Syntactic bootstrapping Bootstrapping (linguistics)

1 syntactic bootstrapping

1.1 acquiring lexical categories

1.1.1 acquiring verbs
1.1.2 acquiring nouns
1.1.3 acquiring adjectives


1.2 acquiring functional categories
1.3 evidence

syntactic bootstrapping

syntactic bootstrapping theory process of how children identify word meanings based on syntactic categories. in other words, how knowledge of grammatical structure, including how syntactic categories combine phrases , constituents in order form sentences, bootstraps, or encourages acquisition of word meaning. children not need rely solely on environmental context understand meaning or have words explained them. theory, children infer word meanings observations syntax, , use these observations comprehend future utterances hear.

one of earliest demonstrations of existence of syntactic bootstrapping experiment done roger brown @ harvard university in 1957. brown s experiment beginning of framework needed in order theory thrive. took nonsense word (like sib) , asked children 1 of 3 specific sentences containing word (1-3). through experiments, showed children acquire grammar , semantics simultaneously. led linguists lila gleitman, coined term syntactic bootstrapping in 1990, argue syntax pivotal language learning, gives learner clues semantics.

(1) see sib?
(2) sibbing?
(3) see sib?

acquiring lexical categories
acquiring verbs

an demonstration naigles (1990) of syntactic bootstrapping involved showing 2-year-olds video of duck using left hand push rabbit down squatting position while both animals wave right arms in circles.

initial video: duck uses left hand push rabbit squatting position while both animals wave right arms in circles

during video, children presented 1 of following 2 descriptions:

(6) utterance a: duck kradding rabbit.
(describes situation duck rabbit)
(7) utterance b: rabbit , duck kradding.
(describes situation duck , rabbit perform same action)

children presented 2 distinct follow-up videos.

follow-up video 1: duck pushing rabbit
follow-up video 2: duck , rabbit both waving arms in air.

when instructed find kradding , children looked video illustrated utterance heard during initial video. children heard utterance interpreted kradding mean act of duck pushing on rabbit, while children heard utterance b assumed kradding action of arm waving. indicates children arrive @ interpretations of novel verb based on utterance context , syntactic structure in embedded.

in 1990, lila gleitman took idea further examining acquisition of verbs in more detail. in study, found children differentiate between verbs take 1 or more arguments , knowledge used them narrow down potential meanings verb in question. discovery explains how children can learn meaning of verbs cannot observed, ‘think’.

acquiring nouns

the acquisition of nouns related acquisition of mass/count contrast. in 1969, willard van orman quine claimed children cannot learn new nouns unless have acquired semantic distinction. otherwise, word “apples” might refer individual objects in pile or pile itself, , child have no way know without understanding difference between mass , count noun. nancy n. soja argues quine mistaken, , children can learn new nouns without understanding mass/count distinction. found in study 2-year old children able learn new nouns (some mass, count nouns) inferring meaning syntactic structure of sentence words introduced in.

acquiring adjectives

in 2010 study, syrett , lidz show children learn meaning of novel gradable adjectives on basis of adverbs modify them. gradable adjectives have scale associated them: example, adjective “large” places noun modifies on size scale, while adjective “expensive” places noun modifiers on price scale. in addition, gradable adjectives (ga s) subdivide 2 classes: relative , maximal ga’s.

relative ga’s words “big” in (5), , require reference point: big mouse not same size big elephant. shown in (6) , (7), while relative gas can modified adverb cannot modified adverb completely.

relative gradable adjectives
(5) a. big mouse
b. big elephant
(6) a. big mouse
b. big elephant
(7) a. *a big mouse
b. *a big elephant

maximal ga’s words like, “full” in (8); operate on close-ended scale. shown in (9) , (10), while relative gas cannot modified adverb can modified adverb completely.

maximal gradable adjectives
(8) a. full pool
b. full tank
(9) a. ?? full pool
b. ?? full tank
(10) a. full pool
b. full tank

in 2010 study, syrett , lidz showed children pictures of objects described in terms of both relative , maximal ga’s. example, picture of container described both tall (a relative ga) , clear (a maximal ga).

