Language Acquisition Flashcards
What is FLA?
The process by which young children come to know and use the language of their caregivers.
What are the biological foundations of language acquisition?
- left hemisphere of brain specialised for language use
- fairly predictable timeline for development
- biologically programmed sensitive period
What are the cognitive foundations of language acquisition?
- need to be able to perceive, comprehend, plan, produce and recall linguistic units
- understand cause&effect relations to use language to communicate
- form mental representations of reality to match words to concepts
What are the social foundations of language acquisition?
- need to experience language use in interaction
- typically developing children communicate with caregiver (proto-conversation: gesture, gaze, noise)
- repetitive social routines which caregivers accompany with speech
First 12 months
first words
12-24 months
v early combinations of words
by 36 months
utterances comprehensible to strangers
features of early combinations?
lacking inflections, auxiliaries missing, no ‘I’ nominative subject, correct word order, correct negation placement
how is FLA achieved (what we know)
- no explicit teaching (even if offered often not obeyed)
- on basis of intake (positive evidence)
- under a variety of circumstances (some don’t receive child directed speech)
- limited amount of time
- very similar process cross-linguistically
what is the task of FLA?
- segment the soundstream to identify components
- learn the lexicon
- learn how to combine words (syntax, semantics, discourse)
The imitation hypothesis and its problems
The idea that children learn language by imitating what adults say.
Problems:
- no 1-1 correspondence between input and output (75% of adult utterances to children are imperatives and interrogatives, children mostly use declaratives.
- children use novel forms such as overregularised forms. Newport, Gleitman & Gleitman 1977
Usage based learning hypothesis
Children learn on an item by item basis
statistical learning + cognitive skills e.g generalisation
Innateness hypothesis
Claims poverty of the stimulus
-principles and parameters rather than language specific infomation.
UG+PLD=I-language
explains cross-linguistic similarities
What kind of data can we use?
Experimental data
Spontaneous speech
Comprehension data
Segmentation of the soundstream
(how?!)
up to 6-8 months can make lots of detailed differentiations
beyond that begin to differentiate just phonemes
Inflections
24-30 months - finite forms only
eventually learn inflections
errors are of overregularisation (use of rules!)
U shaped development: imitate, overregularise, learn irregular forms
negation
1 - negative marker external to the sentence
2 - internal but no finite verb
3 - full form
questions
1 - intonation (on declarative) only
2 - question marker attached to noun e.g where doggy
3 - question marker and non-finite verb
4 - full form
production: word order
generally respect the basic word order of their language (Brown, 1973)
can do recursion
Which features of grammar are the latest acquired? hypotheses as to why?
functional category (including aux) and movement rules.
‘no functional category?’
w regard to verbs this is the q of whether TP exists.
NO: pivot model, ‘small clause hypothesis’ (Stowell, 1983, VPs rather than TPs - are functional cats avaliable in UG but after maturation?)
YES: truncation model: functional category just isn’t realised at this stage (perhaps due to memory capacity)
evidence for the truncation model
also perception data
Guasti:
correct word order in V2 langs implies TP
some examples of finite forms - could just be unanalysed chunks, but Weissenborn 1990 found that negative placement dependencies on finiteness were maintained.
Phonetic processes in child speech - possible explanations:
basically all articulatory simplification
fronting: (why!? - potentially arbitrary distinctions on palate difficult?)
anticipation & preservation - difficulty in motor control prevents quick movement?
reduplication - simplified - same as above?
preference for cardinal vowels?
omission of unstressed vowels - (hard to segment speech stream? simplifies length of word?)
cluster simplification (often continuant that’s deleted - articulatory surely)
stopping (why?!)
CV bias?
What are the differing ideas on BFLA development?
