4>Morphology & syntax Flashcards
Summary of generative-nativist approach to morphology & syntax (5)
- poverty of stimulus
- no negative evidence
- innate grammatical knowledge based on abstract rules & categories
- children “only” need to link rules to abstract catgories
- once link is made, syntactic rules should be applied to all item within a category
Summary of usage-based approach to morphology & syntax> (4)
- both words & syntactic structures carry meaning and can be learned
- linguistic constructions lie on a continuum of size and abstraction
- children derive abstract rules & categories from the input by applying domain-general mechanisms, (i.e. intention reading, analogy & pattern finding)
- children start with small & item-based schemas that gradually develop into more complex & abstract schema
2 sub points of ‘poverty of stimulus argument’>
1>syntactic rules to learn/discover cannot be seen in input we receive
(i.e. yes/no qn formation> hypothesised ‘rule’ doesnt apply to all> so must be innate)
2> children learn grammar even though their input is often ungrammatical
(i.e. false starts/unfinished sentences> so must be innate)
G-N & idea of no negative evidence>
- overcreativity: children going beyond input & producing errors (“he goed”)
- children recover from these, so grammatical knowledge must be innate
G-N what is the innate knowledge?> (2/2)
> structure dependency (principle)
- knowledge that syntactic rules are based on abstract categories & abstract sturctures
(i.e. CP/DP etc)
> word order (parameters)
- as differing across langs but in fairly systematic way
what are the feature of ‘principles and parameters’ in G-N >
- both are based on ABSTRACT syntactic categories
- both are innate
- abstract categories & syntactic sturctures are innate (but links between parameters & categories have to be learnt)
semantic bootstrapping=
theory that children can acquire syntax of a lang by learning & recognising SEMANTIC elements & building upon that
usage based view on grammar>
as a continuum between words & syntactic structures, both carrying meaning & can be learned from input
(development from: small>big & concrete>abstract)
U-B & early development of syntax>
early in development children learn syntactic rules based on ONLY ‘semi-abstract’ patterns that are based on SPECIFIC lexical items
(i.e. “he saw___” <but perhaps not “she/they/i saw ___”)
U-B later in development of syntax>
- gradual development where children discover & form abstract syntactic categories & rules based on what they hear in the input & their pattern finding skills
U-B pattern finding & generalisation> e.g.> (dog)
- 1>see overlap & variation “the dog chased the cat” “the dog chased the mouse”> “the dog chased the”
- 2> see further overlap & variation of prior “the dog bit the cat” > “the dog ACTION PATIENT”
3> see further O&V of prior “he hit me”> “AGENT ACTION PATIENT”
4> see further O&V of prior “she saw him”> SUBJ VERB OBJ
early production syntax: B&F telegraphic speech> (3)
- found lack of grammar in early 2 word utterances, but word order regularities “big balloon”
- some systematicity: possessor followed by possed “daddy coffee”; adj followed by N “big shell”
- overall lack of morphological items & function words
early production syntax> pivot grammar>
claim that children do have function words such as “off” and “on” which they form ‘semi-abstract schemas’ with
(“sock off, shoe off”> __off; “all gone, all dressed”> all___)
Early production syntax: verb islands>
- claims we have patterns that are formed around SPECIFIC verbs & each has own schema; not generalised to all verbs
(e.g. “__is pushing__” A/P; “__ is rolling” just A)
mean letter of utterance (MLU)=
statistic that measure morphemes (free+bound) by counting no of morphemes & dividing by no of utterances
G-N explanation of productivity data>
once a rule is learned (a parameter set) it is applied to ALL lexical items belonging to a specific category (e.g. all verbs)
(“he pushed me”> interpreted as having acquired abstract rule SVO)
U-B explanation of productivity data>
rules start on an ITEM-specific level and GRADUALLY become more abstract
(e.g. “he pushed me” “__pushed__”> indicates that “push” can be used in transitive setence & with different SUBJ but not ALL types of intran too)
Testing productive knowledge of morphological rules: WUG test >
- show 1 WUG labelled
- then showed 2 & state there are two of them, there are two __
- see if ans= “wugs” or not
- finding: difference between preschool & older kids in producing, thus not instant development
problem with WUG test>
-different items tested gave different result
- i.e. for “glasses”, almost all kids got right because was a familiar obj
- indicates rule as “item-specific”
Evidence for productivity: familiar vs novel verbs study>
- 1st a video of lion feeding dog; asked “where is lion feeding dog”
- 2nd a video of frog performing action on monkey; introduce new V “tamming”; ask “where is the lion tamming the monkey”
- ans can only be distinguishing by knowledge of word order in eng (AGENT at beinning)
- children around 24 months performed better with familiar Vs (indicates productivity developes from semi-abstract>fully abstract)
Early comprehension syntax> G-N bunny task>
- bunny & duck on screen; 1 part when bunny does something to duck & 1 part where duck does something to bunny
- asked “where is bunny ‘gorping’ the duck”
-tested ‘prefential looking’ - found early abstraction at 21 months
Early comprehension syntax> U-B bunny task>
- followed up on G-N study
- tested preferential looking
- trained kids in 2 ways: either heard (1) “the bunny is pushing the duck” or (2) “the lion is pushing the horse”; then both followed up with “where is the bunny gorping the duck?”
- if trained with familiar> were able to do task; if trained with different>werent able to do task
What is outcome of U-B bunny task>
- evidence for children being trained to see/find patterns
- rules are not fully abstract but based on what seen & heard before (“the bunny is VERBing the duck”)
Features of senteneces with Overgeneralisation errors>
- cannot be imitated
- evidence that children derived & applied a linguistic rule
- appear around age 3;0
what is the shape of development of overgeneralisation errors>
- u-shaped
- 1st correct based on imitation (“went”)
- 2nd incorrect as become overcreative (“goed”)
- 3rd recovered from being overcreative (“went”)
Alternative factors for how can recover from being over-creative> (3)
- semantic class
- entrenchment
- pre-emption
Alternative factors for how can recover from being over-creative> 1.semantic class
semantic class=if know what V means can guess syntax:
(e.g.”*she giggled me”> ungrammatical as “giggle” is not a causative V)
Alternative factors for how can recover from being over-creative> 2> entrenchment>
- entrenchment= frequent constructions, less likely to extend to novel constructions
(e.g. “she giggled me”> ungrammatical as child only hears giggle in instransitives)
Alternative factors for how can recover from being over-creative: 3.pre emption
- pre-emption= if hear V in construction that serves the same communicative function as a possible overgeneralisation, may infer the overgeneralisation is not conventional
(e.g. “she giggled me”> ungrammatical due to mother using “she made me giggle” this BLOCKS “she giggled me”)