Lecture 5

Question 1

Nativist approach

Answer 1

(generativist)
assumes that children approach the task of learning language with innate machinery that is specific to language, sometimes described as a Language Acquisition Device or Universal Grammar (UG)

Question 2

Universal Grammar

Answer 2

children have to be able to learn any language in the world, so it is universal in all children

Question 3

N vs. C creative

Answer 3

• N: children’s utterances are creative because they have access to grammatical rules
• C: - children’s utterances are creative because creativity is based on the use of lexical frames learned from the language they hear, with new items inserted into variable ‘X’ slots. e.g. I want X

Question 4

N vs C: observing word order

Answer 4

• N:- children observe adult word order because thy have an abstract rule (subject-verb-object)
• C: children observe adult word order because they pick up high frequency lexical frames from their input (which, of course, follow the adult word order) e.g. I want a drink

Question 5

N vs C: generalisations

Answer 5

• N: generalisations (e.g. adding inflections to words e.g. wug vs wugs) provide evidence of abstract (innate) rules
• C: generalisations demonstrated that children learn these patterns gradually from distributional analysis of the language they hear

Question 6

Nativist assumptions

Answer 6

• assume grammar is a symbolic computational system which process the relationships between abstract and variables
• assume that grammatical categories and rules are preprogrammed in the child’s brain from birth (UG)
• predict that acquisition of a particular aspect of grammar should have an all-or-nothing quality e.g. as soon as an item is assimilated into a class, the item automatically inherits the privileges of that category - being able to generalise novel words to old words already known e.g. novel verb used in verb context if in that category

Question 7

general prediction from nativist approach

Answer 7

1) children should learn these innate specified aspects of grammar very early on (helping hand from early on with word order)
2) children should show consistent treatment of members of a particular grammatical category

Question 8

the nature of UG: principles and parameters

Answer 8

• all rules for language are innate
• UG rules apply to all languages
• where grammar rules differ across languages, do so in a way encoded by highly constrained parameters (not much difference?)
• children need to work out which parameter settings apply for the language they are learning

Question 9

examples of parameter settings

Answer 9

• word order: verb-object in english whereas object-verb in japanese
• subject-use: in some languages subjects are obligatory (english), in others they are optional (italian)
  (it is raining (eng), is raining (it))

Question 10

theoretical advantages of UG

Answer 10

• avoids problems of explaining how children acquire complex grammar rules
• allows unified theory of acquisition whilst explaining how languages differ

Question 11

empirical evidence for principles and parameters for nativist theory

Answer 11

• children’s early utterances (usually) observe adult word order - taken as evidence for how the relevant parameter is set
• children are productive from early on - evidence they are applying rules of grammar (treat all words in a category in the same way leads to production)
• some evidence that children understand the role of word order from 2 yrs or earlier

Question 12

preferential looking/pointing studies for children understanding the role of words from an early age

Answer 12

• if they understand the sentence, they will pick the right picture even if novel verb
• Getner et al 2006, children below 2yrs can correctly match the picture to subject-verb-object sentace
• evidence for setting word order parameter
• disagreement from constructivisits as to what these results mean - can comprehend vs can’t productce

Question 13

theoretical problems for UG

Answer 13

• parameters not clearly speified: how many are there? which aspects are coded by UG and which are not?
• unclear how children avoid setting parameters incorrectly e.g. informal speech - want a drink?
• bilingualism - how do children set two or more versions of the same parameter?

Question 14

empirical evidence against principles and parameters nativist view

Answer 14

• children display limited knowledge of S-V-O word order in production and act-out studies
• naturalistic data studies provide evidence of a partially-lexically specific knowledge within a grammatical category - don’t extend to all members of the category
• lots of studies show a close relationship between what children hear, how often, and what and when they learn (so do they need these innate rules to learn grammar or can they learn from just listening etc.)

