information, shouldn't it
supply enough information to learn the basic
principles?
Consider the acquisition of vocabulary:
Webster’s dictionary:
500,000 words
Average educated person’s vocabulary: 40,000 words
(+ another 40,000 proper names, idioms, sayings)
thus: monolingual speakers acquire about 4,000 words
per year or about 10 words every day to age 20
Even within the Chomskyan scheme, there's
debate on whether the principles and parameters are
complete in the newborn child (like the heart) or
whether they develop over time (like the teeth,
which slowly grow and appear, then are replaced
by an adult system).
And even if we "know innately" the principles of
language, are they separate from or integrated with
other abilities?
How the debate developed:
Chomsky develops Poverty of Stimulus Argument
• posits Language Acquisition Device (LAD) and
Universal Grammar (UG)
• claims input is flawed and insufficient for acquisition
input too complicated, contains ungrammatical
structures
 Child Language researchers counter by showing
caregivers using simple, grammatical sentences as
well as repetitions & expansions
Kevin (20 months, 21 days) takes puppet:
Kevin: Dougall. Dougall, Dougall.
Mother: He's a lovely Dougall, isn't he?
Eileen (24 months, 8 days) points puppet
toward television:
Eileen: Skippy a telly.
Mother: That's Skippy on the telly.
Chomsky argues competence would require
"negative evidence“ as basis for grammaticality
judgments--without LAD and UG innate
Al is easy to please It is easy to please Al
Sue is eager to please –
*It is eager to please Sue
Child Language researchers claim kids learn to
make grammaticality judgments only later
(in school) & argue that judgments are based
on semantic factors
It is eager –
unacceptable for nongrammatical reasons
Gold proves mathematically that natural languages
are unlearnable in principle without negative evidence
(unless LAD and UG are innate)
Child Language researchers show caregivers making
corrections, especially in expansions (as above) and
responses
Billy:
Father:
Daddy fixit?
Yeah. Daddy'll fix it for you.
Janik (4,8):
Mother:
Janik:
Mami, ich will mit dich.
Mit?
Dir.
Chomsky-ites argue that caregivers are more
concerned with truth and appropriateness of
kids' talk than grammaticality
They find language communities where kids
receive little if any controlled input or feedback
from caregivers, and they learn language
anyway
Child Language researchers went back to study
input, e.g. as a register like foreigner talk
 this led to study of interaction and hence to kids
developing pragmatic competence, including
interaction between kids
We find kids correct each other from ca. 4 ½ years on:
Nick (4;3):
Coco (2;7):
Nick:
I'm his- I'm a Santa. Who are you?
Santa Mrs.
No, Mrs. Santa.
Come on, let's break the other
people's house down.
Coco (3;2):
Und sie waren in dem Wald in die
Nacht.
In der Nacht.
Nick (4;10):
Due to influence of linguistic pragmatics
(from philosophy of language: Austin, Searle, Grice),
frame theory and richer theories of learning,
Child Language researchers re-emphasized input,
feedback and strategies of learning, especially
negative evidence we find kids not only provide
negative evidence, but even engage in
metalinguistic talk:
Nick: Daddy, Coco hat gesagt güter. Das kann man
nicht sagen, oder?
Me: Nee, was muss man sagen.
Nick: Coco meint besser.
Coco: Nein, güter.
Nick: Nein, Coco, besser. Du musst besser sagen.
Coco: Lass mich, das ich sage.
Nick: ((lacht)) Jetzt hat Coco wieder Unsinn gesagt.
Nick (5;9): Coco, look at these mouses.
Coco (4;1): Mice.
Nick:
Same thing.
5.5 Slobin's Operating Principles &
Universals of acquisition
Whether parts of language acquisition are innate or
not, developing kids seem to follow specific strategies
and their acquisition processes reveal universals
Operating Principles
A. Identify word units.
B. Forms of words may be systematically modified.
C. Pay attention to the ends of words.
D. There are elements which encode relations
between words.
Universal 1:
• postposed forms learned before preposed
forms
• articles before nouns less salient than noun
suffixes
E. Avoid exceptions
Universal 2: Stages of linguistic marking
1. no marking: bird, birdy - singular & plural
2. appropriate marking in limited cases:
bird, birdies – plural
3. overgeneralization of marking: mens, sheeps
4. appropriate marking everywhere
F. Semantic relations should be marked clearly.
Universal 3:
• one-to-one marking is acquired earlier than
compound markings.
