Lecture 10 Pod Flashcards
attachment ambiguity
two sentences that seem to have the same surface structure (N->V->N->N->),
they actually have different syntactic trees
(you don’t know until you have additional info what that syntactic structure is)
ex: the police man hit the thief with the stick or the police man hit the thief with the wart
MLU - mean length of utterance
how many morphemes does the kid produce per speech event
synthetic morpheme
pronouns: I gave the test to HER (where “her” tells you female person and object case).
two pieces of meaning in one
problem with synthetic languages and synthetic morphemes
hard to say what the kid actually knows: do they know all the pieces of meaning or just one?
telegraphic speech
stage 1
novel utterances: not exact mimics of adult form: NEVER HEAR ADULTS SAY THESE THINGS
“abby down” for “abby wants to get down”
built around schemas (formulas or templates)
tend to be open class words
phonology
is a branch of linguistics concerned with the systematic organization of sounds in languages.
telegraphic speech schemas
- “gimme x”
closed class words are produced sooner in languages that are…
….more morphologically complex (russian or german)
children have a small range of semantic relations
Bloom
kids have a small range of meanings that they can express and they use templates to produce these short phrases
- agent + action (hulk + smash!)
open class vs closed class
open class are also referred to as “content words”
- they're open class because you can continue to add them indefinitely - nouns, verbs, adverbs - open because they're open to introduction of new words
closed class words (function words)
- articles, determiners, auxiliary verbs, prepositions
- closed to introduction of new words
acquisition of closed class morphemes, which ones come first and why?
- ing
- in
- on
- plural “s”
- possessive “s”
why? NOT FREQUENCY. has to do with syntactic and semantic complexity
can kids generalize the use of a morpheme?
could learn dog and dogs as two separate words or they know dog and know that dogs is plural dogs
experiment? testing whether kids can produce forms that they’ve never heard before by using morphemes
“the wug test” : passing the test means that you have learned which morphemes mean particular things
showed kids a drawing of a wug: “that’s a wug”, “these are two ____?”
kids say: WUGS
nativist views of syntax
overregularization
what the language learning module is doing is acquiring a set of rules
example: over-regularization: evidence that humans are naturally equipped for language because they’re clearly learning rules
- regular past tenses formed by rules: walk + ed : don’t ever store “walked”
- irregulars stored individually (have to store “ran”) because they don’t use rules
connectionist model and learning verbs
rumelhart and mcclelland
basic learning model with no specifically language geared properties
got it to learn both verb rules and irregulars straight from the input
says: you don’t need a brain or mind, that has any sort of special rule learning apparatus built in, you get it just out of a network that’s just forming associations
got it to produce over-regularization: suggesting it’s just the dynamics of the learning process
concluded that: no need to postulate separate learning mechanisms for learning things in language and in other domains
can kids use word order to tell who is doing what to whom?
experiment
YES
1) they showed kids a video: cookie monster tickling big bird or big bird tickling cookie monster
2) then they presented kids with a verbal sentences: either a congruent one with the video or an incongruent one that doesn’t match the video they saw
results: at 17 months (before two word speech!!) they look at the correct picture more than the incorrect one: can use word order to determine the role that each individual is playing the sentence
EXAMPLE THAT COMPREHENSION PRECEDES PRODUCTION