Question 15

interim summary - natavist summary

Answer 15

• N’s approaches provide an account of children’s early multi-word utterances that emphasizes their similarity o adult language
• continuity in P & P accounts, grammatical rules from the very beginning explain development in terms of limitations on performance rather than limited knowledge (adults and children born with same innate grammatical rules, children just have limitations)
• maturation accounts explain why children’s language develop, while maintaining innate grammar knowledge

Question 16

maturation accounts

Answer 16

allow children to show development while still having an innate grammar

Question 17

maturational models

Answer 17

• children’s language develops over time so many researchers argue that this provides evidence that they do not start out with a full innate UG (no continuity)
• solution to this for nativists is to build a part of UG that matures over time according to the biologically-determined timescale (same as puberty) (Radford 1990)

Question 18

Radford’s 1990 maturational model

Answer 18

• at the LEXICAL stage of development (20mo), utterances are mainly content words e.g. verbs, nouns, adjectives, prepositions with other parts corresponding to adult utterance ommited
• at FUNCTIONAL stage (24mo) child’s innate grammar ‘matures’ and parts governing more complex grammatical concepts switch on e.g. auxillary verbs (modality, future, certainty e.g. can/will/must), determiners (a/the) and inflections (-es/–ed)

Question 19

theoretical advantage & evidence for Radford 1990 model

Answer 19

• explains why early utterances are not fully grammatical (not fully adult as does not fit yet)
• allows for development over time so more likely to fit empirical data
• similar trajectory of learning for deaf/blind children despite dissimilar experiences in the world (maybe because of innate grammar)

Question 20

theoretical and empirical problems for maturational model

Answer 20

• can’t easily identify precise point in development when maturing aspects of grammatical system come ‘on-line’
• early stages show use of grammatical functions, but inconsistently and varying across languages
• but 24mo children’s use of ‘functional’ words relate to lexical frames e.g. Don’t X

Question 21

interim summary

Answer 21

• UG as innate abstract grammar, explains language development in children in terms of biological maturation of grammatical system
• BUT does the development reflect gradual learning from input with the ability to be productive in limited ways (constructivist) or whether they have abstract innate grammar

Question 22

the linking problem

Answer 22

• caregivers don’t label particular words as nouns, verbs etc. so how can they categorize them
• (UG supposes that these are defined innatley)

Question 23

what is the solution to the linking problem

Answer 23

semantic bootstrapping

Question 24

what is semantic bootstrapping

Answer 24

• Pinker 1984;1989
• assumes grammatical (syntactic) categories and rules are innate
• children use semantic (meaning) to map words in the input onto these innate syntactic categories by using innate LINKING RULES to map semantic onto syntax (grammatical category)

Question 25

Linking rules

Answer 25

child links individual words to innate categories e.g. noun verb via meaning and grammar

e.g. action = verb

Question 26

problem with linking semantic and syntax

Answer 26

• not all verbs are actions
• not all nouns are concrete objects
• not all subject are agents

Question 27

solution to linking problem and semantic botstrapping problem (can’t always use meaning to map)

Answer 27

• use a form of distributional analysis to determine word order for the language from prototypical sentences. Then apply knowledge of word order to work out the grammatical category of more abstract items

Question 28

advantages of semantic bootstrapping

Answer 28

• explains how children break into innate system
• explains why early utterances follow adult word order
• explains how children learn verbs which are not actions, nouns which are not objects

Question 29

problems for semantic bootstrapping

Answer 29

• many children’s early lexically-specific utterances are not semantically prototypical and therefore are unlikely to be based on innate knowledge of semantic linking rules
  e.g. I want a drink (not prototypical)
• and in passive sentences the noun prase which is usually the object of an active transitive becomes the subject e.g. the cat chased the mouse vs the mouse was chased by the cat
  BUT
• children do hear and use passive sentences from fairly early on, especially in other languages

Question 30

interim summary

Answer 30

• Nativist approach claims innate abstratc UG
• but how do children map grammatical categories - semantic bootstrapping
• based on perceptual understanding of the world around them, and innate linking rules
• approach faces problems in terms of fit to empirical evidence of children’s early utterances

Question 31

Overall summary

Answer 31