• unchanging singular articles like French le are
acquired faster than der/den/dem in German
G. Grammatical markers should make semantic sense.
Universal 4:
• if a group of inflections all mark the same relation,
the child will tend to use the single commonest
form for all cases
• irregular past tenses reduce to dental suffix -ed
Universal 5:
• semantically consistent grammatical rules are
acquired early and without significant error
• kids don't overextend the progressive to
stative verbs, as in
*She's knowing the answer or
*I'm liking cookies
6. Second Language Acquisition
6.1 Contrastive Analysis
 growing out of work by Fries (1945) and Weinreich
(1953) most work on Second Language Acquisition
in the 40's and 50's shared the assumptions of
Contrastive Analysis (Lado 1957)
Contrastive Analysis based on transfer
• from Native Language (NL) to Target Language (TL)
or First Language (L1) to Second Language(L2)
• shared structures facilitate acquisition
• distinct structures cause problems
• positive transfer when L1 and L2 share structures
e.g. Det Adj N structure in NP in English
and German
the mean dog - der böse Hund
negative transfer when L1 and L2 have
different structures
e.g. Adv V NP in German versus Adv NP V
in English
Morgen fahren wir nach hause
Tomorrow we go home
 so research in Second Language Acquisition
tended to revolve around comparison of language
pairs
Language Acquisition was seen as developing a set
of habits to be practiced in accordance with
Behaviorist Theory
but researchers found errors not predictable by
language differences, and the psycholinguistic
process of language acquisition can't be described
solely in terms of linguistic products
6.2 Approximative Systems and Interlanguage
In the 1960's, linguists rejected Behaviorism and
became interested in mentalistic theories
evidence was mounting for a third system between
L1 and L2
Nemser (1971) recognized an approximative system
for the learner with features of both L1 and L2
Selinker (1972) introduced the term Interlanguage
for this individual language system
Interlanguages are highly variable, due to:
• limited cognitive attention, given so much to learn
and remember simultaneously
• learners lack of knowledge of rules
• simultaneous pull from L1 and L2
• they represent transitional stages of development
but L2 tends to fossilize at some stage, due to:
1. Negative transfer from L1
e.g. putting temporal Adv before locative Adv
They went last week to Berlin.
2. Overgeneralization of L2 rules
e.g. extending progressive pattern to stative verbs
I'm knowing him a long time
3. Simplification of L2 rules
e.g. failure to apply sequence of tenses
(or back shift)
I thought it is a joke
•
•
it's often difficult to tell what causes an error, since
these three factors interact
the concern with rules and errors makes
interlanguages spill over into error analysis
research
6.3 Error Analysis
concern with interlanguage and errors it contains and
their relation gave rise to research in Error Analysis
1. Researchers first look for idiosyncrasies in learner's
production
when a learner says: I want to know the English
we must first determine the intention behind it:
either correct expression of desire involving
knowledge of English people or incorrect expression
of desire involving the English Language
2. Then they try to describe the structure in terms of
the grammars of both L1 and L2
I want to know the English
involves an overuse of the definite article from
the point of view of English grammar;
does it reflect the grammar of the learner's L1,
where abstract nouns take definite articles?
3. Finally, they seek to explain the structure as
interference or the learner's hypothesis-testing
if the learner uses this sort of construction
systematically, it's part of an interlanguage;
but it may be a single careless mistake or an
attempt to test this particular structure as well
this attempt at explanation can get muddled, due to
the unclear distinction between competence and
performance
 Error Analysis ends up as a method of describing
data, but not a psycholinguistic theory of language
acquisition
Error Analysis loses sight of the whole picture of
developing competence in L2 by focusing on errors;
• we could instead equate knowledge of L2 with
fluency and understandability rather than lack of
errors;
• or we could instead focus on what learners do right
and test to see if they do it right intuitively
6.4 Innateness, Input, Natural Order of
Acquisition in L2
The Innateness Debate from child language research
carries over to research in second language acquisition
Does the Language Acquisition Device (LAD) work for
L2 as for L1 ?
If the LAD is at work, there should be a Natural Order of
Acquisition in L2 as in L1.
Could L2 learners simply reset the parameters from L1 ?
Dulay & Burt (1973) posit natural order of
acquisition in L2 parallel to what Brown (1973)
found for L1
at least learners with the same L1 background go
through the same stages in acquiring L2
1. plural -s on nouns:
the books
2. progressive -ing on verbs: they driving
3. forms of main verb be:
this is London,
she was there
4.
5.
6.
7.
8.
forms of auxiliary be:
articles a and the:
irregular past tenses:
3rd person sing pres -s:
possessive -s:
she's driving
a cat, the dog
went, ate, came
she waits
Sally's truck
Dulay & Burt (1974) found even greater regularity of
order if features were ordered into groups
Group 1: progressive -ing, plural -s, copula be
Group 2: auxiliary be, articles
Group 3: irregular past
Group 4: regular past, 3rd pers -s, possessive -s
Dulay & Burt use cross-sectional testing, i.e. what
percentage of which forms show up for a group of
learners, while Brown used longitudinal testing, i.e.
at what stage do kids control (90% correct) certain
forms
other problems with tests for order of acquisition in L2
• tests based purely on English: what about other
languages with lots more inflection or no inflection?
• tests failed to distinguish variants like a versus an,
and degrees of irregularity e.g. in past tense
told, bought, went
• if no firm order of acquisition can be shown, then
there's no reason to assume that acquisition of L2
and L1 are alike.
Even if LAD makes input unimportant in L1
acquisition, the status of input in L2 a remains a
problem:
• What kind of input should learners receive?
• Does correcting errors help?