when showing these objects children, novel adjective used describe them prefaced either adverb (which modifies relative ga’s) or adverb (which modifies maximal ga’s). control, in contexts, no adverb present. when novel adjective presented adverb very, children assigned relative ga meaning it, , when presented adverb completely, maximal ga. when no adverb present, children unable assign meaning adjective. shows that, in order children learn meaning of new adjective, depend on grammatical information provide adverbs semantic class of novel adjective.

acquiring functional categories

there basic contrast between lexical categories (which include open-class items such verbs, nouns, , adjectives), , functional categories (which include closed-class items such auxiliary verbs, case markers, complementizers, conjunctions , determiners. acquisition functional categories has been studied less lexical class, remains unknown. 1998 study led rushen shi shows that, @ young age, mandarin , turkish learners use phonological, acoustic , distributional cues distinguish between words lexical categories words functional categories. 11 20-month old children observed speaking mothers evaluate whether speech directed @ children contained clues use categorize words lexical or function . compared lexical category words, functional category words found have following properties:

simpler syllable structures
simpler vowels (monopthongs opposed diphthongs)
shorter duration
lower amplitude
occur more in speech

evidence

a) brown (1957) -- children between ages of 3 , 5 shown various pictures accompanied novel english words. nonsense words included singular nouns, mass nouns, , verbs. brown showed these pictures child , asked them tell him specific nonsense word picture depicted; data collected either noun, mass noun, or verb. when novel words repositioned within sentence , children asked question, focused on different aspects of image shown , adjusted answer. example, when brown wanted child identify mass noun, ask children see sib , , child point @ pictured mass noun or noun indicating quantity. identify verb, ask sibbing , sib verb stem. in order identify singular noun, ask see sib? when children made guesses, correct more half of time. shows children sensitive syntactic position of words, , can correctly associate novel word syntactic category.

b) landau , gleitman (1985) —upon experimenting both blind , sighted children, landau , gleitman found these children differentiate between , see versus touch, despite blind child not being physically capable of looking or seeing. children found associate , see perception, , touch exploration. blind children able learn meanings of vision-related words though not have vision shows used syntax , context infer meaning of these verbs.

c) papafragou, cassidy, gleitman (2007) —participants asked identify verbs within context of video. papafragou et al had children watch 12 videotaped stories. 4 stories subject s desires , 8 stories varied in subject s beliefs , framing of novel verb. @ end of tape, hear sentence describing scene sentence s verb replaced novel word. children asked respond thought word meant. responses categorized 4 ways: action, belief, desire, , other. found action words interpreted children. however, false belief scenes complementizer phrase caused children respond belief words more often. results showed participants in experiment identified verb accurately when use both video , sentence contexts. when comes attitude verbs, children sensitive syntactic framing of verb in question.

d) wellwood, gagliardi, , lidz (2016) —showed four-year-olds can understand difference between quantitative or qualitative word, based on syntactic position within sentence. in “gleebest of cows barn,” novel word “gleebest” in determiner position, , inferred mean “most” or “many.” in “the gleebest cows barn,” “gleebest” in adjective position, , children infer mean “spotty” or quality. these results significant because show children using syntax understand word meanings.

e) gillette et al. (1999) —researchers tested adults see difficulties face when asked identify word muted, videotaped scene. found adults had trouble identifying word, verbs, when refer scene. performance increased once given syntactic context mystery word. these results indicate word learning aided presence of syntactic context.

f) harrigan, hacquard, , lidz (2016) —found children s interpretation of new attitude verb depended on syntactic frame in introduced. in experiment, children heard word hope presented in same syntactic frame want (i.e. followed infinitival verb) connected new verb hope meaning of desire. on other hand, heard hope presented in same frame think (i.e. followed finite verb) made no such association between desire , new verb, instead associating novel verb belief. provides evidence children use syntax extent in learning meaning behind these sorts of abstract verbs.

g) waxman, s. r., & booth, a. e. (2001) —children heard nouns focused on object categories , children focused on adjectives focused on object s properties , categories. shows children sensitive different syntactic categories , can use observations of syntax infer word meaning.