Learning to communicate
" A precursor to communication is gaze coupling
that the child and the caregiver engage in.
" Mothers respond to infants earliest behaviours
as if they were communicative (Snow 1981):
Oh, what a nice burp!
" Infants learn to communicate crying, smiling,
using gestures and intonation.
" Towards the end of the first year a child can
express a rich repertoire of communicative
functions, even though not all of them are at this
level expressed verbally.
2
Prelinguistic development
" Pre-linguistic milestones:

cooing (2-3 months) - earliest vocalizations using non-
language sounds

babbling (6 months) - repeated CV sequences,
e.g. bababa, dadada, mamama

protoword - an idiosyncratic, phonetically consistent signal
with an identifiable meaning.
3
From signals to words
" Learning to communicate verbally depends
crucially on the child's learning to signal. A signal
is a stable form of behaviour produced with the
expectation that it will have some predictable
effect on the interlocutor. The child's signals
often take the form of a gesture, a prosodic
pattern without the segmental phonemes of a
word, or a protoword.
" The first words appear usually around the 9th
month of life.
4
Early syntactic development:
" one-word utterances (9-12 months),
" two-word utterances (12-18 months),
e.g. Daddy shoe, Allgone milk,
" three-word utterances (18-24 months) -
telegraphic speech, i.e. speech lacking
grammatical functors
e.g. Mommy no go.

Mean length of utterance (MLU) is a measure of
syntactic complexity.
5
The wug test
(Berko 1958)
Designed to test the
knowledge of rules of
grammar in children
6
L1 acquisition research contradicting
Chomsky s ideas: The role of imitation
" For the advocates of the Chomskian approach, imitation
was not an important device in language acquisition since
what the child could imitate was surface structures
(performance), while what he had to learn was the
underlying structure and transformational rules
(competence).
" Psychological variables, such as memory, could not affect
the course of language acquisition. Psychological factors
were seen as affecting the performance only, but not
linguistic competence.
" In brief, the child had to acquire a linguistic competence
before he was able to perform in a language.
7
Clark, Ruth. 1974.  Performing without
competence . Journal of Child Language 1: 1-10.
" Case study of Adam (aged 2;9-3;0), focusing on syntactic
development.
" At that time Adam had a tendency to incorporate in his
utterances parts of his own or an adult s previous
utterance, e.g.
 Baby Ivan have a bath, let s go see baby Ivan have a bath.
 Mother: We re all very mucky.
Adam: I all very mucky too.
 Mother: That s upside down.
Adam: No, I want to upside down.
" This strategy was used to reduce the processing load and
allowed Adam to produce longer utterances.
8
Referential (analytic) and expressive (Gestalt)
children (Peters 1977, Nelson 1981)
" Two learning styles have been differentiated in early L1
acquisition.
" Referential children:

early vocabularies consisting mostly of nouns
" Expressive children:

early vocabularies contain a lot of formulaic language,
e.g. stop it, I want it, don t do it
also memorized sentences, such as:
I don t know where it is. (Jonathan, 17-20 months, Brannigan 1977)
What d you want? (Rebecca, 16 months, Nelson 1973)
Open the door. (Minh, 14-19 months, Peters 1977)
9
Referential (analytic) and expressive (Gestalt)
children (Peters 1977, Nelson 1981)
"  These gestalts (postać) have the characteristic of being
wholistically produced without pauses between words, with
reduced phonemic articulation, and with the effect of slurred or
mumbled speech but with a clear intonation patterns
" Gestalt styles are often characterized by dummy words (fillers).
These are nonsense syllables, with no clear referent and unstable
phonetic form, whose only function is to fill an unanalyzed gap in
the sentence frame.
" Sum-up: Referential children learn the language as if constructing it
from smaller building block. Expressive children go for larger
targets, aiming to produce whole sentences, and later analyze
these into component units.
" These two styles can be used by the same child in varying
proportions.
10
Prefabricated patterns (prefabs) in early L2
acquisition
" Prefabricated patterns are parts of language that are learned as
a whole, without internal analysis into the component parts,
e.g.
Lookit, like that.
Looky, chicken.
Lookit gas.
Lookit four.
(Wong-Fillmore 1976)
" Prefabs  enable learners to express functions which they are
yet unable to construct from their linguistic system, simply
storing them in a sense like larger lexical items. [...] As the
learner s system of linguistic rules develops over time, the
externally consistent prefabricated patterns become
assimilated into the internal structure (Hakuta 1976: 333).
11
Imitative speech vs. spontaneous
speech
" Imitative speech is usually more correct grammatically.
Grammatical structures appear in imitative speech earlier
than in spontaneous speech.
" Children often memorize sentences from the input.
Gradually, such chunks are analyzed into component units.
First, different words are substituted for the word used in
the model sentence. Then, the elements of the sentence
frame are gradually  freed .
" Examples of modified sentence frames in L2 acquisition
(Ewert 1998):
I is jumping.
The Simon and Lottie are going in Tom house.
12
L1 acquisition research contradicting Chomsky s
ideas: The input (motherese, child-directed
speech)
" Mothers' speech to their infants is simpler and more
redundant that their speech to adults (Snow 1972). The
MLU is shorter in speech addressed to infants than in adult-
addressed utterances.
" Motherese also contains more repetition and paraphrases,
as well as numerous expansions and extensions.
" Mothers' speech is also more correct and more varied in
that it contains more questions and imperatives than adult-
directed speech (e.g. Newport et al. 1977).
" The recasts serve the purpose of engaging the child s
attention. They are also a way of focusing the child s
attention on the grammatical structures, while no overt
correction is being used (Moerk 1985, 1992).
13