Language
processing in mind
I. the scope of psycholinguistics
--two
core questions:
what knowledge of language is needed for us to
use language?
What cognitive processes (perception, memory
and thinking) are involved in the use of language?
II.
Psychological Mechanisms
2.1
information
processing system
2.1.1
sensory stores
--incoming
information first enters sm, which retain them , for a brief time, in raw
unanalyzed form.
2.1.2
working memory
--limited
in size; it can hold approximately 7 plus or minus 2 whatever units of
information, average 5
2.1.3
permanent
memory
--acts
as a permanent storehouse for meaningful information
--have
a limitless storage capacity
--information
is stored on the base of meaning and importance , not sound.
--when
new information enters wm, some old information will be retrieved from pm to
the working memory so that this new information can be coded as meaningful or
not by relating it to the old information.
2.1.4
summary
--the
information processing mechanism consist of sensory stores, working memory,
permanent memory and a set of control processes.
2.2
central issues
in language processing
2.2.1
serial
processing vs. parallel processing
Serial processing
--Based on the electronic computer, which tend to
executive processes rapidly in a serial manner.
--If a group of processes
takes place one at a time, with none overlapping, it is called serial processing.
2.2.2
modularity
--modularity
has 2 meanings.
1)
independence of the lang processing system
from the general cognitive system. --Eg. Chomsky
--the speech perception is a special lg module . the
properties belong to the perception of speech but not to the perception of say,
music.
2)
the ling subsystems, such as semantics,
syntax, operate independently rather than interactively.
--eg.
in comprehending a sentence, we apply syntactic principles first, than semantic
knowledge. (the interactive view is that we use the knowledge in syntactic and
semantic simultaneously.)
3.1
dimension of
word knowledge
3.2
organization
of the internal lexicon—spreading activation models
3.2.1
the concept of
a semantic network
--Biologically,
the brain is composed of neurons that are connected to other neurons and that
these connections can be either facilitative or inhibitory.
3.2.2
spreading
activation models
--Collins
and Loftus (1975)
--the
organization of the internal lexicon
--such relations are not
equal; some nodes are more accessible than other and the degree of
accessibility is related to factor such as frequency and typicality.
--by
Bock and Levelt (1994)
1) our
knowledge of words exists at three levels.
2) The
conceptual level: similar to the Collins and Loftus model. Consisting of nodes
that represent concept; nodes are connected to other nodes by various
relations.
3) Lemma
level
--a lemma refers to syntactic aspects of word
knowledge (such as part of speech, gender, and subject requirement such an
animate subject?)
--e.g.
4) lexeme
level: phonological properties of a word (phonemes and pronunciation)
3.3
lexical access
3.3.1
cohort model
--by
Marslen-Wilson (1987, 1990)
1)
--designed specifically to account for
auditory word recognition.
2)
finally the recognized word is fit into the
connected discourse.
3.3.2
variables that
influence lexical access
1) word frequency
Low frequency words take
longer time to be retrieved.
2) morphological complexity
Response time
(in experiments such as eye-fixation, retrieving affixes to change word part of
speech) was longer for affixed words than
words without affixes.
E.g. in retrieving
‘decision’, we retrieve ‘decide’ and ‘–sion’ and then combine them.
3) semantic priming
(semantic association)
--Semantic priming: ~ occurs when a word presented earlier
activates another, semantically related word.
4) lexical ambiguity
Lexical
Ambiguity has generated a substantial amount of research because it raises a
number of intriguing Questions.
Summary
(Lexical Access)
1.
Lexical access influenced by factors
including frequency, phonological structural, syntactic categories,
morphological structural, the presence of semantically related words and the
existence of alternative meanings of the words.
2.
Common words and meaning appear to be in a
state of greater readiness then less frequency word and meaning.
3.
ambiguous word: we briefly consider all
meanings of an ambiguous word. However, when a preceding context primes the
most dominant meaning of ambiguous word, lexical access may be selective.
IV.
Sentence Comprehension and
Memory
Most,
you barely notice their structure.
Some,
the wording is so cumbersome that you’ll struggle to unravel what’s been said.
We
often forget the exact words used to convey a message.
Some
sentences linger in our memories for years.
n Comprehension
involves attention to syntactic, semantic, pragmatic factors.
4.
Immediate Processing of
Sentences
4.1
Parsing
4.1.1
Def. – The first step in understanding a sentence is a procedure to assign
elements of the sentence surface.
4.1.2 Immediacy Principle
Tust
& Carpenter (1980): in Parsing we are making decisions though not
necessarily
in a
conscious manner, about which to place incoming words in the tree diagram we
are
building.
4.1.3
Parsing Strategies
4.1.3.1
Late closure strategies
e.g. Tom said that Bill had taken the cleaning out
yesterday.
---
Whenever possible, we prefer to attach new items to the current constituent,
i.e., to reduce the burden on working memory during parsing.