Unitary language system hypothesis (Genesee 1989)
Separate development hypothesis (De Houwer, 2009):
Volterra & Taeschner 1978 believe that it begins as one system, then the lexicons diverge, then the grammars diverge. suggested to happen between 2-3 years of age. (no-one is saying they never differentiate lol)
possible ways in which BFLA languages could interact (not saying they do because they don’t lolll just suggestions)
- transfer - incorporation of grammatical property from language into the other (e.g due to language dominance)
- acceleration - a certain property emerges in the grammar earlier than it would in monolingual FLA
- languages are both acquired slower than in monolingual FLA due to greater memory load
problems with the maturation hypothesis?
that language development is biologically determined
- diff langs have diff rates
- more input -> faster acquisition
language mixing vs code switching
language mixing is a problem of competence whereas code switching is the result of a pragmatic decision (or to fill lexical gaps)
When evaluating evidence in language acquisition, what conditions does your evidence have to meet?
- be systematic
- cannot be explained by error
- cannot be explained by particularities of the input received by the specific child.
BFLA Vocabulary Size
ULS would expect BFLA children to have a combined vocabulary the same size as that of a monolingual child. According to Pearson et al (1993) this is the case, however there are many reasons why this might not indicate a ULS: child’s memory capacity, the fact that they probably do not receive DOUBLE the amount of input as a monolingual child.
BFLA translation equivalents
Because of the ‘mutual exclusivity’ bias (Markman, 1989), ULS children would surely avoid learning TEs until the differentiation. BUT De Houwer et al (2006) found that all children studied knew at least one pair of translational equivalents.
(Villa (1984) notes the possible use of TEs as units with different meanings due to underextension, however Leopold (1970) gives examples of children using TEs during an interaction to refer to the same physical object).
BFLA bilingual blends
Leopold (1970) - where children merge words from different languages ‘bydersehen’.
Suggests ULS however could just be a performance error. (choosing between langs just as monolinguals create blends when choosing between words).
BFLA mixed utterances re lexicon
- generally only involve nouns from another language (Cantone, 2007)
- Paradis & Genesee 1996 don’t find them to be systematic
- Celce-Murcia (1978) finds that the child often doesn’t know the TE
- Deuchar & Muntz (2003) find that children who do not produce mixed utterances have parents who encourage single language utterances as opposed to those with parents who code mix.
Can’t be used as compelling evidence for unitary language system.
BFLA rate of development
Marchman et al (2004) note that a larger vocab in one lang corresponds to greater structural complexity in that same language but not cross-lingustically - this supports SDH as it suggests words used are reserved fro structures in the same language and rate of acquisition (affected by input etc) is limited to each different language.
BFLA structural features
Kaiser (1994) found that from the emergence of func cat, french-german bilingual children had separate verb placement, agreement, tense & case marking in the 2 languages. (Anti ULS)
This said - Dopke (1998) found that German-English bilinguals create target deviant structures in German which reflect the syntactic rules of English. This is not systematic however (De Houwer, 1990 - very small proportion) (could also be influenced by learning from L2 speakers.)
BFLA mixed utterances re structures
De Houwer 1995 does find instances where a bound morpheme is attached to a free morpheme of another language (e.g schraubing) - though not systematically. Also, it is always the free morpheme which is inserted into the utterance rather than the bound morpheme, so a case of lexical gap filling?
Maiwald and Tracy (1994) found a Ge/Eng child who regularly inserted German functional elements into English but this was before the emergence of the Func Cat in English so due to lack of knowledge probs.
Adult SLA: Phonology
- attaining native-like pronunciation is the hardest feature of adult SLA
- also intonation, timing and prosodic features are difficult to acquire
Critical/sensitive period
In UG we would say this is when parameters are set.
Just in general, Lenneberg (1967) finds there is a biologically determined period for language to be learned perfectly (and naturally) due to loss of brain plasticity.
Supposedly this period ends at puberty
Adult SLA and the lexicon
natural input is better than explicit teaching for the lexicon because not only do we need to know word meaning but also collocations.
See transfer phenomena such as I have very hungry (J’ai tres faim)
