S1L8 Grammatical Development

Grammar: Lang is species-specific – to humans - Is it lang that makes us different? Specifically, is it grammar that makes us different?

Successful lang use requires knowledge of: Vocab (words used). Compare John likes Mary w/ John loves Mary.

Grammar: Syntax (word order): Compare John kicks Mary w/ Mary kicks John.

Morphology (word shape): Shapes of word parts, Compare John loves Mary w/ John loved Mary

Morphology: process of combining morphemes to build complex words e.g. indisputable (in-dispute-able)

Morpheme: smallest unit of lang carries meaning. Free morphemes (‘words’) in large numbers & in isolation. Bound morphemes in smaller numbers & attached to free morphemes

Newport (1990) ASL study. 30 Congenitally/prelingually deaf adults - primary lang ASL. ‘Native’ learners - exposed to ASL from birth by deaf signing parents. ‘Early’ learners - exposed to ASL from deaf peers at 4-6yrs. ‘Late’ learners - exposed to ASL from deaf peers from 12yrs. Initially at strict ‘oral’ schools (no ASL allowed). tasks: ASL verbs of motion. Battery of tests of production & comprehension: syntax, morphology Production tasks; short videotaped events, elicit 1 sign or simple sentences, score use of correct morphemes or syntactic forms

Newport (1990) – Age of acquisition: 1st lang Chinese or Korean – uni students in U.S - varying in age of arrival. 23 native speakers of Eng. Gram judgement test; 276 simple short sentences in Eng, half sentences were grammatical, half contained a grammatical violation. Tasks: say whether sentence grammatical or not. Score decreased as age of arrival increase.

Early word combinations: Brown 1973: MLU = Mean length of utterance (number of morphemes), my dog = 2 morphemes

grammatical8 Grammatical stages: Stages of syntactic development (Brown, 1973) Evidence for systematic acquisition of syntactic rules

What errors do children make? Children’s early utterances surprisingly error-free. make errors of omission, not commission: e.g. Put floor, What boy doing? I going come in fourteen minutes. Put it to table. 2 box-ed. Children’s utterances look like simplified versions of adult form – theories of gram dev explain how children progress systematically towards adult lang

Changes in performance: U-shaped learning: over regulation errors e.g. wented, feets

• 2 routes past tense? (1) Pinker, Marcus: Child initially learns words in context but unrelated to eachother. Child learns rule but overextends. Child w/draws from overextension. Listing of exceptions  blocking, regular route  apply rule. (2) Mature state: Detect word, look up in lexicon. If exception, apply learned mapping. Otherwise, inflect by rule(s)

theoretical issues in lang development

Behaviourist account – learning theory: Children learn lang through experience: Skinner’s Behaviourism: Children learn to produce grammatical sentences in same way y learn words – through imitation & reinforcement. “In all verbal behavior under stimulus control therere are 3 important events to be taken into account: a stimulus, a response & a reinforcement. se are contingent upon each or… 3 term contingency… is exemplified when, in presence of doll, child frequently achieves some sort of generalised reinforcement by saying ‘doll’.” (Skinner, 1957). Parental approval or disapproval indicates gram of child’s utterances, reinforcing correct utterances.

Chomsky’s Universal Grammar & Lang Acquisition Device: Chomsky (1986): Experience of lang cannot account for linguistic productivity = ability to produce & understand sentences never produced or heard before. Productivity implies implicit knowledge of syntactic rules e.g. Colourless green ideas sleep furiously *Green sleep colourless furiously ideas. Thus, children genetically endowed w/ a Lang Acquisition Device (LAD): LAD specifies fundamental gram principles shared by all langs (Universal Grammar). LAD ability to acquire lang, is innate, domain-specific & uniquely human

LAD: Lang Acquisition Device: Universal, Nativist, Modular, Minimal role played by child’s experience of lang

Principles & parameters: Principles: universal rules w/ which child starts out. Parameters: set by lang child hears. Parameters are set, principles applied e.g. Principle: Ev sentence has a subject. Principle: Some langs allow subject to be dropped. Parameter: Whether subject must be specified in lang. Child hears ‘I don’t speak English’ or ‘je ne parle pas français’ or ‘_ no hablo español’ & sets parameter as null subject or non-null subject accordingly. Principle can be applied to all instances in lang

Argument from Poverty of Stimulus: Chomsky’s claims about Universal Grammar stem from argument Poverty of Stimulus: + evidence is under-specified, No - evidence. Yet children progress rapidly to full mastery of syntax, making surprisingly few productive errors along way

BUT does negative evidence exist? (1) Explicit disapproval? (Brown & Hanlon,1970); No evidence that approval or disapproval are contingent on syntactic correctness. Parents only correct semantic errors (2) ‘Subtle negative feedback’: Parental repetitions/ recasts more likely after errors (3) Direct Contrast Hyposis (Saxton, 1997) parents correct grammar

Is Negative Feedback Used? Direct Contrast Hyposis (Saxton, 1997); 5-year-olds trained w/ novel verbs e.g. neak/noke (repeated clapping motion in which an object is trapped btwn palms) – like speak/spoke. 2 testing groups; - feedback & + demonstration by adult. Negative evidence (correction) resulted frequent subsequent correct use of past tense form compared w/ + demonstration. So: - feedback is used! (sometimes)

Specific Lang Impairment: Larry Leonard ~7% 5-year-olds show specific lang impairment (SLI; Tomblin, 1996), defined as lang skills significantly below age norms, plus normal-range IQ, hearing, motor & neurological function. Characteristics include: Poor vocab knowledge (link to ‘late talker’ status at age 2). Omission of gram morphemes (“I need go now”). Diffi interpreting subtleties in lang (‘pragmatics’) In some cases, articulation problems. Delayed, rather than deviant lang but unlikely to catch up (link to learning disability in adult)

What is underlying problem in SLI? - 4 candidate explanations of SLI (Leonard, 1998): 1. Incomplete grammatical rule systems? immature gram dev (Rice & Wexler, 1975). inability to induce grammar, & hence reliance on rote learning (Crago & Gopnik, 1994). controversial argument is genetic i.e. a specific impairment to Lang Acquisition Device itself 2. Difficulty processing grammatical morphemes? Surface Hyposis (Leonard, 1992) difficulty perceiving & using low salience or low frequency morphemes e.g. box vs. rocks. grammar-specific processing impairment – because of unique difficulty of hearing, processing & linking se low salience, brief bits of info 3. Auditory processing deficit? difficulty w/ processing rapid auditorily-presented info (Tallal & Stark, 1981) i.e. impairment of auditory processing, not specific to lang (‘domain-general’) 4. Generally limited processing ability: Generalised Slowing Hyp (Kail, 1994; Miller et al., 2001). Limited working memory capacity (Garcole & Baddeley, 1993) i.e. general processing impairment