4.1.3.2
Minimal Attachment Strategy
---
We prefer attaching new items into the tree diagram being constructed using the
fewest syntactic nodes.
5.
Memory for
sentences
In natural discourse, it’s unlikely that we can retain all of
sentences accurately since one sentence follows another in a successive flow.
In this part, we’ll examine what we remember and what we don’t
remember for sentences.
5.1 Memory for meaning VS. surface form
5.1.1 Question: Whether we retain the exact form of a sentence
or simply its meaning?
5.1.2 Retention interval in memory
5.1.3 Pragmatic factors
In some case, we seem to remember the
exact form of what’s said, puzzling, confusing, insulting liger for years in
our memory.
--- Holtgraves 1997
6. Discourse processing
6.0 strategies to establish coherence
The comprehension of discourse depends
less on the meaning of individual sentences than on their arrangement.
6.0.1 Given/new strategy
--- this strategy status: a process of
understanding a sentence in a discourse context involves 3 stages:
6.0.2 Direct Match
e.g. Zak hopped into a waiting car and
sped around the corner. He swerved to avoid the parked car and smashed into a
building.
e.g. Zak hopped into a waiting car. The
old car lost a wheel and smashed into a building.
6.0.3 Bridging
--- In some cases, people must make an inference to bridge the gap
between the target sentence and antecedents.
e.g. Last Christmas he went to a lot of parties.
Comprehension: Last Christmas he got absolutely smashed.
This Christmas
he goes very drunk again.
6.0.4 Reinstating old info.
Example:
S1: I’m trying to find a black dog. He’s
short and has a dog tag on his neck that says Fred. Yesterday he bit a little
girl. She was scared but she wasn’t really hurt.
S2:
yesterday, a black dog bit a little girl. It got away and we’re still
trying to find it. He is short…. She was scared but…..
6.1 Schemata and Discourse Processing
6.1.1 Def.
Schema: a structure in semantic memory
that specifies the general or expected arrangement of a body of info. Context schema
Structure schema –
genre
6.1.2 Activation of appropriate schemata (context schema)
--- Lack the appropriate schema: comprehension and memory are
poor.
Experiment: Bartlett (1932) found it’s hard for British college
student to understand Eskimo folk tales.
--- Even if one has the appropriate schema, but if one fails to
activate them, comprehension and memory are poor. (course book P.207)
6.2 Narrative discourse Processing—story grammar
--- Some schemata concerns certain forms
of discourse. A type of discourse which has a characteristic structure genre. Genres provide general
expectations regarding the way info in a discourse will be arranged.
6.2.1 Story grammar
--- ~ is a schema in semantic memory that identifies the typical
or expected arrangement of events in a story.
7. Perception of Written Language
7.1 Levels of written language process
Processing written language exists at 3 levels—feature level,
letter l, and word l
1)
At feature level, the stimulus is
represented in terms of the physical features that comprise a letter of alphabet.
2)
At letter level, the stimulus is
represented more abstractly as an identity separate from its physical
manifestation:
3)
As the word is recognized,
various properties of the words (spelling, pronunciation meaning) become
available to us.
7.2 Eye moments during Reading
All these three above pieces of visual info are extracted
through a series of eye movement.
1)
Saccades:
2)
Regression:
3)
Fixation:
4)
Span of fixation (perceptual
span):
5)
Reading speed is determined by
the duration of our fixations. The span of material fixated and the proportion
of regressive eye movement.
8. Production of Speech and Writing
8.1 Slips of Tongue
--- Collect speech errors in spontaneous speech (live TV or radio)
--- Determine whether there are consistent patterns in when and
how they occur.
--- Why study errors?
8 basic types of speech errors
1)
Shift
e.g. That’s so she’ll be ready in case she decide to hits
it
2)
Exchange
e.g. Fancy getting your model
renosed. (nose remodeled)
3)
Anticipation
e.g. B(T)ake my bike.
4)
Preservation
e.g. He pulled a p(t)antrum.
5)
Addition
E.g. I didn’t explain this
c(l)arefully enough.
6)
Deletion
e.g.
I’ll just get up and mutter (un)intelligible.
7)
Substitution
e.g. At low speed it’s too
light (heavy).
8)
Blend
e.g. That child is looking
to be spaddled. (spanked + paddled)
8.2 Garrett (1984) 5 stages in speaking a sentence
e.g to produce
sentence: She’s already packed two trunks.
1.
Massage-level representation
2.
Functional-level R
3.
Positional-level R
4.
Phonetic-level R
9. Production of Written Language
9.1 Issue: writing product vs. writing process
--- writing-as-a process approach
9.2 F&H
9.3 B&S’s model of the writing-process
--- proposing that: the writing process cannot assume a single
processing model, but should consider different processing models at different
developed stages.
--- focus more an describing why and how skilled and less-skilled
writer compose differently.
--- 2 models: Knowledge-telling model
Knowledge-transforming model.
